The concept was simple. The public were invited to submit their “refreshing ideas that change the world” to the Pepsi Refresh Project website. The first 1,000 posted within the month were  eligible for  funding. Grant money ranging from $5,000 to $250,000 was awarded at the end  of each month  depending on which idea got the most votes from the general public. Facebook  Connect, a Facebook app and the usual share buttons were all used to encourage participants  to share their idea with the world and garner support.

A noble idea but did it work?

Assessing the campaign using the staple social media metrics would suggest that the Pepsi Refresh Project was an overwhelming success:

• Over 3 billion earned media impressions
• Nearly 5 million site registrations
• 63 million votes

A ten-fold increase in the number of Pepsi Fans on Facebook.

Wow.

But did the campaign actually motivate people to buy Pepsi? The company’s 2010 trading numbers tell a different story:

• Sales declined by 4.8%
• Market share declined by 0.6%
• For the first time in history, Pepsi-Cola fell below Diet Coke from no.2 to no.3 in the market share rankings.

There is no evidence that can categorically point the finger at the decision to shift budget from traditional to social media but it would be foolish not to heed lessons from this: social media success does not necessarily mean success for your bottom line.

The problem with the Pepsi Refresh Project is that the common interest that united the brand and its campaign participants was a desire to do something good for the world, not a mutual love for Pepsi Cola. Was the association with the product itself strong enough to compel participants to break habit, overlook the Coca-Cola can on the shelf and reach for the Pepsi Cola instead? Did those who consumed the brand in one of those 3 billion earned media impressions sufficiently identify enough with the brand and its product to become a new Pepsi customer (or increase their consumption of Pepsi)?

Compare this to the Skittles Facebook Page. While creating Skittle covered American Football helmets and dresses made out of Skittle wrappers, the Fans of the Skittles Facebook Page are united by one thing: their love of Skittles. And with every Like, every comment and every share these 21million Fans are sharing their love with their network of Friends and Family on Facebook.

Not every brand is going to command the same level of passion as Skittles but every brand can follow this simple rule: ensure that your social media proposition is consistent with your brand proposition. While driving earned media and interactions with your brand is no doubt valuable, you should not do so for the sake of it. Interactions and earned impressions are meaningless unless they pull together to support communication of – and bring to life – the overall brand proposition.

About Rob White

Mail | More Posts (1)

While the challenge of informing investment decisions has always been a part of advertising – from George Gallup introducing research techniques in the 1930s to the development of yield management tools for online publishers – it has acquired a real sense of purpose and scale through such methods as Real Time Bidding (RTB).

Big Data comes hand in hand with a Big Story.

Billy Beane’s players and staff did not adopt his statistical methods until he took the time to tell each of them a story around the numbers. The A’s front office not only recorded and analyzed every pitch thrown to Oakland A’s hitters, but related every plate appearance to a hand of blackjack; the tone of it changed with each dealt card (pitch). The same applies with online media and the footprint of data we deal with every day. When numbers acquire the significance of language, they are able to do all things which language can do: become persuasive and tell a story. I personally have found hypothesis planning to be quite effective – framing a commercial challenge on the basis of a series of story based assumptions that are continually tested and validated through data. Real Time Bidding has magnified the challenge of good storytelling in our industry, be it through the growth of Data Management Platforms or aggressive data monetisation strategies by publishers. The most important challenge, though, comes from the ability to transform, interpret and visualize data to make better decisions.

The relative value of experience.

At the heart of the Moneyball story is the premise that experience is over-rated. Put simply, the collective wisdom of baseball insiders (including players, managers, coaches, scouts, and the front office) over recent decades was subjective and often flawed. I dare say the same applies in our industry. In the current ecosystem, we no longer can afford to use our ten, fifteen, twenty years of experience as the single point of differentiation. The democratization of information, the demand for interoperability and the changes in structures, resourcing, skills and culture, forces us to continually learn and challenge ourselves – not through our past experience but with our ability to anticipate behaviour. Anticipating behaviour is hard as a baseball manager or a marketer – human beings do not always act on or make rational decisions. However, behavioural economics can provide some valuable clues and insight. And while behavioural economics demands a separate posting (given the significant breadth and scope), it shows by and large that every decision people make (be it as a baseball player or a consumer) is massively affected close to the moment of decision by the context in which they decide.

The value of information sources.

Moneyball chronicles not only the initial indifference of baseball insiders to statistical analysis, but also the progressive availability of baseball statistics. For decades, baseball statistics were collected through rudimentary score sheets and box scores controlled by a handful of organizations such as the Elias Sports Bureau. The creation of STATS Inc., which aimed to “set down the primary events that occurred in a baseball game as completely as possible” started to change things. Faced with low adoption by the organizations that should have embraced this new body of knowledge, STATS Inc. gave up on trying to sell the data directly to teams and began selling it to fans.

The evolution of baseball statistics mirrors the evolution of audience and inventory data in online media. In the early days, audience and inventory data was owned and controlled by individual publishers and research organizations that were more likely to approach clients directly. Fuelled by RTB, the availability and volume of audience and inventory data today has grown exponentially.

Yet issues still remain: while we have access to more data, questions about who owns the data are yet to be addressed; audience data is not easily transferable between Data Management Platforms and Dynamic Creative Optimization platforms; a large portion of audience data U.S based (a concern when dealing with global clients); no standardisation of data (classification and taxonomy) across multiple data sets, and there is limited social media audience data at hand. Increased access to quality audience and inventory data provides us with two fundamental benefits – the ability to generate new data and inform decision making about investments.

The “power of no”.

Billy Beane and the Oakland A’s created a competitive advantage through better decision making – statistical analysis allowed them to exercise the “power of no” over the market by cherry picking undervalued talent based on their own set of criteria. RTB provides us the same “power of no” – by allowing us to purchase on an impression basis, and completely bypass on impressions that do not meet our own specific criteria and goals. Buying at the impression level also means de-averaged pricing, allowing buyers to set individualized price parameters for each impression based on expected return. The ability to refine our buying decisions – taking full advantage of the impression “long tail” – is only the start of the process. The real challenge is building scale and sustained return on investment across time.

Focusing on the metrics that uncover hidden value.

The principles adopted by Billy Bean in Moneyball were first developed by Bill James, the author of The Baseball Abstracts, a book that unknowingly became the scriptures of the Sabermetric movement that has shaped the landscape of player evaluation in baseball. James questioned the value of using highest batting average to judge the success of a team, player or manager, so he set about developing an alternative metric that would recognize the single most important objective of an offense – the number of runs a given player was able to generate or contribute to.

Even more valuable than the equation itself was the realization that most baseball organizations did not place enough value on the actions that led to a run such as walks and extra base hits. The same dynamic applies to the way in which most online media investment objectives are measured and managed. When discussing and planning return on investment goals, for example, most organizations ignore the importance of the conversion path in achieving that return. While considerable progress has been made in segmenting the value of prospecting vs. re-targeting investments, attribution planning and analysis seemed to be relegated to analytics reporting as opposed to marketing plans. The challenge with attribution planning is that there is no single best model or “one size fits all” formula: it requires a bespoke model that fits the dynamics of each individual organization and the behaviour of its customers. Return on investment is the sum of individual actions, and those actions must be identified, attributed and more importantly continually optimized to uncover value.

About Alain Portmann

My job is simple. To inspire staff, clients and consumers.

Mail | Web | More Posts (65)

When you’re on a fast break, a lot of good things are happening. You’re rewarding your defense for their hard work. You’re outrunning your opponent. And you’re scoring points very effectively with a high percentage layup or dunk. But the metaphor doesn’t end there.

A successful fast break requires communication and coordination among teammates. You move the ball with a lead pass. That is, you don’t aim for where your teammate is standing, but rather ahead of them, to a space between them the basket. It’s often hard to get your two points without the lead pass.

What’s the significance for digital marketers?

In the marketing world, the equivalent of a lead pass is a strategic roadmap that takes future growth into consideration. Knowing that the fast-moving digital world can lead to unforeseen surges in demand for your product, you score fast break points by thinking a few steps ahead and coordinating your assets to best position your brand for success in the future, and not just the present.

For example, consider a trend like gluten-free foods. Millions of Americans suffer from celiac disease, which makes it difficult or impossible to digest a certain protein found in wheat. The food industry has latched onto this trend, bringing a variety of gluten-free products to market. However, a very small share of celiac sufferers are presently aware of their affliction, and thus the search behavior associated with a keyword like “gluten free” is nowhere near what we’ll see in the future.

Yes, there are a great many instruments out there to size up search engine marketing opportunities, but what you’ll usually find is a range of resources built around historical data. These offer a great way to gain perspective on how search activity has built up over time, and can certainly put you on the path of setting realistic objectives (and budgets) today, but what is it doing to prepare you for tomorrow and the next day?

Not much. Backward-looking insights will only provide a limited amount of forward momentum. Think of it this way: If you were President, would you nominate a Civil War buff to the post of Secretary of Defense?

Yet it is possible to take two halves and a make a whole. It starts with the Google AdWords Keyword Tool and Google Insights, two of the richest and deepest troves of data available to digital marketers today. But these two don’t always play nice: one tool quotes numbers in raw query volumes, but only provides a maximum twelve-month trend; the other goes back as far as 2004, but only reports an opaque index metric useful only to judge relativity.

If there were a way to marry the two – measuring a long historical trend on an apples-to-apples basis, and then projecting forward – that would certainly add color to a great many marketing discussions.

With our Search Query Volume Estimator, that’s exactly what we’ve set out to do. Below is an example using the keyword “gluten free”.

How does the Search Query Volume Estimator work?

Actually, it requires just a handful of ingredients:

1) Your search keyword of choice (your brand, your category, a core customer action reflected in search behavior, etc.)
2) Last full month’s search query volume figure for that keyword (as reported in the AdWords Keyword Tool)
3) Pasted CSV output from that same keyword’s Google Insights data

And what do you get in return?

What you get is a visual report with a nice dotted line telling you where projected search query volumes will go in the coming months (our current version projects all the way through the end of 2014). You’ll also get treated to a table showing yearly estimates of raw search query volumes, as well as an aggregated chart indicating the seasonality profile based on available data.

It comes out looking like this:

In other words, this is your lead pass – and if you’re taking the steps to capture this growth across digital channels… well I don’t have to tell you what comes next.

Limitations

Like any tool in its first iteration, clearly there are some weaknesses to the Search Query Volume Estimator. The most obvious is that by basing future predictions on past growth trends, we overlook the effect of emerging technologies or other interruptive changes in the marketplace. If anyone out there has premium access to that kind of tool, let’s do lunch.

Another weakness is that version 1.0 currently requires the user to use the entire trend (Jan 2004 to present) as the basis for calculating growth trajectory. But if you’re selling Jeremy Lin jerseys, well… in 2004 he was still in Driver’s Ed. We’ll look to enhance this in a future update.

There are plenty of other things we could do to build on version 1.0. Wouldn’t it be nice to aggregate multiple keywords, or even show competitive keywords side by side? AdWords Keyword Tool also gives us CPC and competitive data, which could one day be rolled in.

If you think of any others, please let us know in the comments.

Conclusion

The Search Query Volume Estimator is directly applicable to search marketers, but is relevant far beyond just the search channel. After all, what causes growth in raw search queries in the first place? It is channels like display advertising and social media, where the brand extends its reach and builds a following. With this tool, you get an empirical basis for making bigger and bolder investments for your brand – based on what’s happening tomorrow, and not just today.

So the next time you’re trying to close a complex sale, just ask yourself where you should be aiming your next pass on the fast break. The answer is forward.

To download the Search Query Volume Estimator, please click below. (If you run into problems using this tool, please leave a note in the comments.)

About Paul Burani

Paul Burani - Partner, Web Liquid Group. Connect with me on Google+

Mail | Twitter | LinkedIn | Google+ | More Posts (55)

It’s January, and all the industry pundits are buzzing with their biggest and boldest predictions of what’s in store for the year. But let’s talk about what we’re not going to see in 2012: the return of user search query data from individuals signed into Google accounts.

Back in the fall, Google made a landmark announcement: in an effort to provide greater security to users of Google products, they would by default encrypt search queries entered by any individual who is signed into their Google account. All of a sudden, things got a little blurrier for digital marketers.

The search marketers’ neighborhood of the blogosphere immediately went ablaze, with the SEO community in particular crying foul over the deprivation of a significant amount of critical user data.

Research by Conductor indicated that about 9 percent of search queries were affected by the changes. Google spokespeople said single-digits were to be expected, but then HubSpot put the figure at 11 percent. Some people reported 20 percent or even higher. [For a comprehensive overview, see Jonathan Allen’s summary, “SEOs Strike Out as Google Encrypts Signed-in Search Data.”]

There are numerous implications to this announcement (and the industry repercussions that tend to follow big announcements from Mountain View), but this article isn’t intended to lament or speculate on the changes. Rather, the industry needs to think more about how to embrace the changes and move forward.

Marketers need to think more concretely about search queries as user-generated content (as we argued back in May 2010, “Estimating Word-of-Mouth Activity from Search Query Data”). As search engines and social media environments begin to look more alike, the lines will blur between these data inputs as well.

We know that Google’s strategy has been to always put users first, and this move toward encryption of search query data should be considered a long-term investment in the loyalty of Google’s user base (or, as Allen says, “an attempt by Google to barter web data in order to leverage security as a quality control factor”).

Consumers today are a savvy lot; in a short time they’ve learned how to meet most of their critical needs: hunting down obscure information, filtering content according to preference, engaging in deep exchanges of value. Privacy is not just a buzzword in Silicon Valley, Silicon Alley or Capitol Hill; it’s also one more basic human desire that consumers are smart enough to protect (or trade, for the right offer).

While it remains to be seen whether this secure data will make a comeback (perhaps in Google Webmaster Central or in the enterprise edition of Google Analytics), in the short run the brands that are able to wean themselves off this data will be the ones best positioned to forge ahead. The proactive solution is to build a comprehensive profile of the “Secured User”, to gain a better understanding of the faces who will be absent from the crowd:

Based on this sample data set, we can make a confident statement about our Secured Users on this particular website. Compared to their Unsecured counterparts, they are:

• More familiar with the website (and possibly the brand itself).
• More likely to reside in North America.
• More likely to land on a deep URL after a search query (instead of the home page).

This type of comparative insight paves the way for other executions that can help close the gap, for example:

• Reach more Secured Users in another Google environment, such as Google Plus, and then use social media monitoring tools to collect data on their interactions. It might not be search query data, but word-of-mouth data is still essential for meeting a wide range of objectives.
• Ramp up your efforts to attract more visits from Mobile devices – because for now the encryption doesn’t yet apply on phones and tablets. Create dedicated mobile-specific campaigns in AdWords, and offer mobile-optimized content for a good user experience and SEO profile, to help to make this happen.
• Adapt your destination URL strategy in AdWords paid search and/or Google Display Network. You might test for different URLs tied to branded terms. Or maybe there is a specific deep URL that skews high with Secured Users; building a retargeting campaign specifically around visits to this one URL is one way to wrestle back some influence with this crowd.
• Connect your Google Webmaster Central and Google Analytics accounts. It’s not as pretty as what we’d gotten used to, but you can still use Google Analytics to browse top daily search queries and landing pages, with the standard array of performance metrics: impressions, clicks, click-through rate (CTR), and average position.

Through all of these growing pains, remember one important thing: there’s virtually nothing about Google’s change that will impact the demand for your products and services. Users will continue to use search engines as a primary component of their online experience.

Google will continue to be the search engine of choice in the U.S. And it’s unlikely we’ll even see a change in the share of visits sourced from organic search, compared to paid search.

Let’s just take this change for what it is: a gift from Google to its users. Users will repay Google with their loyalty, which ensures their long-term status as the stable audience the digital marketers need to keep brands happy, growing… and secure in their future.

About Paul Burani

Paul Burani - Partner, Web Liquid Group. Connect with me on Google+

Mail | Twitter | LinkedIn | Google+ | More Posts (55)

Fortunately, most Search Engine Watch readers know at least the basics of using a search engine, and we’re not 7-year-olds drooling over the hottest toy of the 2011 holiday season. (Although, with just the right holiday cheer, we too can go into orbit like the kid who got the Nintendo 64.)

But the truth is, with the way search engines are evolving toward personalization and predictive modeling, it might not be long before any one of the aforementioned keywords might suffice in delivering relevant search results to users.

There’s a curious trend going on behind the scenes, one that hasn’t received a whole lot of publicity: the growing prominence of one-word search queries. We’re seeing one-word search queries’ share of total search activity growing fast: from 20.3 percent of all search engine queries  in January 2009 to 27.2 percent in October 2011, according to Hitwise.

That’s more than a one-third increase in percentage terms – in less than three years. Are things changing so fast in our digital ecosystem, that our impatience can be measured in such dramatic terms?

Maybe not. Let’s see if we can break down this trend and isolate the behavior for what it really is.

Growth of Mobile Search?

The first explanation that comes to mind might be that the meteoric rise of mobile search. With almost a quarter of mobile searchers inputting queries almost every day, and with the short form that lends itself to the mobile format, you might have a real insight on your hands. Alas, that Hitwise data excluded mobile devices.

Belated Effects of Google Instant?

There’s no doubt that the launch of Google Instant in early September 2010 touched off a sea change in search behavior, as hordes of search engine users gleefully rescued nanoseconds back into their calendars – and the occasional multi-word search query got truncated by the pre-emptive delivery of some useful search results.

Various people tried to quantify this effect. ComScore’s Eli Goodman probably came the closest, and in his verdict, only 3 percent of search queries saw a reduction in word count. Maybe that’s why Hitwise data showed a decline in the incidence of one-word search queries in September 2010, to 22.7 percent.

So we’re barking up the wrong tree there, too.

More URLs as Keywords?

When faced with the home screen of Google.com, a lot of users still are compelled to enter a URL as a keyword input – and those are generally counted as one-word keywords. With browsers emulating Google Chrome’s singular input field for both URLs and search queries, maybe this is in fact changing the face of search behavior.

But knowing what you know about the sophisticated online consumer, are you going to settle for that explanation?

Improvements in Geographic Targeting Search Engine Users?

Search engines have become a lot smarter about understanding the similarity between a search query for “leappad brooklyn,” and “leappad” queried from a computer or mobile device located in the Brooklyn area. If users have started to catch on that simply typing “leappad” delivers them a litter of Nearby Stores links among the search results, then yes, that could help to explain the downward pressure on query length.

Something Happening Further Down the Conversion Funnel?

We can deconstruct these trends until we’re blue in the face, but in all likelihood the behavioral insight here has less to do with the search engine, and more with the resulting action. About a year ago, we wrote about the relationship between search query length and predictive keyword research, and cited research that illustrated a correlation between the number of words in a search query, and its resulting conversion potential.

Yet one would assume that data analysts and economists at Google, Bing, and the like would have spotted this trend and started clamoring for a re-engineering of the search algorithm toward this behavioral sea change.

If long search queries are the highest converters, and search engines are seeing more one-word search queries than ever… are we at the nexus of the universe?

Or if you’ve got a simpler explanation, let’s hear it!

About Paul Burani

Paul Burani - Partner, Web Liquid Group. Connect with me on Google+

Mail | Twitter | LinkedIn | Google+ | More Posts (55)

______________________________

• Background
• What is the role of cookies?
• What is the new legislation on cookies in the UK?
• When is the legislation on cookies coming into effect?
• What should you do?
• Our view on the UK cookie legislation
• The different forms for obtaining consent
• Recommended next steps
• How we can help
• References & Sources

______________________________

Background

Earlier this year, the UK government implemented an amendment to the EU’s 2003 Privacy and Electronic Communications Directive. The “UK cookie directive” requires businesses running web sites in the UK to get informed consent from visitors to their web sites in order to store and retrieve information on users’ devices. {1} While the new legislation is a significant change from previous policies, the government’s view is that there should be a phased approach to the implementation of these changes.

What is the role of cookies?

Cookies are the most obvious manifestation of ‘store and retrieve information’ and are the centre of the new legislation. Cookies are text files that are dropped in users’ hard drives by the server being accessed (through the browser) when they access a site or view an ad.

It is important to note that the consent rule does not apply to all cookies. At the present time, the only exception to this rule is if the cookie is ‘strictly necessary’ for a service requested by the user. This applies to cookies used throughout a shopping cart, including ‘add to basket’ or ‘proceed to checkout’. {2}

What is the new legislation on cookies in the UK?

In simple terms the new “UK cookie directive” require a user’s consent if an organization wants to store a cookie on their device. The revised rules replace the requirement of the 2003 Regulations that users must be 1) told how cookies are used and 2) given an opportunity to refuse cookies (an “opt out”) with a requirement for user consent.

When is the legislation on cookies coming into effect?

The legislation came into effect on the 26th of May 2011. Given the difficulties in establishing specific guidelines, organizations have until May 2012 to comply. {3} These timelines might change but on the date of publishing this document, these are the timeframes set forward.

[Click here for Part 2 of 'The future use of cookies in the UK: From informing to obtaining consent']

About Alain Portmann

My job is simple. To inspire staff, clients and consumers.

Mail | Web | More Posts (65)

What should you do?

Unfortunately there is not a defined set of exact guidelines to ensure compliance. The main reason is that the Department for Culture, Media and Sport who are responsible for drafting the exact steps that businesses should follow have not finished these.

However a series of industry bodies and organizations including the IPA and more importantly the Information Commissioner’s Office (ICO), have issued initial recommendations to ensure organizations are prepared. This document serves to outline key considerations as well as recommend a set of initial actions to ensure readiness once specific guidelines have been defined.

Our view on the UK cookie legislation

The legislation will not stop users receiving online ads. More importantly, the legislation does not forbid the use of cookies that are not considered “strictly necessary” – it simply requires organizations gain consent by giving the user specific information about what they are agreeing to and providing them with a way to show their acceptance. In most cases consumers will be happy to allow you to track them if it makes their experience more enjoyable and useful.

When considering the requirement to obtain consent it is important to note:

1. Organizations are required to provide information about cookies and obtain consent before a cookie is set for the first time. Provided you get consent at that point you do not need to do so again for the same person each time you use the same cookie (for the same purpose) in the future. {4}
2. The use of cookies for “functional purposes” including remembering previous settings or collecting information about how users’ access and use a site (e.g. analytics cookies) at the present time still require consent. Possible ways of obtaining consent for these functional cookies is discussed in the next section of this document.
3. Organizations using cookies provided by a third party in their sites (placing pixels from media publishers such as Specific Media on their sites) may need to make users aware of this and point them to information on how the third party might use cookies and similar technologies so that the user is able to make an informed choice. This is by far the most challenging area in which to achieve compliance and have yet to be defined. {5}

The legislation in its current form and wording highlights challenges that will require further interpretation and clarification including:

• The definition of “user consent” based on their access device. In many cases a device is utilized by multiple users (e.g. home computer accessed by a family of three). Cookies are set at the device level as opposed to the individual user level. For example, if three people access an organizations’ site via the same device and browser it would require each user to be indentified individually.

• The legislation does not explicitly define the consent requirement relative to an “opt-in” or “opt-out”. The current interpretation leads to the “opt-in” option which is an important consideration.

• The definition of cookies serving “functional purposes” requires further validation in the context of Digital Marketing. The use of cookies for analytics and site optimization do serve a functional purpose in the fact they improve the user experience. However, their classification is yet to be completely defined.

• Larger third party organizations such as Google and Microsoft could gain higher acceptance due to their brand recognition as opposed to their specific privacy policies. If Google’s cookies used for their search engine, checkout service, analytics and advertising network gain wide acceptance over other providers, it could establish a monopoly of cookie data in the future.

• The legislation will accelerate the adoption of the ‘Advertising Option Icon’ (being developed by the Internet Advertising Bureau Europe, the World Federation of Advertisers, the European Advertising Standards Alliance, the Direct Marketing Association and the Incorporated Society for British Advertisers) which will be shown on or near behavioural advertising and will provide consumers with further information about behavioural advertising in the relevant European language when clicked.

The users will also be able to manage information preferences or stop receiving behavioral advertising via a new pan-European website.

[Click here for Part 3 of 'The future use of cookies in the UK: From informing to obtaining consent']

______________________________

• Background
• What is the role of cookies?
• What is the new legislation on cookies in the UK?
• When is the legislation on cookies coming into effect?
• What should you do?
• Our view on the UK cookie legislation
• The different forms for obtaining consent
• Recommended next steps
• How we can help
• References & Sources

______________________________

About Alain Portmann

My job is simple. To inspire staff, clients and consumers.

Mail | Web | More Posts (65)

The different forms for obtaining consent

Various methods for obtaining consent are currently being discussed and evaluated by the ICO and industry bodies. Below is a description of the most significant.

The use of browser settings. Unfortunately at the present time most browsers do not include advanced privacy setting which would allow for the user to consent individual cookies. While this is an option that could advance quickly it’s not one organization should rely on. One of the most significant challenges with this approach is that not all website visitors will have the most up-to-date browser with enhanced privacy settings. {6}

The use of pop ups and splash pages. While the use of pop-ups has been suggested as a way of gaining consent, it is an option that could seriously compromise the users’ experience on a site.

The use of terms and conditions. While the use of specific terms and conditions can be used to gain consent, it is important to note that changing the terms of use alone to include consent for cookies would not be good enough even if the user had previously consented to the overarching terms. To satisfy the new rules on cookies, you have to make users aware of the changes and specifically that the changes refer to your use of cookies. {7} This is most commonly obtained by asking the user to tick a box to indicate that they consent to the new terms when they register to their existing account or open a new account.

The use of footer or header copy. When related to the use of functional cookies, one possible solution being discussed includes the placing of text in the footer or header of the web page which is highlighted or which turns into a scrolling piece of text when you want to set a cookie on the user’s device. This could prompt the user to read further information (perhaps served via the privacy pages of the site) and make any appropriate choices that are available to them. This applies to cookies deployed when a user makes a choice about how the site works for them. In other instances, some objects are stored when a user chooses to use a particular feature of the site such as watching a video clip. In these cases, consent could be gained as part of the process by which the user confirms what they want to do or how they want the site to work. {8}

The functionality used to gain user consent is not yet defined nor have best practices been defined. It is important all organization research and keep abreast of developments in this area. On their site http://www.ico.gov.uk the ICO has implemented a very rough form of consent from users accessing the site.

[Click here for Part 4 of 'The future use of cookies in the UK: From informing to obtaining consent']

______________________________

• Background
• What is the role of cookies?
• What is the new legislation on cookies in the UK?
• When is the legislation on cookies coming into effect?
• What should you do?
• Our view on the UK cookie legislation
• The different forms for obtaining consent
• Recommended next steps
• How we can help
• References & Sources

______________________________

About Alain Portmann

My job is simple. To inspire staff, clients and consumers.

Mail | Web | More Posts (65)

Recommended next steps

Given the lack of specific guidelines on the implementation of the new legislation we recommend undertaking the following steps:

1. Undertake an initial audit of your site and the way cookies are being used. The audit should classify the individual cookies based on the role they currently play – from “strictly necessary” to “not necessary”. The audit should also aim to define how third party cookies and partners are approaching the new regulations.

2. Review your site privacy and terms of use to ensure details are provided about how cookies are being used. Your site cookie policy should accurately describe how you use your customers’ data as well as the data practices on your site, including your use of analytics software and your advertising practices. Furthermore, the policy should include links to cookie advisory sites such as www.allaboutcookies.org.

The privacy and cookie usage policy adopted by the BBC is one of the most comprehensive – including a detailed description of all cookies used and for what purpose. http://www.bbc.co.uk/privacy/bbc-cookies-policy.shtml. British Airways has actually included a stand-alone “View our cookie policy” link on their global landing page instead of simply adding the cookie usage information within their legal information section.

That’s what happened with the Information Commissioner’s Office (ICO) implemented the new law with existing technology, over 90% of site visitors declined to accept a Google Analytics cookie, thereby disappearing from their analytics.

3. Involve your legal council to ensure there is a common view within your organization.

Every organization needs to ensure they are doing everything they can to get the right information to users and that they are allowing users to make informed choices about what is stored on their device. Not taking action is not a choice. Every organization affected by this legislation must demonstrate that it is taking specific steps to comply with the new requirements on user consent.

If a complaint about a website is received by the ICO they are likely to take a negative view of an organization that cannot show that it is taking steps to change current practice to bring about compliance with the new laws.

How we can help

In response to the legislation and the needs of our clients, we have developed a Cookie Audit solution which provides organizations with an audit of the cookies on their site. The audit will provide the following deliverables:

1. List and categorize all cookies being served across a specific site including a) first party cookies and b) third party cookies.

a) First Party cookies
b) Third party cookies:

i) Analytics (e.g. Google Analytics)
ii) Ad-Server (e.g. Doubleclick, CheetahMail, Commission Junction)
iii) Advertising (e.g. Specific Media)
iv) eCommerce
v) Surveys
vi) Blogs
vii) Widgets
viii) Server Functions

2. Identify the specific data points captured by first party and third party providers including ad networks and media publishers.

3. Define the impact on the current Digital Marketing practices resulting from a possible low level of user consent on specific cookies

4. Present findings and recommendations on an MS Word brief.

The cookie audit has a one-off cost of £4,950 and takes 10 working days from the day of sign-off. If you are interested in discussing the contents of this document further, please contact Alain Portmann, Head of Strategy at 0207 253 4133 or by email: alain [at] webliquidgroup [dot] com.

[Click here for Part 5 of 'The future use of cookies in the UK: From informing to obtaining consent']

______________________________

• Background
• What is the role of cookies?
• What is the new legislation on cookies in the UK?
• When is the legislation on cookies coming into effect?
• What should you do?
• Our view on the UK cookie legislation
• The different forms for obtaining consent
• Recommended next steps
• How we can help
• References & Sources

______________________________

About Alain Portmann

My job is simple. To inspire staff, clients and consumers.

Mail | Web | More Posts (65)

References & Sources

(1) Source: ICO “Advice on the New Cookies Regulation, 09/05/2011, Version 1
(2) Source: ICO “Advice on the New Cookies Regulation, 09/05/2011, Version 1
(3) Source: ICO “Enforcing the revised Privacy and Electronic Communications Regulations (PECR), 25/05/2011, Version 1
(4) Source: ICO “Advice on the New Cookies Regulation, 09/05/2011, Version 1
(5) Source: ICO “Advice on the New Cookies Regulation, 09/05/2011, Version 1
(6) Source: ICO “Advice on the New Cookies Regulation, 09/05/2011, Version 1
(7) Source: ICO “Advice on the New Cookies Regulation, 09/05/2011, Version 1
(8) Source: ICO “Advice on the New Cookies Regulation, 09/05/2011, Version 1

______________________________

• Background
• What is the role of cookies?
• What is the new legislation on cookies in the UK?
• When is the legislation on cookies coming into effect?
• What should you do?
• Our view on the UK cookie legislation
• The different forms for obtaining consent
• Recommended next steps
• How we can help
• References & Sources

______________________________

About Alain Portmann

My job is simple. To inspire staff, clients and consumers.

Mail | Web | More Posts (65)

Last time, we began our look at behavioral search query insights around gift giving from a search marketing perspective by looking at trends in the use of “gift” as a search keyword in Google – and learned a few things in the process:

• When holiday shopping activity begins to heat up (October, going into November), we saw the Photo/Video, Food/Drink, and Beauty/Personal Care categories rising the fastest.
• Among gift purchases with a longer sales cycle, the Recreation category stood out, showing the fastest growth in gift queries from September to October.
• During graduation season in the spring, gift-givers frequently search for Recreation and Photo/Video gifts, but they also use keywords like ‘graduate’ or ‘student’ to qualify their research.

As we look beyond the spike in shopping around the holiday season, a number of interesting gifting occasions show their own unique patterns.

Reminder: the key metric in this analysis is called Average Search Volume Index, or ASVI. For more insight into our research methodology, see here.

With a New School Year Approaching, People Can’t Resist Giving More “Studious” Gifts

While we may not traditionally think of the back to school period as a big gifting occasion, whatever gift giving is going on indeed skews toward items of some educational relevance (the Arts & Humanities category). This is where book retailers really shine; are we perhaps seeing adult gift-givers trying to get school-aged gift recipients a little more excited about returning to the classroom?

The back to school season also doesn’t seem to diminish searchers’ interest in recreation: outdoor activities like boating, cycling, running, hiking, and camping, as well as crafts and hobbies. Note that this category doesn’t include sports like baseball, basketball, football, or soccer, which get their own separate category in Google Insights.

Several Key Categories Characterize a Slow Gifting Season in July & August

July and August can be relatively quiet from a gifting perspective. Summer birthdays notwithstanding, this slow period sits right between an active spring (Mother’s Day, Father’s Day, graduations) and the busy fall, which is characterized by an elongated shopping run-up to the holiday season.

What are the biggest drivers of this decline? For starters, it looks like the Recreation category, highly active during the spring and fall, enjoys a bit of quiet in July and August. We know that outdoor activities are popular in the summer, but according to the data, the relative demand for those gifts in fact bottoms out in the summer.

A few other categories are not far behind:

• Food & Drink: while overall search interest in beer and bbq are high in the summer, for a great many others the data is scattered elsewhere around the calendar. The trend just may be that the more “giftable” food & beverage items are simply not associated with warm weather.
• Beauty & Personal Care: Fitness queries, and to a lesser degree weight loss, tend to reflect New Years’ resolutions. People love to give spa gift certificates, but they generally do so around the holiday season or Mother’s Day. It could be that the prevailing wisdom is, summer is when we need to look our best… so winter and spring are the time to be thinking about those gifts.

Gadgets & Computers are Popular Last-Minute Gift Purchases During the Holidays

Computers & Electronics is a little misleading here. Because we’re using a tricky metric based on averages of a normalized index, we’re exposed to a big weakness in the data.

If there happens to be an abnormally high figure for a certain data period, it pushes down the index value for all the other values, which dilutes the average. And that’s exactly what happens with this category, the week before Christmas. This could be an indication that digital cameras, computers & laptops, home video & audio, and personal electronics like iPods (but not iPhones) are very popular last-minute purchases for the stressed-out gift giver.

The same holds true for categories like Games (especially video games). As for the others, that’s anyone’s guess, including the outliers at the top of the chart.

To Be Continued…

Our analysis continues next time when we’ll drill down into the Shopping category, where a lot of important products and services reside: apparel/clothing, flowers, luxury goods, and collectibles. We’ll also get to see more about gifters’ demand for information like product reviews, coupons, price comparisons, and general customer service.

After that, we’ll wrap everything up and provide some best practices for how you can put this data to good use – as well as dig up an insight or two of your own.

Footnote. The following categories of industry classification are used in this data set: Arts & Humanities, Automotive, Beauty & Personal Care, Business, Computers & Electronics, Entertainment, Finance & Insurance, Food & Drink, Games, Health, Home & Garden, Industries, Internet, Lifestyles, Local, News & Current Events, Photo & Video, Real Estate, Recreation, Reference, Science, Shopping, Social Networks & Online Communities, Society, Sports, Telecommunications, Travel.

About Paul Burani

Paul Burani - Partner, Web Liquid Group. Connect with me on Google+

Mail | Twitter | LinkedIn | Google+ | More Posts (55)

Even more than in the prior Excel posts, the accompanying excel file will be critical to clarifying many of the following tips. Link to download:

Excel Tips for Marketers – Part 3 Excel File

So, let’s get started!

GENERAL TIPS

Using Excel to Build URLs for Media Campaigns

We frequently find our clients prefer Google Analytics, and if you are only running a Google Adwords campaign on the paid media side, you merely have to hook the two platforms together with a single check box. But what if you have more than just Google Search or Google Display Network in your media mix? Google offers all Analytics users the option of manually tagging incoming paid media using a certain set of URL parameters (more information here). Using Excel, one can quickly build all the URLs one needs for the rest of their media mix (Facebook advertising, Ad Networks, Microsoft Bing, etc).

The Excel demonstration file attached to this blog post outlines the actual steps required to take landing pages and media campaign information and turn it into click through URLs for your ad server of choice (whether it be Microsoft AdCenter, Mediaplex or Facebook’s adserving platform). The functions are relatively simple for this process; it’s mainly a matter of splitting out what changes and what stays the same for each click through URL, followed by a large number of concatenations (my preference is the ampersand connector over CONCATENATE(), as outlined in Excel Tips for Marketers – Part 2)

The end result of using Excel and URL parameters is the ability to see all your paid media broken out in Google Analytics right in line with the more commonly reported traffic sources:

Using Excel for SEO: Parsing URLs & Backlinks

Just the other day we provided a backlink profile for a new client of ours (a fairly standard SEO-related piece of work that we do standard). The results came back in URL form, namely a domain (such as “www.makerfaire.com”), followed by a slug (such as “/newyork/2010”). By separating the domain from the slug, one is able to begin several interesting different types of analysis. When one has just the domains, one can aggregate them to see which domains are particularly prone to linking back to the website on which the backlink analysis was done. With the slugs, one can aggregate them to see which pieces of content are most likely to be linked to. Next, a textual analysis of that most-linked-to-content can lead one towards an SEO content development strategy that better matches the audience of the website on which the backlink analysis was done.

The formulas in the accompanying spreadsheet may look rather complicated, but they become that way due largely to the myriad of minor ways a URL can be written that would break many an unsuspecting MID / FIND formula combination. Here is a more basic example of a MID / FIND formula combination that extracts the domain (~95% of the time without error*):

=MID([Cell to look in], 8, FIND(“/”, [Cell to look in], 8 ) – 8 )

*The domain has to have a URL that begins with “http://”, and follows with either a “/” or a slug.

Cross-worksheet Validation in Excel

If you use a reference to identify the list of items a cell needs to validate against, that reference has to be on the same worksheet as the cell for which you are trying to setup validation. This first time I encountered this I wanted to pull my hair out. Luckily, it turns out there is a better solution: Named Ranges.

Being able to store the list of valid values a cell can have on a worksheet separate from that cell is hugely helpful when setting up Excel workbooks where the input tables and the output tables are on separate worksheets. By using this technique, one can lock users from modifying output/presentation/results worksheets while still allowing them to add and remove categories/values for VLOOKUP() and SEARCH().

The example outlined in the demonstration Excel file attached to this is straightforward, as most of the work is in menus and not formulas. The one formula involved is this:

=OFFSET([Cell reference to the first item in the list of valid cell values], 0, 0, COUNTA([Cell range reference to all the items in the list of valid cell values), 1)

FUNCTIONS

Histograms with FREQUENCY()

Histograms are wonderful ways of simply showing the distribution of a set of data, whether it’s days during the year above a certain volume of revenue, or in the example spreadsheet, the mozRank of the websites that mentioned the client’s brand.

There are two critical aspects of FREQUENCY() not directly mentioned in the demonstration Excel spreadsheet. First, one has to select a range of cells that is the same size as the cell range that defines the buckets, before one begins typing the formula. Second, one has to press CMD+SHFT+ENTER when they finish typing the formula rather than just ENTER. This is because FREQUENCY() is known as an array formula (I discuss array formulas in the first installment of Microsoft Excel Tips for Marketers, and this is where I learned how to use them).

MORE ADVANCED TIPS

Microsoft Excel + Wordle

Wordle is an amazing, free tool for generating word clouds. Word clouds are a neat way of visualizing the frequency of individual words in a larger body of text. We use them as a method for analyzing word of mouth data, and discovering what the critical emergent topics are in the conversation that may affect our clients. Here is an example word cloud created with Wordle using the text of this blog post:

Simple word clouds can be created by selecting "Create" on Wordle’s home page, but we are concerned with the "Advanced" section. In the advanced section one is able to bring Microsoft Excel to bear to help create, in this example, word clouds that color topics by their sentiment.

The starting data is fairly standard Social Media monitoring (our take) data, from an aspect of our current monitoring software that allows us to identify trending topics. This monitoring tool also analyzes the sentiment of each instance of the discussion on each topic, so we are able to get percentages of positive, neutral and negative commentary.

The key step in this process is not actually turning the sentiment numbers into Red/Green/Blue color values, but the step in between where we index the positive, neutral and negative sentiment percentages in order to produce actually meaningful color values (as opposed to ending up with all brown/grey words in the final product). The basic formula is (example here is for positive sentiment):

=(([Topic’s % positive] – [smallest % positive out of all topics]) / ([largest % positive out of all topics] – [smallest % positive out of all topics])

This indexing method maps the smallest positive % in the range to 0, and the largest to 1, and maps all the remaining topics’ positive %s in between 0 and 1, and levels the playing field such that, for example, a 1% to 4% jump in negative mentions is noticeable compared to a 27% to 40% positive mentions.

Many of the above tips involve more than just Excel formulas, and as such I can’t recommend enough downloading the excel file accompanying this post and trying some things out yourself. On the flip side, I know not all of us are interested in rushing home and spending our free time in Excel, so, if you post your questions, comments, grievances in the comments we’ll see if we can figure them out!

Prior Excel tips posts: Part 1, and Part 2.

And that’s all for now, but there will be more!

About Nate Walton

Mail | More Posts (5)

In the early days, the prevailing wisdom was that once search had established itself as a viable marketing channel, higher budgets would follow and CPC’s (cost per click) would understandably rise.  What has followed is a more quantifiable account of the same view, with a number of voices in the search world proclaiming a future of prolonged rises in costs per click.  In late 2010, Efficient Frontier published a report charting significant growth in CPC’s across a number of verticals:

• Travel: +21%, from January 2010 to January 2011
• Automotive: +8%
• Retail: +7%
• Finance: +7%

By the end of the year, SEO.com came through with a not-so-bold prediction: “Only three things are certain in a search marketer’s life: death, taxes, and rising CPC’s.”  The panacea, this story goes, is to invest in process automation from the likes of paid search management platforms.  But in truth, there are a number of lower-tech solutions to stave off hyper-inflation within the paid search economy:

1)    Pay more attention to Average Position. This metric often takes a backseat to click-through rate (CTR), CPC and conversion rate.  By capping average position, however, search marketers retain control over ballooning costs.  Google AdWords even offers advertisers a position preference functionality which can be helpful in structuring test scenarios.

2)    Blow the lid off your keyword lists. The 80/20 rule is alive and well in the paid search world.  While the top tier of keywords will always provide the bulk of your impressions and clicks, they also attract the most attention from advertisers and thus are most prone to rising CPC’s.  By doing your research and extending way into the long tail, you’ll not only extend your reach, but you’ll find that better conversion rates await.

3)    Experiment with search networks. Google segments clicks in paid search between searches conducted on Google.com and its search network partners, licensees of Google’s search technology who provide custom search engines within their own sites.  Often times the difference in CPC between the two is substantial (in one direction or the other).

4)    Daypart bidding. As PPC tactics are concerned, dayparting is an oldie but a goodie – but too often we think of it in a black & white context: either your ads are running or they’re not.  However, you can also adjust bids based on dayparts.  If your analytics data shows that your conversion rates are lower in the mornings, shave a portion off your morning bids – this way you’ll manage your costs without sacrificing the reach of your advertising.

5)    Use ad extensions whenever possible. Sitelinks, locations, product ads, seller ratings… these all have one thing in common.  They are non-standard ad formats which Google has tested and re-tested, with the intention of assuring high click-through rates and efficiency for advertisers.  Given the nature of PPC auctions, these can be effective in managing CPC as well.

6)    Invest more in Mobile Search. Searches conducted on mobile devices come at a heavily discounted cost, less than half of what we pay in desktop & laptop search (on average).  Meanwhile, the surging sales of smartphones have grown the mobile share of total search impressions to over 10%.

7)    Scale up to a more advanced solution. Managing multiple accounts can be an immense amount of work, but a good bid management tool goes a long way in simplifying the process.  That will make your campaigns work more efficiently on the media side, but what about the performance of your website?  Even a small incremental gain in landing page performance can yield huge returns – and a systematic approach to conversion rate optimization can pay for itself.

When all else fails, you can also accept your fate and just steal some budget away from offline channels – there’s plenty to go around!

About Paul Burani

Paul Burani - Partner, Web Liquid Group. Connect with me on Google+

Mail | Twitter | LinkedIn | Google+ | More Posts (55)

With the 2010 holiday season now over, the search marketing industry engages in a common ritual – measuring from all angles the most recent bump in spending, and its estimated effect on the overall e-commerce trends of the gift-giving blitz.  But what about the e-commerce sales that never happened?  Is anyone looking into these?

Virtually every consumer is handy with a search engine.  Roughly two-thirds of them are using search engines to qualify purchase behavior, and yet online sales represent at most 7% of total US retail sales.  So we’re not exactly going out on a limb here, saying that a lot of commercially-motivated search behavior is ending up in a black hole… instead of the shopping cart.

One explanation for this is the ROPO effect (research online, purchase offline).  In 2011, the overwhelming majority of offline purchases still include some online research component.  But ROPO is not a trend which will die off as digital consumer culture continues to evolve; Google estimates that offline sales influenced by online research will grow 12% this year, while online sales begin to flatten out.

The trick, for search marketers, is to acknowledge the offline world’s contributions to marketing innovation in the online world.  This starts with proper measurement of the ROPO effect to understand its contribution to revenue and profit.

Measure.

Passing from online to offline, the data attribution challenge becomes significant.  In one study, Google measured ROPO by monitoring different configurations of online & offline advertising in three German markets, and comparing resulting revenue levels.  This method alone is prone to various biases, but could be improved upon with an additional qualitative component.  A simple point-of-sale survey, asking the question “have you ever researched this product online?” would provide insight into the share of ROPO customers (and how much they spend).  Pivoting this against credit card data from the e-commerce side of the business de-duplicates multichannel consumers, and voila, you’ve sized up the ROPO opportunity.

Search.

If the ROPO share of your brand’s revenues is 25%, then it stands to reason that a sum equal to at least one-fourth of the search marketing budget should be earmarked for ROPO applications.

But what are these, exactly?  Any marketing tactic which bridges the gap between the consumer’s online and offline worlds constitutes a ROPO opportunity.  There are a number of tactics which come to mind, but for now we’ll focus on the low-hanging fruit: Mobile Search.

Tell me if this sounds familiar.  You’re walking around a store or shopping mall, and all you see is the tops of people’s heads as they’re buried in their phones.  Many of these people are researching their next purchase.  What’s the name of that thing again?  Can I get a better price by comparison shopping?  Does it come in fuchsia?

Today, the Mobile Search market (worth $250m in the US in 2010) represents only 2% of the entire search marketing industry.  That share will more than triple in the next four years – and it remains to be seen whether the researchers adequately take into consideration the radical innovations occurring in the mobile space.  It could be 10% or more, for all we know.  The first mover’s advantages are crystal clear.

How to target the mobile ROPO audience with Search:

1. Get into the habit of segmenting all your search campaigns, into Mobile vs. Non-Mobile.  It’s not just for budget control; you’ll see differences in searchers’ affinity for keywords and ad creative as well.
2. Aim for granularity with your keywords, right down to the product’s model name and number.  These will present uncompetitive, low-CPC (cost-per-click) opportunities for visibility.  (Tip: Remember how small those mobile keyboards are; you’ll want to add typos to your keyword list as well.  I’m a fan of Aaron Wall’s Typo Generator.)
3. Reach out with mobile coupons.  Roughly 8 of 10 smartphone users are receptive to receiving coupons on their phones (compared to less than half on their computers).  These can be used to push foot traffic to offline destinations using services like ShopText, Opt It, or Google Local Business Center.
4. Build up your efforts on the organic side as well.  Keywords like ‘reviews’ and ‘compare,’ when paired with your product category, will turn up popular sites for consumer feedback.  This is an opportunity to build a trusting rapport with the shopper when they’re close to the checkout.

Destroy.

How does the ROPO effect translate into success for a brand?  Let’s count the ways.

1. Showing an ad, absent competitors and so close to the point of purchase – that alone is enough to make marketers salivate.
2. Given the relatively low CPCs, generating incremental revenue from Mobile Search improves ROI overall, which makes you a lot of friends.
3. Running ads on Mobile Search, over time you’ll amass a vast repository of search query data.  This is in effect a snapshot of consumer behavior focused specifically on their needs when they’re somewhere between their home and the cash register.  In a channel where virtually every unknown ties back to consumer behavior, this could be the biggest catch of all.

Next time around, we’ll talk about a more qualitative approach to ROPO – at the intersection of Search & Social Media.

Update: Food for thought — research from Performics indicates that roughly 1 in 6 of all Google search clicks will be from Mobile Search by the end of 2011.

About Paul Burani

Paul Burani - Partner, Web Liquid Group. Connect with me on Google+

Mail | Twitter | LinkedIn | Google+ | More Posts (55)

Digital marketing is blessed by a high degree of accountability; unfortunately accountability comes at the price. Somebody once told me managing online data is like trying to drink water from a fire hydrant. In addition to the sheer volume of data, marketers are faced with identifying relevant metrics – metrics that help manage the short term and long term performance of a brand and marketing investments.

Since 2003 we have been faced with this challenge – clients, publishers and partners requesting a methodology that extracts the most value from their data. Before we molded our silver bullet, we reviewed several business principles to help guide our approach.

Kaplan & Norton’s (Harvard Business Review, February 1992) use of “the balanced scorecard” principle shaped our approach. What they essentially said through the balanced scorecard approach was that when managing the performance of a company and its assets (in this case data), there were many interests and data points to take in account. To be of any value measures have to be inter-related as opposed to exist in isolation. Our interpretation of “the balanced scorecard” principle led to a simple tailored belief.    

In order to combine a long-term roadmap with short-term actionable insight, data has to be interlocked to develop a “balanced scorecard” of marketing and brand data - compared to history, the industry or a competitive set.

 We call it a Brand Dashboard.

Our Brand Dashboard combines proprietary, third-party and public data to provide a full view of your brand’s health. With our Brand Dashboard you can not only capture more insights from the existing sales funnel, but also introduce new sources of data to enrich your consumers’ profile in the digital space – and ultimately widen the funnel to capture more revenue.

The Brand Dashboard includes a series of brand metrics with weighted influence, grouped into four manageable brand facets:

1. Awareness
2. Engagement
3. Loyalty
4. Efficiency 

These four brand metrics are benchmarked against three individually-weighted dimensions: 1) performance relative to competitive set, 2) directional performance trend across date range, and 3) rolling average of competitive or directional performance.

The ATLAS Institute confirms the value of our brand dashboard. Most brands only measure 6% of the consumer purchase funnel, focusing almost entirely on qualitative metrics. While these qualitative metrics are incredibly valuable, they need to be complemented with engagement metrics.

Alain Portmann, author of  ”Building a balanced scorecard of marketing and brand performance” is Web Liquid’s Founding Partner, Head of Strategy. Visit www.webliquidgroup.com for more thought pieces.

About Alain Portmann

My job is simple. To inspire staff, clients and consumers.

Mail | Web | More Posts (65)

While most digital marketers embrace the wealth of data generated by their campaigns, one of their biggest challenges is to leverage that data beyond the specific channel in which it is generated.  It can be a great frustration to make sense of numbers at the highest, most general level – but within that problem lies a significant opportunity.

Let’s use discount wine retail as an example.  A store owner may be proud of their e-commerce site, but if they’re not connecting with the clientele on a more personal level, they’re leaving a lot of potential value untapped.  This is frequently a motivator for having a presence in social media.  Adding a blog as a complement to the e-commerce experience would be a great way to build a presence in these word-of-mouth (WOM) channels.  We know that WOM is highly trusted by consumers (McKinsey estimates it is the primary factor in up to 50% of purchase decisions), but how do we identify the relevant conversations going on?

Tapping into related search query activity is one easy way to do this.  Google’s AdWords Keyword Tool is a free resource offering a robust view of how frequently people search for different keywords.  Let’s take it for a spin.

[Note: Earlier in May, Google updated its AdWords Keyword Tool interface.  You’ll find it is easier to use the old version; just click Previous Interface.]

1)    Go to the Keyword Tool and type “discount wine”.  (Make sure “Use synonyms” is checked, and both options beneath “Filter my results” are unchecked.)

2)    You’ll get two lists (Related terms, Additional terms).  Click “Add all” for each, so that both show in the green section on the right.

3)    Output that list to a text file; this is now your Seed List.  You’ll then take this and drop it back in where you typed “discount wine” beforehand.  (This time, make sure “Use synonyms” is unchecked, and “Don’t show ideas for new keywords” is checked.)

4)    Export this new list to CSV, and open in Excel.

5)    Extract “discount” and “wine” from the whole list (first, go for related terms like “discounts,” “discounted,” “wines”).  This leaves you with the most common search modifiers related to discount wine.

6)    Re-sort the list by keyword to remove blanks, and sort again, descending by Global Monthly Search Volume.  At this point, it’s useful to cap the list to only those terms with a search volume X or higher (for “discount wine” I used 1,000 as my cutoff).

7)    Next to the search volume column, divide the number by 1,000 for each term.  You’ll understand why in a second.  Let’s call this Adjusted Search Volume.

8)    Create a column next to the keywords where you fuse each term with a space (use the CONCATENATE function with the keyword followed by a space in between quotes.  Example: =concatenate(A1,” “)

9)    Now create one last column for your final output, where you multiply the keyword by the number shown in Adjusted Search Volume.  This will give each keyword the proper relativity in the finished product.  You can use the REPT function in Excel to do this.  It will look something like this:

10)    Take that output and drop it into your favorite tag cloud generator (at Web Liquid, we like Wordle).  And voila’, here’s a visualization of the top search modifiers related to discount wine (click to enlarge):

Is the word “red” cluttering things up too much?  You can go back, and paste everything but the top keyword.  Watch how the cloud changes:

Still not inspired?  You could just limit to the terms that have monthly search query volumes between 1,000 and 10,000:

This process offers an objective, relative view of the conversation surrounding a root keyword, as reflected in search activity.  It’s a very scalable approach as well, with room for conversion, geography, seasonality and other filters.

When adding search-enabled WOM insights to the other creative priorities surrounding new content development, the net result is an airtight method for populating an editorial calendar and positioning its content for maximum visibility in search engines.

About Paul Burani

Paul Burani - Partner, Web Liquid Group. Connect with me on Google+

Mail | Twitter | LinkedIn | Google+ | More Posts (55)

A special report in the Economist, referenced in the post “Data Exhaust, from the Woods to the Highway”, had this to say about the overwhelming amount of data being collected:

Sophisticated quantitative analysis is being applied to many aspects of life, not just missile trajectories or financial hedging strategies, as in the past. For example, Farecast, a part of Microsoft’s search engine Bing, can advise customers whether to buy an airline ticket now or wait for the price to come down by examining 225 billion flight and price records. The same idea is being extended to hotel rooms, cars and similar items. Personal-finance websites and banks are aggregating their customer data to show up macroeconomic trends, which may develop into ancillary businesses in their own right. Number-crunchers have even uncovered match-fixing in Japanese sumo wrestling. (from “Data, Data Everywhere“)

After the awesome size of “225 billion” wears off, I start to wonder: what if I had that dataset of 225 billion flight and price records, and decided the price of farm machinery and the population of zebras in central Africa were the deciding variables in price prediction? You would think it a terrible price predictor and probably tell me so.

But, what if I decided to include the time of day of the flight as well as the weather conditions, for which I used a proxy of temperature?  What if including both caused variance inflation factors to skyrocket, because lower temperatures are correlated with nighttime?  What if I didn’t have time to check, or didn’t think to check the variance inflation factors – and this model was published on the Internet for consumers to use?  What if customers, upon inputting their preferred time of departure, received improper predictions because the weight applied to that variable in this model was biased? What if nobody could tell anything was wrong?

The science of data mining — systems, networks, and complex multivariate statistics — is critical to bringing understanding to the new data age, but the bottom line is most people do not have the time to learn and/or apply the most complicated algorithms and statistical methods to their data.  In these day-to-day situations, the hidden problems of a strictly data-driven approach grow larger without ever becoming more apparent. The human mind is still the best analytic thinker around and an analytic framework, again, should always come before the data processing.

It can be very tempting to defer to algorithms or statistical techniques far enough beyond our comprehension that we trust they will work away behind the scenes, providing us with deep insight as we ask them basic questions (think Google). But, as the current literature on computer learning algorithms and statistics attempts to demonstrate, we still have a long way to go before we have computers with any human semblance of theoretically sound pattern-matching, and thus any statistical methods which can deduce real causation.

While computers don’t have an intuitive grasp of how theory should be applied to patterns, we humans do, and we have to be careful to apply our intuition to the problem before we apply our computers to the data.  The order is important due to the risk of overlooking the problem at hand, focusing instead on the output from the computer.

If we directly apply our analytical skills to the output from the computer, we are hunting for a solution without ever properly defining the problem.  The actual causal process becomes a needle in a haystack of other temptingly similar looking needles.  And at this point, I think all of us would just choose the needle we like the most.

Researchers have a term for this: it’s called confirmation bias.  And marketers, like researchers, earn their keep by making decisions that are as free of bias as possible.  While data mining has become an essential weapon in the marketing arsenal, consumers remain complex in their behavior, and human intuition is still required to foster their brand relationships.

About Nate Walton

Mail | More Posts (5)

As the pay-per-click (PPC) channel continues to mature for search engine marketers, it’s become more and more important to anticipate search behavior in the process of meeting existing demand.  The objective of this analysis is to assess the viability of using URLs as keywords, in order to scale PPC programs and drive larger volumes of conversions.

Hypothesis

Individuals vary widely in the way they interact with search engines.  Some users are prone to querying broad terms such as “digital camera” and then drill down based on the results they find. Others use more focused queries such as “buy used nikon d200 digital camera.”  With the evolution of local search we see more queries like “dry cleaning wicker park chicago,” and with blended search results becoming more popular, users will search terms such as “youtube video will ferrell landlord.”

An important subset of search engine queries request entire URLs (or parts thereof), instead of keywords in the traditional context.  Search engines generally read periods as spaces, so querying the keyword “www.google.com” ends up looking like “www google com” — which is a fundamentally different investment within a marketer’s PPC (pay per click) advertising program.

Hypothesis: The presence of extensions such as “www” and “com” presents a unique commercial and competitive profile for keywords, and marketers should look to these as opportunities to improve overall ROI within PPC (pay per click) advertising programs.

Definitions

Root Keywords – these represent the single term shown between the periods in a root URL.  Example: for www.google.com, the Root Keyword is “google”.

Extension Keywords – these represent the prefix/suffix used in the URL.  For the sake of this analysis, the Extension Keywords under consideration include the prefix “www” and the TLD’s (top level domains) “com”, “net” and “org”.

Cost-per-click (CPC) in this analysis is defined as Google’s value for Estimated Averaged Cost Per Click, which is defined as the maximum CPC bid necessary to show an advertisement in one of the first three positions within Google’s sponsored links.  It should be noted that this estimate is privy to the unique characteristics of the auction, of which billions take place every day.  Therefore, we default to Google’s best guess.

Competitiveness in this analysis is defined as the quantity of advertisers worldwide bidding on a particular keyword, relative to every keyword.

Methodology

We analyzed thirty-one domains whose Root Keywords combined for over 2.2 billion global average monthly queries in Google.  The domains were chosen from among the most popular sites across the U.S. audience (measured in unique visits), with several notable exclusions, including compound names (e.g. “huffingtonpost” vs. “huffington post”), domains with multiple semantic applications (e.g. “live.com” vs. the verb “live” or the adjective “live”).  Monthly query volumes are based on Google data at the broad-match level, indicating raw volumes with or without other terms as part of the search query.

Using Google’s estimates for Average CPC and Advertiser Competition, we aggregated data across the experimental set, to generate unique figures for Root Keywords and Extension Keywords.

Results

• The presence of any or all Extension Keywords increased the average monthly query volume of the experimental set from 2.21 billion to 2.47 billion global monthly queries — an increase of 11.7%.

• The aggregate CPC for the Root Keyword set was $0.72, while the Extension Keyword set showed an aggregate CPC of $0.47 — a discount of 35%.

• The Competitiveness for the Root Keyword set was 99.94%, while for the Extension Keyword set we observed 98.59%.

<click to enlarge>

Analysis

Bidding on keywords which include extensions presents a significant opportunity for marketers to extend reach in the PPC channel at a relative discount of approximately one-third, measured in Estimated Average CPC terms.  The competitive profile of these keywords is not substantially different, but this does not offset the notable cost savings of this ROI-driven keyword strategy.

Appendix

The following domains were considered in this analysis: adobe, amazon, aol, bing, cnet, cnn, dell, digg, ebay, facebook, flickr, hulu, imdb, latimes, microsoft, mlb, msn, myspace, netflix, nfl, ning, nytimes, pandora, paypal, photobucket, twitter, ups, usps, wsj, yelp, zynga.

Limitations

We acknowledge that the reporting and analytical tools used in this study have limitations which might impact our conclusions.  Keywords estimated by Google at $0.05 CPC generally have a high standard deviation relative to this number.  Relying on Broad match, while essential to accumulating a large enough sample to derive meaning in this analysis, also introduces the remote possibility of extraneous terms occasionally appearing within the query.

About Paul Burani

Paul Burani - Partner, Web Liquid Group. Connect with me on Google+

Mail | Twitter | LinkedIn | Google+ | More Posts (55)

Before we get excited and start pulling display budgets, let’s bear a few facts in mind:

1) On the User side: When a consumer sees a display ad, a click is only one of a multitude of actions they may take in response.  Telling a friend or colleague, typing the URL into a browser, or simply filing away that brand interaction for future reference — these are all actions of interest to marketers who are looking to drive true engagement with their brands… to say nothing of search engine activity, which is demonstrated to be a direct by-product of display advertising.

2) On the Marketer side: Clicks are just the tip of the measurement iceberg — beware the temptations therein!  The click may be widely-appreciated for its ease of measurability, but as marketers we need to resist giving in to “Clickthrough Determinism” — the tendency to boil down consumer behavior into the simple arithmetic of click patterns.  When a click on a display ad leads to a purchase, we need to be taking into account all the communications and actions that transpire upstream in the purchase funnel.  It’s here that we really impact the value of media, using more profound metrics such as conversions (post click and post impression), cost-per-conversion and ultimately, dollar-in/dollar-out ROI.

Microsoft’s Atlas Institute found that over the course of a ninety day sales cycle, only 6% of accumulated data is taken into consideration for marketing purposes:

Perhaps this is why, as the comScore/Starcom report indicated, this year’s study “focused more on alternative measurement, suggesting that a low number of clicks doesn’t necessarily mean banners don’t work, but that marketers are looking at the wrong success metrics.”  Acknowledging that the digital “channel” is actually comprised of a few dozen unique channels (search, display, affiliate, mobile, video, social, etc), we prefer to use a holistic approach to brand health assessment.

Acknowledging that marketers can dig far beyond click activity for insights, we also took note of a 2009 Eyeblaster study which indicated that 2 out of 3 senior marketing executives run cross-channel campaigns, but 88% do not integrate performance data across channels.  This has been the basis for developing our comprehensive reporting system for Digital Brand Management, with a concise and actionable dashboard.  By measuring brand health on a standardized relative performance scale across specific metrics — and comparing to history, the industry or competitive set — it is possible to answer the question asked by many a brand manager, “How is my brand doing in digital?”

We’re glad that Advertising Age has shed light on the dilemma of Clickthrough Determinism.  The big issue for 2010 is, what do marketers plan to do about it?  Reducing display budgets would minimize the impact of one of the most efficient, accountable and innovative channels in the digital landscape, but it’s hard to ignore the raw numbers.

When it’s time to report to the CMO on the health of the brand, a standardized approach removes all the bottlenecks from the assessment phase, leaving more resources available for action.  In any halfway competitive category, this kind of competitive advantage can translate into significant gains in market share — and that’s the ultimate, measurable return on investment.

For more about our brand dashboard, or our philosophy on measurement and analytics, drop us a line in London or New York.

About Paul Burani

Paul Burani - Partner, Web Liquid Group. Connect with me on Google+

Mail | Twitter | LinkedIn | Google+ | More Posts (55)

One of our favorite resources to measure brands’ awareness and engagement in digital media is Google Insights, which shows trends in Google search query activity across a wide range of parameters.  Acknowledging that key players in the Pharmacy Retail segment have a significant opportunity to connect with their customers during this time, we chose to take a deeper look at four key players in this space: CVS, Rite Aid, Walgreen’s, and Duane Reade.

Definitions

Before going any further, let’s state the obvious.  By no means do we believe that search engine behavior is 100% congruent to overall consumer behavior.  We are well aware that people interact with brands in a variety of ways, through a variety of channels.  However, in recognition of the trend toward rising adoption of internet-enabled mobile devices, and the near-ubiquity of search engine usage among the US online audience, search engine activity is nonetheless very insightful to marketers.

By measuring volumetric trends in search activity, we can isolate patterns of activity surrounding the aforementioned brand names.  Here are a few types of Google queries that would be reflected in such an analysis:

• Where is the nearest Rite Aid?
• How does Duane Reade’s Rewards program work?
• Are there any good coupons at CVS this week?
• Will Walgreen’s be offering flu shots this fall?

We have chosen to limit our analysis to the New York City metro region, to allow for the consideration of a regional brand (Duane Reade) without skewing data.

Research

This is how relative search volumes in Google have looked since 2004, and project through 2010:

(Click on the chart to enlarge.  This image is a static depiction of Google’s search query volume estimates in October 2009.  For an updated estimate, see Google Insights for Search.)

As recently as early 2008 (and defined strictly by branded search queries), CVS has a strong #1 position in New York, with the other three all clustered in a distant second place.  In other words, CVS had done a far better job of promoting awareness for itself.  But watch what happens around the second half of 2008… Walgreen’s has begun to separate itself from the pack.

Here’s how the big four break down in terms of their social media efforts on Facebook:

We also thought it would be interesting to look into similar efforts on Twitter:

Notice a pattern?  Not only is Walgreen’s leading the charge, but the proportions are fairly constant across both Facebook and Twitter (two of the primary channels using in social media marketing).  This is especially noteworthy in light of today’s news from comScore, M80 and GroupM illustrating a “correlation between the discovery of brands through social media and search behavior, including increased lower-funnel searches and paid search click-through rates (CTRs).”  Here are some of the high-level takeaways:

• Exposure to branded social media increases likelihood to query that brand in search engines by 2.8X
• After querying the brand name:
  • click-through rate on paid search ads increases from 4.5% to 11.8%
  • likelihood of clicking on the brand’s site in organic search results increases by 2.4X

Clearly, a good portion of Walgreen’s growth in branded search query volumes can be traced back to their social media success.  (For more information on the study, see comScore’s press release, “Social media exposure is correlated with search behavior and click-through-rate; Introduces the value of media discovery for advertisers.“)

At Web Liquid, we tend to look at the word-of-mouth/social media space through a paradigm of “service over solicitation.” (We wrote about it back in early 2007: “Service vs. Solicitation“).  Given Walgreens’ social media dominance compared to the competitive set, their strong performance in the online space is not surprising.  With an audience actively engaged through social media, they enjoy many spoils: crystal-clear feedback loops, rapid deployment of promotions, opportunities to demonstrate quick, effective customer service… we could go on and on.

To Walgreen’s we tip our hats.  After all, in the retail business service is a hefty concern… and their progress in the social media space is an indication that they have embraced Service over Solicitation.  As it turns out, however, the increase in search activity around their brand name will, in fact, also spawn demonstrable commercial gains.

About Paul Burani

Paul Burani - Partner, Web Liquid Group. Connect with me on Google+

Mail | Twitter | LinkedIn | Google+ | More Posts (55)

• Two-thirds of users do not click on any display ads
• 16% of Internet users account for 80% of all clicks
• The demographics of clickers are skewed towards users aged 25 to 44 earning less than $40,000 a year
• Even search ads, are only clicked in average on 4% of the time on Google and 2% on MSN

So are low click rates evidence that a display ad or search query has not had any impact on consumer behaviour? Display ads do impact and influence consumer behaviour, in fact there is latency and branding effects even when click rates are minimal. Furthermore, the combination of display and search delivers considerable synergy as the combined elements are greater than the sum of the individual components.

Let review the evidence put forward by Comscore:

• Display ads lift visitation to an advertiser’s site by at least 46% over a four week period
• Display drives a lift in brand searches of at least 38% over a four week period
• Display lifts the likelihood of purchase on the advertiser’s retail store by 17% over a four week period

The problem is that if clicks are not the appropriate metric, what’s the alternative? Clicks are easy to measure, so it’s a good default. In reality the question is not about effectiveness but attribution. The current “last click wins” reporting model is flawed as it ignores all prior interactions the consumer has with a marketer’s message. Have a read of  “Why the click is the wrong metric for online ads” for a view on the current “last click” attribution model.

Lets’ consider alternatives to measuring the effectiveness of advertising based on clicks. After all, if you know most users that are exposed to your media will not click though – clicks provide a very limited view on the impact of your advertising. On the other hand, post impression metrics can also be challenging due to conversion latency.   

Unfortunately it is not as simple as replacing clicks with another metric, it is a fundamental change in the way media is tracked and evaluated – based on the cumulative impact of multiple media on consumer behavior. Some recommendations to consider:

• Measure the value of clicks and media relative to site visitation; specifically the value of new visitors and returning visitors
• Build an attribution model that assigns value based on the ability of media to 1) trigger a transaction, 2) assist a transaction and 3) complete a transaction
• Begin to consider “time spent” as an engagement metric – the growth of video advertising will force brands and agencies to look at media consumption metrics 
• Evaluate a first party ad-serving solution such as TruEffect which provides cookie ownership and a much richer set of engagement and conversion data than third party ad-servers

Alain Portmann, author of  ”Clicks as measurement” is Web Liquid’s Founding Partner, Head of Strategy. Visit www.webliquidgroup.com for more thought pieces.

About Alain Portmann

My job is simple. To inspire staff, clients and consumers.

Mail | Web | More Posts (65)

Then why does Ad Position matter?

The simple answer is, as Varian acknowledges, is that when ads show higher up on the page, they generally garner more clicks, which theoretically leads to more conversions.

Digging beneath this surface-level insight, however, we observe several other meaningful implications.

1) It is well-known that the clickthrough rate (CTR) impacts Quality Score on Google AdWords.  The historical CTR of each keyword/ad combination, as well as CTR of all individual keywords, ads and display URLs, factor heavily into this score — which interacts profoundly with bid settings to determine how much the advertiser will spend.

In other words, maintaining a strong CTR program-wide is one way to mitigate costs and economize a pay-per-click (PPC) advertising program for the long term.

2) An ad showing high on the search results page will often be perceived more favorably than another lower-placed ad.  In 2008, iProspect conducted its Blended Search Results Study and found that 39% of search engine users perceive the companies showing ads among the top search results as the leaders in their field.  This figure has also shown a noticeable upward trend over the years.

So not only are we talking about budget management — we’re talking about brand perception in search engines.

What does Ad Position reflect in the AdWords auction?

Ad Position does not directly correlate with keyword bids or effective CPC (cost per click); this is a fundamental principle of Google’s quality-based auction system (see this video of Varian’s introduction to the AdWords auction).  However, by no means does this a direct inverse correlation — just because the spend level doesn’t directly affect ad position, we can’t say it has no impact at all.

How should Ad Position figure into PPC optimization strategy?

Where are we going with this?  Let’s start with some “ifs”:

IF we acknowledge that ads showing lower on the page will rack up lower costs,
IF we have some tolerance for lower CTR, and
IF we have other branding measures in place (ideally from display advertising or social media)

…then what we should be doing is placing all of our keywords into one of three categories, based on historical PPC data indicating keyword-level ROI:

1 = keywords with high ROI
2 = keywords with low ROI
3 = keywords with zero conversions

Group 1 keywords should receive generous bids (and of course be optimized for quality as well), in order to push high Ad Positions.  They will convert at a certain rate, but our goal is to push upward on budgets to attain maximum click volumes.  Nothing special here.

Group 2 keywords should be scrutinized the heaviest.  We look closely at the bid history and routinely push downward; the goal is to register clicks at a very low overall CPC.  Showing ads at the bottom of the page won’t generate many clicks, but thanks to Hal we can now be confident that these keywords’ conversion potential won’t be profoundly changed by low Ad Position.  This tactic allows us to actualize ROI even among the poor performers, maintaining a conversion stream which would otherwise be forfeited.  All keywords, even the weakest ones, have an effective CPC at which they are worth bidding on.

Group 3 keywords are a tricky lot.  Some of them won’t have enough accumulated clicks or spend for the marketer to make a confident decision.  The others, however, may have branding potential, or perhaps act as a gateway to other more profitable content.  These should be abandoned in PPC, and become the focus of a targeted SEO campaign.

Aim for the middle!

Segmenting a PPC program is nothing new, nor is throwing a lot of budget at the best performing keywords.  However, with this research, Varian has given us a fresh perspective on how to treat the average performers.

We can be sure that the aggregate quality scores of each competitor in a given category will improve over time; this is inevitable as agencies continue to evolve their search engine marketing (SEM) capabilities.  As a result, all other things equal, the bid price will slowly become a bigger predictor of Ad Position — and if Varian’s findings continue to hold true, marketers who master this type of segmentation will retain the maximum ROI from paid search.

About Paul Burani

Paul Burani - Partner, Web Liquid Group. Connect with me on Google+

Mail | Twitter | LinkedIn | Google+ | More Posts (55)

So why does traditional online advertising not work in social media? First, most consumers decisions on social media sites take place through the referral of friends. Case in point, 78% of respondents to the 2007 Nielsen Global Survey on Word of Mouth indicate consumer recommendations are the most credible form of advertising. Second, the push marketing approach does not apply to niche, fast-moving audiences found in social media sites. Three, the mindset of social media users is driven by self promotion and community (social engagement) in order to connect there has to be an exchange that is of value to the user. If you chose to advertise on social media sites be conscious of the need to give that audience something of value as opposed to a advertising message.

Twiter, Facebook, Flyertalk and thousands of community boards, blogs and social media platforms empower consumers to communicate, network, contribute and make decisions about their everyday life. The gold dust of social media is consumer generated content and the sentiment it reflects about brands dynamic word of mouth and conversation. The question is how to indentify, harness, measure and act upon the conversational gold dust. In simple terms the challenge is to define and tie key performance indicators to the content and sentiment being measured.

The first thing to understand is that social media and consumer generated content serves as a repository and public record of conversations and consumer sentiment about a brand. A number of companies (Nielsen Buzzmetrics, Brand Watch, Market Sentinel) have popped up offering technology to capture those conversations for analysis but much of the application of that technology is still quite tactical or focused on brand reputation management. In fact, the consumer generated insight research opportunity is to use that insight on a regular basis and with quality marketing-based analysis to derive valuable insights that can improve marketing communications.

In our experience, the single most important measurement metric for social media is a variation of Net Promoter Score ©. Creator, Fred Reichheld of Bain & Co., measured the number of consumer “Promoters” (those who would recommend a product or services) and “Detractors” (those who actively discourage the use of a product or service) of hundreds of companies across many categories finding that on average the NPS leader grows at 2.5x. While the Net Promoter Score © metric can provide a foundation metric, it is important to define additional metrics aimed at benchmarking the contribution of social media against other channels of consumer interaction.

Our work with Avis and OpenSkies on the social media space has allowed us to define a measurement approach to consumer generated content; based on five measurement modules.

1. Involvement
This set of metrics is aimed at understanding the propensity of involvement for either your own social media environment (community board, blog) or third party social media sites that drive users to your site. Assigning a unique referring ID to your social media property or third party social sites within your site analytics tool (Web Trends and Google Analytics) should facilitate this data.
- Site visits
- Time spent
- Pages viewed
- New vs. returning users
- Navigational path

2. Engagement
These metrics are aimed at understanding the specific actions and behaviors of users. Every engagement such as a comment should be equated to a call centre engagement or store visit. This approach allows for social media engagement to be viewed as part of the overall customer journey, providing a degree of benchmarking.
- Commentary volume
- Frequency of comments
- Sources of commentary
- Commentary creator segments
- Topics/Subjects

3. Sentiment
This set of metrics is aimed at identifying changes in positive and negative feedback across time and against competitors. This information can be used by Customer Service teams to proactively address specific issues being openly discussed by consumers.

- Net Promoter Index variation

4. Propagation
This set of metrics is aimed at indentifying the distribution of social media content and its impact on how other consumers find and participate on the discussions about your brand. A key metric is search engine indexing which is related to the incremental impact of natural search traffic and visibility derived from social media.
- Citations
- Trackbacks
- Search Engine indexing

5. Transaction
While not applicable to all brands, transactions from social media can be measured by leveraging third party ad-serving and site analytics data. Establishing a control test against your average transactional site data can be an effective way of identifying variations on behavior and value.
- Registrations
- Shopping cart instances
- Conversion rate
- Revenue per transaction
- Revenue by user

We have been working with one of our travel clients to define the contribution of social media to their bottom line. Our approach was to benchmark revenue by blog visitors versus non blog visitors. The results are not surprising and have strengthened our clients belief on the value of social media. The average transaction value of a blog customer is 1.16 times higher than average site user.

Alain Portmann, author of  ”Measuring the gold dust of media” is Web Liquid’s Founding Partner, Head of Strategy. Visit www.webliquidgroup.com for more thought pieces.

About Alain Portmann

My job is simple. To inspire staff, clients and consumers.

Mail | Web | More Posts (65)

In this post I reveal the 9 most significant factors to consider when launching or evaluating an existing Pay-per-Performance program that contributes incremental revenue and return on investment (ROI).

1. Remove conversion duplication.
Conversion duplication occurs when users are being reached by the same advertiser across multiple sites and tracked across different reporting systems. An Atlas Institute study entitled The Hidden Costs of Pay per Performance Media estimates the rate of duplication to be between 20% to 25%. The issue of conversion duplication is further magnified, given publishers typically sell their remnant inventory to multiple ad networks and by ad-networks buying, selling and trading ads from each other to maximize reach and reduce frequency. If you are working with multiple ad-networks such as Advertising.com, Blue Lithium or DrivePM, your conversion duplication increases significantly. Conversion duplication not only has a significant financial impact as advertisers over-pay for conversions but it skews media inventory optimization and investment allocation decisions.

While there is no single ad-server or site analytics solution that offers the best tracking and ad management solution across all channels, it is possible to piggy-back all third party tracking platforms into a single tracking tag. This effectively allows advertisers to utilize their preferred channel-specific platforms (e.g. BidBuddy for paid search, Trade Doubler for affiliate marketing) while having a single de-duplicated count of conversions.

2. Focus less on the buying of media and more on the optimization of media.
Early in my career I learned that the secret to sustainable and effective online media solutions what not in the buying, but in the optimization. What you buy the media inventory for is the starting point, what you do with that media inventory is what really counts.

On-going optimization is the single most important factor is the success of a Pay per Performance advertising program. In our experience, the buying model is secondary to the end return on investment (ROI) metric with CPM programs at times working more effectively than CPA based programs. One way to improve optimization is by increasing the granularity of your data send your cost per performance partner multiple tracking tags as opposed to a single tracking tag.

3. Assign credit where credit is due.
It is likely your Pay per Performance advertising program is based on payment to the publisher that drove the last click. While this is an accepted metric, it negates the value or contribution of everything that preceded that click. Imagine if the winning Track & Field Relay Team in the Beijing Summer Olympics was only given one gold medal the medal given to the last runner to carry the baton and cross the line. It would not be fair, given the other three runners contributed equally to the winning performance.

This is how conversions are attributed under the last click rule. For more on the last click rule read a previous post entitled “The last click gets to much credit”. If all credit is given to the last click action, it is only a matter of time until the media that drives that prospect down the conversion funnel is discarded (and that funnel becomes smaller and smaller). Furthermore, a study on online advertising has shown that click performance is the wrong measure of ad effectiveness in the long run. According to the study, 6% of the heavy clickers called “Natural Born Clickers account for 50% of all display ad clicks and cannot represent the total online population.

In a Pay per Performance advertising environment it is essential to understand the role of every publisher and partner in order to make informed decisions. Exposure to conversion mapping provides the information needed to potentially weigh conversion payments, depending on the role each publisher played in driving a conversion.

4. Get in the habit of re-engaging your visitors.
When you establish a new personal or business acquaintance it is customary to communicate after the first engagement. Re-targeting users that have visited, transacted, or researched your site is key to drive efficiencies from your pay per performance program. Measuring the return conversion rate of users is a useful metric to establish a pool of valued prospects. Furthermore, re-targeting provides you with the opportunity to tell your prospects and customers a story about your brand as opposed to simply broadcasting the latest offer in hope it meets their needs.

Our work in the travel sector indicates 56% of users identify “visiting a site on the past” as the most compelling reason to purchase from a specific travel provider. Furthermore, a re-targeting program for one of our travel clients has increased click rates by 73% and revenue by ad impression served by 96%.

5. Optimize your site landing pages.
Imagine receiving a beautifully crafted and targeted invitation in the post. Your expectation is that the event will mirror the craftsmanship and value of the organizers’ invitation. Upon arriving to the event your expectations are not met, you are not properly greeted, introduced or entertained. A great invite. Poor party. The same dynamic applies to your cost per performance advertising programs and the expectations of your potential customers. While you can spend substantial resources crafting a great creative and offer, you have to invest in developing landing pages that will facilitate and fulfil the mission of your potential customer. Consider the impact of increasing your site conversion rate by 5% versus increasing your marketing spend by 25%.

6. Understand your optimal conversion window.
It is essential to determine the conversion window that provides the optimal balance between volume and cost of conversion. While the temptation is to establish a uniform conversion window for all channels including paid search, email marketing, display advertising and affiliate marketing, the approach is short-sighted. The reality is that consumers interact differently with each channel; each channel plays a different role in driving the users through the conversion funnel. Determining the optimal post click and post impression window goes beyond deciding on a 7, 14, 21, 30 or 90 day parameter. A good understanding of your optimal conversion window can allow you to negotiate payment on post click conversions, as opposed to post view conversions.

7. Differentiate between conversions from existing and new customers.
In the majority of cases cost per performance advertising agreements are driven by the volume of conversion, as opposed to the quality of conversion. If your business tracks, communicates and measures the contribution of existing and new customers differently so should your Pay per Performance advertising program.

In my experience, all advertisers are greatly concerned with the incremental contribution of cost per performance advertising – am I paying for conversions that would have occurred regardless of advertising? Establishing a payment model that recognizes conversions from existing and new customers can help tie advertising investments to larger commercial imperatives such as market share, cost per customer and revenue per customer.

8. Focus on large ad-units above the fold.
Should you pay for conversions driven by a 120 x 60 display ad at the bottom of a publishers site? No, because size does count in Pay per Performance advertising. Small display ads do nothing more than “drop cookies” onto users accessing a page; with users not even viewing the ads themselves. Research indicates that larger ad formats can lift metrics such as branding awareness and message recall by as much as 25 percent, even after just one exposure. For pay per click programs such as Google AdWords set-up message groups within your campaigns; rotating different message as different times of the week.

9. Track site to call centre traffic.
Your pay per performance program needs to recognize that online is at times a tool of researching and information gathering; driving users to complete their transaction offline. Click-to-call solutions are widely used by affiliate marketing networks and offer an effective way of tracking users beyond their online interactions. Other services such as AdInsight integrate call centre data into your Google Analytics account.

Alain Portmann, author of  ”Pay per Performance Advertising – 9 Tips to Drive Incremental Revenue” is Web Liquid’s Founding Partner, Head of Strategy. Visit www.webliquidgroup.com for more thought pieces.

Click below for links to relevant agency/consulting services:

Affiliate Marketing & Pay on Performance
Securing and growing pay-on-performance investments.

Media Optimization
Ensuring your money follows the best performing sites, messages and prospects.

Conversion Optimization
Continuous improvement of conversion by focusing on dynamic landing pages, conversion optimization and site analytics.

About Alain Portmann

My job is simple. To inspire staff, clients and consumers.

Mail | Web | More Posts (65)

The legend of Picasso is at the heart of a contemporary challenge in the advertising industry – the value and cost of ideas. There lies the problem. As an industry we are obsessed with ideas. We complain when these ideas are not accepted. We feel cheated by having to put a price tag on the enterprise of our ideas ” how can I be asked to price passion and the selfless pursuit of an idea?

When we recognize we are in the business of “worth creation” can we begin to shift our thinking from “‘What does it cost us to generate work and ideas a client wants?” to “What is the value of the services and materials we are creating for the client?”

Worth creation forces us to decentralize the idea creation process. Instead, everyone’s job must become value creation. Worth creation forces us to establish a strong personal and commercial relationship with our clients; truly understand their business as opposed to their latest brief. Worth creation demands we measure and place more value on the outcome of our work.

Our latest idea is more than a campaign concept; it is worth creation. Was the agency responsible for creating the Staples’ plastic Easy Button, a $4.99 gadget (that’s sold more than 1 million units since its launch in 2005) aware of that? Apparently not because they received no financial reward beyond their original fees.

Financial advisors are paid on the basis of worth creation. This is accepted given their decisions have a direct and measureable impact on wealth. Digital marketing, like no other channel allows us to directly measure the value created for a brand, be it revenue or perception. This is part of the problem with digital marketing, worth creation has been completely tied to quantitative metrics ” sales, revenue, ROI.

If worth creation is proven and measured every day, the degree of compensation then becomes a question of positioning. If clients regard an agency as just another operator on their marketing conveyor belt, worth creation is not possible. Worth creation requires partnership. Unfortunately most clients regard their agencies as just operators in a large conveyor belt.  In response, and to extend their control and influence, agencies try to be the “jack of all trades”, operators in all realms of digital marketing. The focus is then on depth of offering as opposed to worth creation. These new services are generally sold to clients at a discount “ lowering overall compensation levels.

Client and agency must be willing to invest in worth creation. When this occurs, the conversion of two intangibles – time and ideas ” translates into a tangible and sustainable compensation model.

webliquid agency compensation value creation Pablo Picasso
worth creation

About Alain Portmann

My job is simple. To inspire staff, clients and consumers.

Mail | Web | More Posts (65)

In 2007 we generated GBP £125,422,639 (USD $251,655,481) in revenue for our clients, delivering a return on investment of 23. In simple terms for every £1 a client gave us to invest, we gave them £23 back in revenue. This represents a 34% increase from the previous year.We know it is not all about the numbers either. At the core of these results is a comitment to service and taking care of our clients’ budgets, knowledge and personal objectives.

Delivering these results while maintaining operational efficiency is a challenge. A challenge we have met two years running.

 Every day, one simple statement helps us focus our efforts.

“As our clients trust us with a larger portion of their marketing dollars, and the contribution to the bottom line of those investments grows; increased creativity, service, insight and accountability is required – an environment of thinking:returns.”

Webliquid return on investment ROI

About Alain Portmann

My job is simple. To inspire staff, clients and consumers.

Mail | Web | More Posts (65)

As early as 1989, advertisers and research firms such as Millward Brown, proved the value of combining multiple media and advertising programs to maximise effectiveness. The “Media Multiplier” theory emerged, stating that when two or more media are taken into account, the combined impact is more than the sum total of the two individual media. While this theory has been widely accepted and translated into Digital Marketing, most advertisers are blind to its true magnitude and impact.

Why the last click gets too much credit?

This is largely because the large majority of advertisers still subscribe to the industry standard which attributes a transaction to the last ad interaction. Furthermore, the “rule of attribution” set as a default within third party ad-servers is based on the last click.

This attribution rule has lead to what is referred as “click through tunnel vision” – focusing on the last click before a transaction ” under-estimating the value of the media that preceded that last click. From a practical perspective, the “rules of attribution” are taking precedence over the “rules of advertising” providing a skewed perspective on the optimal balance between PPC advertising and display based advertising.

Under the “last click” rule of attribution, click driven channels such as PPC search, are attributed and credited for a disproportionate amount of transactions. While PPC advertising is an effective online advertising channel, it benefits from being widely used at the start and end point of the consumer purchase cycle. By attributing 100% of a transaction to the last click, there is a danger in underestimating the contribution and value of impression based channels. After all, advertisers don’t just measure the success of a poster by how many people buy the product advertised from the nearest shop.

The evidence

A study conducted by the ATLAS Institute, titled “How Overlap Impacts Reach, Frequency and Conversions,” asserts that 90 percent of the consumers that converted were reached by placements other than the last ad clicked, and that far too often the proper credit for the sale is inappropriately given to search. The study also found that two out of three consumers who eventually took a responsive action were reached by ads across multiple sites before actually going on to make a purchase, and that consumers reached across multiple publishers were twice as likely to convert as those reached only on a single publisher.

As discussed earlier, it is not a question of display based programs being more effective than PPC Advertising or vis-versa, it is a about how both channels work together. This synergy between PPC advertising and Display based programs is confirmed by a study conducted by the ATLAS Institute indicating conversion rates from search advertising is 22% better when used in conjunction with display based programs. Further research indicates 80% of users exposed to display advertising and completed a search, completed a booking within 8 days . What is the reason for this uplift in conversion?

1. Display advertising generates brand awareness and increased purchase intent, reinforcing messages from other channels.

2. Display ads can help win over interested users who may be “on the fence” about purchasing. Display advertising “generates demand” for products and services, while search is far more efficient at “meeting demand”.

3. Search may be used as a navigational tool on a repeat visit to sites which users have previously visited via a display based ad.

Furthermore, the last click “rule of attribution” has a significant impact on the reporting of PPC advertising. The “last click” rule under-represents the contribution of generic search terms as part of the research and purchase cycle on consumers on search engines.

A study entitled “Search Before Purchase” published by Doubleclick, provides insight into the way consumers use search; supporting the importance of generic terms as part of the overall purchase funnel leading to a transaction purchase.

Travel buyers conduct an average of 6 relevant searches in the 12 weeks before purchasing a holiday, car hire, flight or hotel.

76% of keyword searches conducted by travel buyers are generic, with only 21.5% being for brand

Implications

The impact of “last click attribution” on the reporting and management of Online Media is significant.

1. Programs and sites that have a significant impact on reaching and driving purchase intent, but are not attributed credit for the transaction; are removed from campaigns and schedules based on attribution to the last click.

2. Investment and budget decisions are purely driven by the ability of the media to generate the last interaction (last click) as opposed to their impact on the entire purchase cycle.

3. Last click attribution negates the impact of overlap. Overlap being defined as users seeing ads across multiple sites or placements. Research indicates that that, while a minority of users are reached across multiple sites, they consume media at a higher rate than users exclusively reached on a single site. More importantly, this overlap group is responsible for the majority of transactions.

4. The true value of behavioral targeting programs and tenancy sponsorships is not considered; by not crediting media placements aimed at driving relevance and consideration prior to the final click.

Solutions

While it is important to identify the issues and implications related to “last click attribution”, it is more important to identify possible solutions. We have listed a series of solutions, providing a roadmap for consideration and enablement.

1. Get “under the hood” of your campaign data. Commission a research study that analyses the “exposure to attribution path” of your online campaigns. By analyzing the third party-ad server log files, including the initial interaction time stamp and conversion time stamp, it is possible to determine a path of conversion and the sites associated with it.

Complement the “exposure to attribution path” with a Time to Conversion Study, which would identify the time lag between first exposure, last click and transaction. This is a significant data point which will reveal the optimal “length of attribution” for your view and click based media. The default post impression and post click window on most ad-servers is set at 30 days. The window of conversion can be modified, ranging from minutes to days.

2. Customize the rules of attribution within your third party ad-server. Leveraging the data from your research and log file analysis customize your attribution rules. The attribution of online sales by third party ad-servers is flexible given they are rule based. Specific rules can be established, replacing the “last click” default. For example, a rule can be defined which states “last click, if exposed to less than two previous views”. This rule would ensure that a transaction was awarded to a click only if it was preceded by less than two views ” acknowledging transactions driven by more than two views have been largely driven by display based programs.

However it is important to point out the final decision on the attribution rules should be determined by your overall online media mix. An emerging trend is the weighting of attribution between click and view based media. Sophisticated advertisers, aware of the synergies between online media channels, have started attributing 50% of a conversion to the last view and 50% to the last click.

3. Complement third party ad-serving with site analytics tracking. Assign a referring ID tag to all your online marketing programs, creating a unique referring ID by channel (e.g. email marketing, PPC search, display advertising). While site analytics should never replaces third party ad-serving for the optimization and management of online marketing, it should be used as a supporting measurement tool to validate figures.

4. Assign a place in your media plan for conversion influencers. By committing investment to key sites at the start of the purchase funnel, your bottom cost of transaction and brand will benefit. However this approach requires budget holders, the client, to buy into the value of these investments. This can be challenging given the propensity to measure the value of online media on the basis of attributed revenue. Furthermore, and as discussed in What is Return on Investment, there is a propensity to view return as a mutually exclusive metric, which does not account for exposure or influence.

More importantly, a model that measures and recognizes consumption, as opposed to just interaction, is required to gain a more balanced and accurate view on the contribution of online marketing to the bottom line.

webliquid PPC search online measurement doubleclick

About Alain Portmann

My job is simple. To inspire staff, clients and consumers.

Mail | Web | More Posts (65)

Finding: Users exposed to three distinct display advertising messages throughout the campaign, produced average revenue 47% higher than the users who only converted through search.

Learning: Sequential serving of multiple messages across display campaigns combined with search will generate increased revenue.

Finding: 60% of users exposed to search only used brand terms before purchasing. 69% of users exposed to both search and display ads used brand search terms before purchasing.

Learning: Display advertising will prompt brand recall, driving increased demand for brand search terms across search engines.

Finding: A frequency of 10 ad exposures generated average revenue of £393, while users on less than three 3 ad exposures generated average revenue of £172.

Learning: Campaigns aimed at “meeting demand” should be frequency capped to 3 to 4 exposures, while campaigns aimed at “generating demand” should be assigned a higher exposure frequency. While low frequency (1 to 3) drives efficient conversion volume, increased frequency will drive incremental conversion quality by reinforcing the brand.

About Alain Portmann

My job is simple. To inspire staff, clients and consumers.

Mail | Web | More Posts (65)

Imagine search engines to be rooms with millions of people. When a specific question is submitted, the people in the room talk among themselves, to determine the experts in the subject. Once the experts are identified these vote over who should speak first, second, third and so forth. Recently which expert speaks first is also determined by their relationship to other experts outside the room. The role of organic search (usually known as natural or search engine optimisation) is to ensure a brand and website(s) are identified as experts and allowed to speak first on a specific question and subject. The room also features a paid members section, in which a selected number of individuals are allowed to speak first based on how much their wiling to pay to be heard first, second and third. The response of these paying individuals is included as part of the expert responses. The role of commercial search (usually known as PPC search) is to ensure your brand and site is featured as commercial responses relative to a specific subject and question.

As the metaphor points out, organic search should be viewed as a long-term foundation program, which supports tactical short and mid term commercial search activities. While the benefits of search are understood, the challenge many marketers face is defining one set of common metrics to evaluate both commercial PPC search and organic search. In our experience there are eight key metrics that can be used for both commercial and organic search.

1. Rank (metric is visibility)
2. Traffic (metric is traffic to site)
3. Persuasion (metric is initiated shopping cart instances)
4. Sales Volume (metric is volume of conversions)
5. Sales Efficiency (metric is cost per conversion)
6. Net Revenue (metric is total revenue)
7. Revenue per Item (metric is revenue per conversion)
8. Return on Investment (metric is ratio of return)

The majority of advertisers measure the contribution of organic search through a web analytics solution (such as Web Trends, Google Analytics or Omniture) – by placing a referral ID on all commercial search links, the contribution of natural search is determined by subtracting all commercial search from the total search contribution to the site. While this approach is effective in the short term, it lacks the sophistication required to answer key strategic decisions such as the optimal media allocation on commercial search given the contribution of organic search. A series of technology companies including Hit Dynamics, Doublelclick and Mediaplex provide in-depth tracking for natural search and commercial search. More importantly, these solutions can allow advertisers to de-jup conversions against all other channels including PPC, display and affiliate marketing.

Following a period of consultation with a client we defined the following reporting structure, which takes in consideration the visibility, engagement and transactional impact of organic search.

Visibility Reporting: How well ranked?
- Searches that returned at least 1 result
- Searches that returned at least 1 Top Ten result
- Searches that returned a number 1 result
- % Search area occupied
- % search are occupied by top 10 results
- % search area occupied by number 1 results

Engagement Reporting: How did visibility translate into traffic, bookings and revenue?
- Actual Traffic generated (Visits)
- Actual Conversion rate (visit to sale %)
- Sales Generated
- Actual Revenue Generated
- Actual Average Booking Value

In the next few weeks we will be publishing a series of knowledge pieces on the best way to manage commercial and natural search; examining the benefits of using different remuneration models other than media commission as a way of having your search agency work harder on exploiting the search tail.

web liquid, organic search, natural search, search engine marketing,sem

About Alain Portmann

My job is simple. To inspire staff, clients and consumers.

Mail | Web | More Posts (65)

In 2006 we generated $133,535,453 in revenue from our clients’ investments in online media, delivering an overall return on investment of $19:1. The 19 to 1 profitability index takes in account all costs including creative, media, agency remuneration and ad-serving. These figures represent a 242% increase from 2005.

We thank our clients for their trust in our proven methodology and look forward to an equally successful year ahead.

About Matthew Cronin

Mail | Web | More Posts (32)

But beyond all that we’ve identified one metric that really matters – the success of our business is, after all, tied to the success of our clients. It’s for that reason we are very proud to have delivered over $57 million in revenue with an average ROI of 20:1 for our clients in 2005.

About Matthew Cronin

Mail | Web | More Posts (32)