Background

We have received numerous inquiries about the paid click data released by Google on 4/17/08. Some inquiries focused on reconciliation: How could comScore and Google figures be seemingly so different? Others, recognizing that comScore’s data reflected domestic paid clicks only, questioned why we release data that only partially represent Google’s global business. Others asked about our plans for offering our own analysis and training on the proper usage of our data. Our answers to these questions are provided below.

Summary

The main difference between the paid clicks trends reported by Google and comScore can be traced to the fact that the comScore paid click data cited in financial analysts’ reports (and subsequently reported by the media) are U.S. data only. Analysts’ efforts to use comScore’s domestic data to estimate Google’s global trends were misguided. In fact, a detailed analysis by comScore of the global paid click data publicly reported by Google indicates that Google’s U.S. trends are:

1) Weaker than their global trends, but
2) Are consistent with comScore’s U.S. paid click data

It also needs to be noted that comScore does not currently provide a measurement of global paid click data and has never claimed that its U.S. data can be used to “predict” global trends. In fact, in late February, comScore posted on its blog a warning of the potential misuse of its U.S. paid click if used to predict Google’s first quarter global financial results. Unfortunately, despite this, many in the financial community and in the media have chosen to do otherwise.

A) Making the Right Comparisons
To do a fair and accurate comparison, it is first important to understand the differences between the comScore and Google paid click metrics. Google’s publicly reported “aggregate paid clicks” refers to global paid clicks that come from search on owned and operated sites and partner sites, as well as paid clicks that come from the Google Content Network (O&O + AdSense). On the other hand, comScore’s paid click report currently refers to domestic paid search clicks only. The following table summarizes the differences:

To reconcile the two data sets, it’s clear that we need to do a comparison based on U.S. trends only and try to make the comparison as “apples to apples” as possible. While comScore is working hard to develop a measurement of AdSense paid clicks, at present we only report on paid search clicks and try to get as close as we can to the definition of what Google reports. When we include U.S. paid clicks for AOL and Ask in the comScore data, the trends in Google’s paid clicks change as follows:

To get to a comparable U.S.-only figure based on Google’s publicly reported data we started with Google’s reported U.S. revenues and adjusted for CPC price increases. While Google reported Q1 ‘08 total revenue growth of 7% vs. Q4 ’07, international revenue was up approximately 14% and domestic growth (excluding the DoubleClick acquisition) was flat. Based on Google’s publicly reported data, one can conclude that Google’s CPC prices increased by 18% Y/Y and 3% Q/Q. The following table provides a trend summary:

Unfortunately, this is as close as we can get to an apple to apple comparison based on Google’s publicly reported data. We cannot, at present, make adjustments for paid clicks on the Google Content Network (AdSense, Gmail, etc.) and search partner sites other than Ask and AOL. Nevertheless, the differences between the comScore and Google data have narrowed considerably and are directionally the same:

This contrasts with what an inexperienced comparison would, at first glance, believe to be a large difference:

B) Key Takeaways

Quantitatively, we have shown that when we compare the comScore and Google data on a consistent basis the differences in the trends of Google’s U.S. paid clicks are small. But, one might ask: what about qualitative analysis? Did comScore’s interpretation of its own data accurately anticipate the key conclusions from Google’s actual first quarter results? We believe the answer is a resounding “Yes”. While many in the media and the financial community had originally interpreted the comScore U.S. paid click data to foretell a negative impact from the U.S. economy and a disappointing Google first quarter, comScore’s own diagnosis can now be seen to have been remarkably accurate. comScore’s analysis, originally posted on the comScore Voices blog and dated February 29, 2008, asserted:

• Google’s growth in U.S. paid clicks has decelerated.
• The softening in Google’s U.S. paid clicks was not driven by the economy:
  “While we do not claim that these concerns are unwarranted, we believe a careful analysis of our search data does not lend them direct support.“
• The key driver was Google’s own quality improvement efforts:
  “The evidence suggests that the softness in Google’s paid click metrics is primarily a result of Google’s own quality initiatives that result in a reduction in the number of paid listings and, therefore, the opportunity for paid clicks to occur.”
• CPC’s are likely to increase due to the reduction of sponsored ad supply:
  “The reduced supply, as well as the higher minimum bids, contributes to an increase in the price per paid click, which is what helps counteract the slowdown in the absolute number of paid clicks. Therefore, Google’s revenue will not necessarily suffer from this.”
• Google’s first quarter revenues will likely be strong despite the sequential decline in paid clicks:
  “It is entirely possible, if not likely, that the improved revenue yield will continue to deliver strong revenue growth in the first quarter.”

It is ironic that this blog posting was criticized by some readers when it was first published. It seems that too much attention was myopically focused on just one data point, the trend in Google’s U.S. paid search clicks, an admittedly important metric, but one which quantifies just a single element of a multifaceted revenue equation. This was accentuated by the coincidence of two major changes that were happening in parallel and which can also have had a negative impact: 1) The deteriorating U.S. economy stoked fears of a slowdown in online advertising that was previously assumed by some to be immune to an economic downturn, and 2) Google’s own advertising quality efforts negatively impacted paid clicks – just as a poor economy might have. In statistics, they call it collinearity. It always wreaks havoc with data analysis!

Regardless of who took what position, there are important lessons for us. While comScore’s first priorities are to provide our clients with the highest quality data and unparalleled innovation in measurement, it’s clear that we can also do a better job of explaining the meaning and the limitations of the statistics we publish, not only to the financial community but also to the media. comScore’s data has become so widely used by so many different constituencies that it is incumbent on us to do this.

In addition, we will continue to strive to become the first company to provide the desired and more complete picture of the global search market, beyond that obtained from U.S. paid search clicks, by providing international estimates and reporting of contextual ad clicks. We have been working on this challenge ever since we released qSearch 2.0 last August, when we became the first and only company to measure worldwide search queries. Though we are getting close to measuring paid versus algorithmic clicks globally, the effort has been no easy task. Nevertheless, we should all remember that paid click data provide only part of the picture when trying to understand and forecast search revenues. For example, in the comScore database we currently have no way to directly measure CPC prices, which are necessary to accurately translate paid clicks into revenues. While we believe that many of the financial analysts using our data understand this, we should all be mindful that the primary uses of comScore’s data are for marketing and media analysis purposes, where being able to accurately measure Internet users’ online behavior is the key objective. We know we do that with excellence. That said, the commentators who pointed out that other data points need to be added to comScore’s data so as to paint Wall Street’s optimal financial picture, got it right. We agree, and have never pretended otherwise.