Campaign gross ratings points (GRPs) have recently been discussed as a breakthrough in online ad campaign measurement and cited as the ‘new online currency.’ However, there is nothing new about it. These metrics have existed for many years, with comScore supplying online GRPs for more than five thousand ad campaigns since 2007. Today, comScore’s Campaign Essentials service, first introduced in 2010, is the global industry leader, being used to analyze hundreds of online campaigns each month across 17 countries. Campaign Essentials builds upon Campaign Post Buy, a comScore service first introduced in 2007, which first delivered campaign GRPs and demographics, and has been used to evaluate thousands of campaigns since inception. This product evolution is consistent with comScore’s track record of leading innovation in all aspects of online measurement.

At the same time, our customers, who have collectively measured thousands of online ad campaigns, tell us that GRPs, while useful in TV, are much less useful unto themselves for online measurement. The very nature of online advertising, where cookies are used by ad servers as the determinant of exposure frequency but which are deleted by about 30% of Internet users in a month, can result in severe delivery skews and exposure overkill, with a small number of users being exposed to hundreds of impressions, while a substantial number of users receive only one impression (that they may not even see because they never scroll down “below the fold”). An illustration below shows an example of two campaigns with the same level of GRPs but more than a 300% difference in audience reach.

Campaign Essentials leverages comScore’s pioneering Unified Digital Measurement (UDM) methodology, which combines person-level data from our proprietary panel with census-level page views or campaign impressions obtained from ad server data, to deliver a highly accurate measurement of people reached with an ad campaign. comScore first introduced UDM two years ago and expanded this approach globally in early 2010. We currently receive census page-level information from more than 90 of the top 100 U.S. media properties, and 80 of the top 100 global media properties. This industry adoption reflects a consensus that any accurate online measurement service must have a combination of a robust person-level panel data with a census-level site-side measurement overlay.

comScore Campaign Essentials’ approach helped overcome the perils of cookie inflation caused by cookie deletion and users using multiple browsers or machines, offering a significant breakthrough in online campaign measurement. Similar to site audience measurement, campaign measurement needs to be unified with ad server counts, which measure every delivered impression but can only report campaign reach based on unique cookies rather than people. This cookie-based reach is typically inflated by 50% to 300% because of cookie inflation. Obviously, reach is a critical objective that advertisers need to have measured accurately.

comScore has solved the complexities of adapting the UDM methodology to campaign measurement, including granular reporting at the creative, placement, publisher and country levels. It is because of the UDM methodology that we are able to provide this granular methodology at scale on a global basis for the 43 individual markets currently reported by comScore. The methodology even allows clients to create custom segments to answer specific business questions related to their campaign objectives. This granularity is available for demographics beyond just age and gender, such as household income, household composition, region and any other demographics reported in comScore Media Metrix.

So the availability of GRPs is not news, but the learning about their applicability in the digital environment has not yet been broadly absorbed. In short, comScore’s position is that GRPs in and of themselves are simply not sufficient for either planning or measuring the effectiveness of campaigns. By leveraging the extensive learning we’ve had from our UDM initiative – which combines panel and census data – over the last several years, comScore is working on a number of initiatives to address this, which we think will have important implications for the industry.

A core ingredient in these developments is the comScore panel, which is the largest continuously measured panel of its kind. This enables us to accurately measure the audience demographics of a campaign at a granular level. Our available sample size is many multiples of the sample size required for accurate statistical measurement of audience demographics. Larger sample sizes are only better if they accurately reflect the overall demographic composition of the population.

Our goal has always been to provide our customers with the highest quality and most useful information possible. Importantly, the accuracy of the demographics used in the sample is significantly more important than the size of the sample alone. In contrast to getting person-based demographic data as used in Campaign Essentials, resorting to the use of demographics from cookie data provided by third-party source sites, can be problematic for a number of important reasons:

1. Publishers each have unique demographic profiles which might be disproportionately represented in the reported campaign demographics. As a result, when comparing campaigns that run across different publishers, the demographics will favor the publisher that was used as the source site.
2. The demographics available from source sites, particularly social networks, are often restricted to age and gender and are missing important variables such as income and household size. Many global advertisers also need to understand dynamics among locally relevant demographics that reflect the way planning and buying is done in country. For example, in France and the UK, reach among different social classes is important, while in many markets, the presence of children is a critical segment.
3. Self-reported age information is not always accurate in an online environment. Teenagers often overstate their age and adults (especially women) tend to understate it.
4. Reach within a typical campaign may be limited to machines logged into the source site at the time of exposure, which represents a fraction of the total campaign audience. In addition, there is no guarantee that the logged-in ID, stored in the login cookie, represents the real person using the machine at the time of exposure. With widespread sharing of machines by individuals within a household (60% of U.S. Internet users are on multiple-user computers), this cookie-based method can result in falsely attributing the demographics of one user to another. Therefore, the cookie approach adds substantial error far overshadowing any marginal improvement in precision that a larger sample size may ostensibly provide.
5. To measure the entire campaign, outside of the reach of the source site, there is a need to ascertain the demographics of cookies with unknown demographics using panel data, and the size and breadth of comScore’s panel provides optimal quality of results.

Perhaps a larger issue is that when data from select publishers is used to generate reporting on other publishers, it is no longer objective third-party measurement. For example, when media planners receive a report on how well site A is reaching the demographic target, site A will tend to look strong if it is the source of the demographic information, because its own data are used for both targeting and demo reporting. If media buyers are buying on percentage of target delivered, site A may appear to hit 100% of the target while site B ends up only delivering 80%. The reality is that site A might have also delivered 80% but appears to have delivered 100% because its data was the source of the reporting. Naturally, publishers, advertisers and agencies are interested in neutral third party measurement which does not favor one site’s audience over another.

Finally, comScore Campaign Essentials is delivered via a powerful but intuitive UI. This UI provides for flexibility to view a host of variables (audience by site, placement, creative size, creative type, creative theme, etc.) and also allows for custom aggregation of placements. For example, if a marketer would like to bucket various placements into themes such as contextual, audience, remarketing, etc., simple drag-and-drop technology enables this powerful insight.

comScore’s perspective is that accurate demographic, reach and frequency reporting is critical, but is only the beginning of a marketer’s need for campaign insights. Through the comScore AdEffx suite, we also currently provide information on creative messaging optimization, branding impacts, online and offline sales, and behavioral interactions with the ad campaign, such as searches for the advertised brands, or subsequent visitations to the advertiser’s site.

True to comScore’s heritage of forward-looking innovation, we have been hard at work for the past twelve months developing the next generation of campaign measurement products tailored to meet fast-changing industry needs. During the next 60 days, we expect to introduce solutions that we believe have the potential to revolutionize the industry and usher in a much more valid new currency. Our strategy has always been to be the first to develop the solutions our industry seeks, and to constantly renew our edge even when our competition eventually catches up with yesterday’s innovation. Our approach to campaign measurement is no different.

comScore is excited to announce that we have released our first beta data set from the comScore Total Universe report, and we have some very interesting results to share. As background, the comScore Total Universe report provides audience measurement for 100 percent of a site’s traffic, including usage via mobile phones, apps, tablets and shared computers such as Internet cafes. This first-of-its-kind measurement capability enables publishers and advertisers to understand digital audience sizes in a cross-platform environment. As media consumption shifts to emerging platforms like mobile devices and tablets, the effects on accounting for audiences and activity levels only becomes more pronounced with time.

In order to grasp the effect that Total Universe reporting has on audiences, let’s examine the impact for a few key publishers. An excellent example is for digital music provider Pandora, which attracts a significant percentage of its users via mobile (which in this case includes web browsing on phones and tablets, as well as access via mobile apps). Pandora’s U.S. Total Universe audience in April 2011 was 31.8 million unique visitors, 74% higher than its audience coming from home & work computers. In fact, 13.4 million of that total (42%) accessed Pandora only through mobile channels. A greater number of people actually accessed Pandora via phone or tablet (20.7 million) overall than did via home & work computers (18.2 million). Pandora’s reach among home and work computers in April was a very respectable 9%, but an incredible 35% among smartphone users.

To exclude the mobile media channel from Pandora’s audience is to ignore the majority of its audience. And while Pandora may still be somewhat of an outlier case, the example highlights the extent to which mobile can drive traffic for certain brands – especially those whose value proposition largely relies on mobility, such as Internet radio. It’s also an indication of where the digital landscape is headed.

Another illustration of how mobile audiences fit into the Total Universe landscape are news sites. Some of the most commonly used media on mobile phones are email and social networking, both of which are popular sources of link-sharing, which can drive significant click-through to news sites. If we look only at mobile web traffic (i.e. mobile browsing of traditional and mobile-optimized websites) as incremental to the existing home & work computer audience, we can see the level to which the mobile channel can attract additional users.

Below is an illustration of the incremental audience reached via mobile web for several leading newspaper brands. The New York Times, the largest digital newspaper, has an incremental reach of 2.3% for its mobile web, which represents nearly one million individuals – an audience certainly too large to ignore. And as significant as the New York Times’ incremental reach might be, other leading newspaper brands like the LA Times (6.7%), Washington Post (4.9%), Wall Street Journal (5.5%) and Chicago Tribune (5.2%) have an even greater incremental reach.

We can see a similar pattern in the UK (compared to the US data) with some of its leading newspapers. The Mail Online has a Total Universe audience of nearly 11 million in the UK, with an incremental web reach (mobile + shared locations) of 4.2%. The Guardian (4.8%), Mirror Online (8.3%), the Independent (6.8%) and Metro.co.uk (6.4%) all have notable incremental reaches, as well.

What these data suggest for publishers is that their total audiences are being shortchanged if they do not account for mobile audiences, which provide access to more eyeballs. In most cases today that incremental reach may be less than 10%, which still should not be ignored, and in certain cases (e.g. Pandora) that additional reach may be even more significant.

But if we take a long-term view of the digital media landscape, we recognize that mobile media usage is only going to become an increasingly large piece of the media consumption pie. Smartphone and tablet users are shifting a meaningful share of their total media consumption away from traditional PCs. Continued adoption of these digital access platforms has significant implications for publishers who get paid by advertisers to reach specific audiences at scale.

The growth drivers are undeniable. Nearly 70% of Americans do not use smartphones, and even with the excitement around the iPad and other tablet devices, they are used by less than 5% of the population. Increased penetration coupled with the explosion of optimized content and services will change the digital landscape.

How will it change? How much digital consumption will migrate away from the traditional PC? Who will capitalize on the opportunity? These are great questions. comScore’s Total Universe report will help you answer them, so stay tuned for further developments.

Importantly, the value of the Total Universe report depends on publishers tagging their digital content entities, including traditional web pages, mobile-optimized web pages, and mobile apps. Any publishers interested in participating in comScore tagging in order to be featured in the Total Universe report, please visit comScore Direct.

When comScore publicly releases its September qSearch reporting this week, there will be a couple of important enhancements to highlight. The first change is that September will be the first month reflecting the impact of Google Instant, Google’s new feature that delivers results in real-time while users type their query. The second change will be the inclusion of added metrics around the U.S. explicit core searches “powered by” Google and Bing.

Google Instant Search Reporting
Google’s “Instant Search,” announced on September 8, 2010, alters the user search experience considerably. Prior to the introduction of Instant Search, a Google user typed in a search keyword or phrase, which dynamically generated a pick list of query suggestions designed to help the user select a query that best reflects his or her intent. The user would keep typing until she found an appropriate search suggestion or hit enter to submit the query she typed in the search box.

With Instant Search, the user is now additionally presented with a full results page that corresponds to the top ranked suggestion in the suggestion list. The page is dynamically updated as the user continues typing, which could potentially generate a large number of result pages that may or may not be of user interest. At the same time, the newly added detailed results provide the user with more information to formulate a satisfactory query and offer new modes of explicitly engaging with a result page. In particular, the user no longer has to hit enter to begin interacting with the result set on the screen.

Fortunately, the comScore panel provides visibility into all events that a user is conducting and all the HTTP calls associated with the user’s actions. Based on this insight, we have developed a priority scoring system that allows us to identify search results with explicit user action and interstitial results with a sufficiently long pause to suggest some level of implicit engagement. User actions that qualify an explicit query include those where a user hits “enter,” clicks on an algorithmic or sponsored result, clicks one of the various refinement links (such as past 24 hours, etc.) or clicks on a vertical search tab to execute the query in a different channel (such as News or Image Search). Any query with such explicit action is counted in Explicit Core Search. Query result pages without explicit user action, but with a pause of at least 3 seconds, are considered as indicating ‘implicit’ engagement and will count towards Total Core Search.

“Powered By” Reporting
Beginning this month, comScore will begin providing insight into the share of Explicit Core Searches that are powered by the two major providers of algorithmic search results – Google and Bing. Google’s “powered by” share is composed of searches conducted at Google entities, as well as searches on AOL and Ask’s MyWebSearch where Google branding is present. Bing’s “powered by” share is composed of searches conducted at Bing entities, as well as Yahoo! searches where Bing branding is present. This primarily consists of queries at Yahoo! Web Search, Yahoo! Image Search and Yahoo! Video Search. Some of Yahoo!’s in-channel searches, such as Movies or Finance, are still provided by Yahoo!

When comScore reports our July 2010 qSearch data this week, we will begin reporting an additional view of our U.S. Core Search data known as “Explicit Core Search” alongside our standard Core Search market reporting. comScore defines “Explicit Core Search” as user engagement with a search service with the intent to retrieve search results.

As I described back in June, the search market continues to evolve with new implementations of search moving beyond the traditional search box and into the context of the browsing experience as the user engages with non-search content. We called these searches “contextual searches” because they are originated in the context of other browsing behavior without explicit intent on the part of the user to conduct a search query.

To help the market assess the impact of these contextual searches on the search market, comScore has provided its qSearch clients with detailed release notes each month that quantified the magnitude of this effect.

Ultimately, comScore is committed to providing our clients and the broader marketplace with increased transparency into the search market. Going forward, we will include in our public release of the monthly U.S. search data both the traditional Core Search and Explicit Core Search views of the market. This will empower each interested stakeholder to determine which view of the market they deem most appropriate depending on the particular circumstance.

We look forward to continuing to provide the most advanced search market reporting in the industry and we are committed to doing so in a way that is meaningful for all involved.

comScore will publicly release May qSearch data tomorrow, tracking the market share of all the major search players in the industry. We have been publishing search reports since 2003 and have watched (and tracked) many changes in the industry as it has grown to dominate the digital advertising space.

In July 2007, comScore revamped search measurement with the introduction of qSearch 2.0, which expanded the view of search beyond the traditional search sections to include searches at retailers, travel sites, directories, etc… anywhere that a user might submit a search. qSearch 2.0 captured the broad adoption of search technology throughout the web to improve the user experience.

Since that time, we’ve seen a wave of change across the web that has changed the very nature of a web page, from an object that is requested and delivered, to one that is a live platform that can integrate content from many sources. And search has changed along with it…

The recent innovation deployed by a number of search engines is to go beyond simply providing a search box along with content, but to weave their search engines into the user experience.

There are at least two reasons they’re doing this. Search can be used as a powerful contextual content discovery technology. By providing search results that are highly relevant to the content being consumed by a user, properties like, MSN and Yahoo! can provide intuitive and convenient content discovery experiences. Also, by providing search results in context across their network, those sites are able to leverage the size of their audience to expose more users to their search services.

Traditionally, the industry has thought about a “search” event as a submission of a query that subsequently presents a set of results to a user. comScore’s definition of search requires that the user be presented with search results and be able to completely change or refine their search directly from the result page. This encompasses the traditional “text box” query, as is the case with the major search engines’ main search entry point.

Some context-driven search experiences also meet comScore’s current criteria for qualifying as a search and are therefore counted in qSearch market share reporting. At the same time, comScore recognizes that these are inherently different experiences compared to traditional web search queries. And because context-driven searches are sometimes monetized at different rates than traditional searches, we believe it is important to provide the marketplace with visibility into how they are contributing to search share. For this reason, we will continue to explicitly quantify context-driven search volume in our monthly release notes to clients.

That said, the continued evolution of search and emerging innovations in how it is used to enhance user experience, calls for a thoughtful review of how we classify various types of searches, count them and report them. We want to ensure that we provide comprehensive and flexible measurement that meets the needs of the various constituencies in the digital marketplace. As our thinking evolves, we will include relevant stakeholders in the discussion and clearly communicate our thinking and rationale to the marketplace. While we will maintain the current method through the end of the second quarter to avoid reporting disruptions, we will aim to implement proposed revisions in the third quarter, ideally starting with the release of July data in the first half of August. Stay tuned to the comScore blog to find out more in the coming weeks.

I’ve run into a handful of cases lately where distinct IP Address counts have been offered as evidence that comScore’s numbers underrepresented an audience. These are very easy cases to address, but their increasing frequency made me wonder if the “Myth of Static IP” – that IPs will become equivalent to Users – was undergoing some sort of revival. I’d like to run through a few facts about what IPs represent with respect to Unique Visitors and show that IPs are a poor proxy for UVs even in a broadband world.

It's common knowledge that ISPs rotate IP addresses. The IP address I have today could be the one you have tomorrow. With a dial-up connection, you get a new one each time you connect. With “always-on” broadband connections, the IP addresses are more durable, possibly even static over long periods of time. As consumers began to adopt broadband technologies in earnest several years ago, a myth arose that IPs would soon become a good proxy for individual households.

We looked into the matter when we were researching cookie deletion in 2007, finding that fewer than half of US machines used the same IP during a one month period, and individual machines used an average of nearly 6 IPs. We asked around the industry to verify our findings and found very little surprise. The general tone of the response was: “You didn’t know that already?” We moved on with our cookie research.

But given the apparent renewal of interest in using “Unique IPs” as a measure of “Unique Visitors,” I thought it was time to update the IT research and make a few facts known. All figures are built from comScore’s research panel activity in June 2008.

US IP Address Counting Inflates counts by more than 5 Times

In June 2008, nearly 60% of all machines used a single IP address, which is slightly more than when we initially saw in early 2007. However on average, US machines used 5.7 distinct IP addresses in the month, which means that 40% of machines that change IP addresses during the month do it with great frequency. What’s most striking about these distributions is how the multiple-IP machines drown out Single IP machines. Of the distinct IPs used by US machines in the panel, just 10% came from Single IP machines. Only 1 in 10 observed IPs satisfies the Myth of Static IP.

When you break the results down by location of use, its clear that the problem is significantly less at Work where broadband is the norm. But even at work, where 75% of machines use a single IP, these single IP users account for just 28% of the total IPs that a publisher would see. If we use work to represent how IP counting would perform in a broadband world, its pretty clear that IP Counting will continue to be too fraught with issues to represent Unique Visitors.

(Note: we are setting aside for the moment the core workplace IPs problem: Multiple users appear with the same IP address to publishers. This situation presents the opposite problem for people counting IPs. A single IP represents many individuals. We’ve heard this fact cited as evidence of why IP counting is actually conservative. If anything this fact renders the problem unsolvable for publishers – there is error in both directions, but you don’t know which direction is any given case.)

Broadband Is Not Static IP

The root of the Myth of Static IP is the dawning of the broadband network with “always-on” connections. It was expected that with more broadband adoption, we would get more reliable IP counting. This assumption also falls short. When we break downthe U.S. Home total by connection type, it’s clear that only Cable approaches the static IP levels seen at work. Dial-up predictably uses piles of IPs for each machine, and DSL falls in between. Only 30% of DSL machines used a Single IP address during the month, and the average DSL machine used more than 10 IPs. Of these residential technologies, only cable is even close to living up to the Myth. But even Cable suffers from the core inflation issues, averaging 2.7 IPs per machine.

IP Inflation Issue is Amplified Globally

If you pull back and look at the IP Counting issue on a global scale, it becomes clear that the Myth of Static IP is simply not true. Outside of North America, average number of IPs per machine in a month are roughly double those in North America, ranging from 10 in Western Europe to 15 in the Middle East and Africa. Worldwide, only 4 in 100 IPs observed in June came from a Static IP machine.