Can Google Be Beat On Competitive Web Intelligence Data?

Banner promozionale per DoubleClick Ad Planner
Happy days: statistics for almost any website from DoubleClick Ad Planner

One of the great promises of digital marketing is a presumed availability of quality statistics on website usage and users. Web intelligence data can be used to benchmark the competition, to evaluate potential new markets and to answer many interesting questions along the way. Many companies are active in this area, most notably Nielsen, comScore, Experian Hitwise, Compete and Quantcast, One supplier stands out for its ability to offer statistics on websites across most of the world at an unbeatable price: Google.

Suppliers of website market research data are generally very good at issuing authoritative sounding press releases and providing reports with pretty graphics, yet an appearance of authority sometimes masks an ugly reality: the vendor has a very limited sample of data to work from. As an example, ’s June 2012 panel size in the Italian market, as reported by Audiweb, was 41,256 people, i.e. less than 1 tenth of a percent of the entire potential internet population in that market. If this data sample isn’t representative of a site’s visitors and usage patterns there is an all too real risk of making business decisions based on dubious data.

Some web statistics suppliers rely on measuring users who have agreed, knowingly or not, to be tracked as a condition to download free software. Others recruit their user panels primarily through random digit dialing, or by buying data from a few ISPs. Each method includes pitfalls which can skew the data sample (will more savvy users knowingly download software which tracks them? Are the users of ISPs that sell their data truly representative of the greater internet population?).

When IAB president and CEO Randall Rothenberg arrived in 2007, he was aghast at the state of measurement affairs. In a letter to the panel measurement leaders at the time, Nielsen and , he wrote:

After decades of media research methodologies premised on early 20th century innovations in statistical sampling of barley yields, we could at last move … into a census-based universe that would let marketers eliminate waste, media companies realize a fair price, and advertising agencies target audiences and analyze their campaigns more effectively.

… our members’ server logs continue to diverge starkly from your companies’ sample-based assessments, by 2x to 3x magnitudes in some cases – far beyond any legitimate margin of sampling error.

(The emphasis is mine). Our colleagues in the IT field describe the problem in rather blunt terms: garbage in equals garbage out.

Enter Google With Doubleclick Ad Planner And Google Trends For Websites

Google Trends for Websites, click to enlarge
Figure 1: Google Trends for Websites, with data for Klout, Peerindex & Kred (click to enlarge)

In 2008 Google entered the digital audience measurement fray by introducing two free tools: Google’s Ad Planner, later renamed Doubleclick Ad Planner, and Google Trends for Websites, for the most part a more limited version of AdPlanner.

Since Rothenberg’s letter most digital measurement players, Google included, have applied to a US industry advertising body, the Media Rating Council (MRC), for accreditation. Most applications are outstanding; Google’s has languished for years – putting the efficiency of the MRC process in doubt.

Several suppliers, such as Compete and , provide tracking code sites can install directly in their web pages, potentially providing good, if not excellent, tracking of these specific sites (data quality still depends on several factors, including the tracking code type (synchronous or asynchronous) and the thoroughness of the code installation). In 2009 comScore began moving to this model as well. Their press release is interesting as they needed to describe a process which was going to significantly improve data quality while trying to minimize potential concerns about the quality of prior statistics within their client base. Rather disappointingly comScore disingenuously discusses perceived issues with server centric statistics capture, which might have made sense in 2000, but by 2010 most companies, encouraged by web analytics vendors, had migrated to browser centric (tags) measurement, exactly the new approach comScore was implementing. Needless to say, the press release makes “interesting” reading to those in the know. Initially comScore required companies to pay to be tracked more accurately. The logic of charging companies to help comScore improve the very data it then sells to other companies proved unjustifiable and comScore relented.

What Sets Google’s Competitive Intelligence Tools Apart

Google’s Ad Planner and Trends for Websites competitive intelligence tools stand out on a number of fronts:

  1. Of all the suppliers of website analytics data, no one can match Google’s data collection ability.
  2. Many tools focus their coverage on a limited set of markets. Google offers as close to world-wide coverage as it gets.
  3. Google doesn’t charge companies to collect nor to consult data.
  4. Google has a strong engineering culture, a data-driven culture which puts a premium on data collection and usage. Google has continued to improve the quality of Ad Planner data over time. It isn’t clear to what extent Google’s competitors in this space put the same emphasis on data quality.

A Few Ad Planner And Trends For Websites Limitations

  1. Google provides little historical data. ’s primary data view is of the previous month; it isn’t possible to consult a data snapshot for prior point in time, an issue for sites which experience seasonal variations in a particular month. A graph displays Daily Unique Visitors over time for multiple months which somewhat mitigates this issue. Google’s related Trends for websites offers the same graph with additional options to filter by year and geography.
  2. Google collects very little demographic data on its own. Where Facebook knows a user’s age and gender, and perhaps their level of education, profession, religious and political affiliations and even sexual orientation. Except in the few markets where Orkut is/was strong, Google doesn’t know much about its users. Google may eventually capture this information through richer Google+ profiles but for now it appears that Google is treading very lightly. The available demographic data, supplied primarily by unspecified third parties, is limited to 10 countries (Australia, Brazil, France, Italy, Japan, Germany, Spain, Switzerland, UK, US).
  3. Google is not explicit about the methodology it uses to calculate the various statistics it offers, nor is it clear what level of accuracy to expect.
  4. Google won’t show data for “niche” sites for which they believe they have insufficient data.
  5. Google won’t tell you about traffic to *.google.* domains. So much for fairness and transparency.

Benchmarking The Benchmarks: How Accurate Is Google Ad Planner’s Visitor And Page View Data?

Google collects website usage data from a multitude of services, but does Google use it? Not necessarily. Google’s data usage policies may prohibit data sharing within Google unless a user has explicitly consented, as in the case of . There may be internal conflicts (the suppression of organic search keywords for logged in users is but one example) which hinder sharing, as in all large companies, and one hand may not necessarily know what the other is doing (as we’re told was the case in the Street View wifi data collection debacle).

What we can do is choose a set of sites and compare their known traffic figures from a source like Quantcast with Google Ad Planner statistics. Quantcast says it has more than 100 million sites under direct measurement, although not all data is publicly accessible. The same type of comparison can be made for any site using Google Analytics or a similar digital analytics system, provided we have access to the analytics data.

SiteGoogle Ad PlannerDirect measurementNotes
AdAge910K / 420K Unique Visitors

2.9M Page Views

632,230 Unique visitors (people)

1,233,473 Visits

4,961,970 Page views

Direct measurement: Quantcast
LinkedIn230 / 98 M Unique Visitors

2.7B Page Views

108,993,456 Unique Visitors

317,647,168 Visits

2,707,373,056 Page views

Direct measurement: Quantcast
Anonymous reference site (Italy)48K / 20K Unique Visitors

510K Page Views

56,619 Unique Visitors

493,524 Page Views

Direct measurement: Google Analytics
Wired.it820 / 350K Unique Visitors

2.9M Page Views

30,629 Unique visitors

65,000 Page views

Direct measurment: Nielsen / Audiweb

Data refers to worldwide coverage for June 2012

Great Tools Still Require Great Analysts

No competitive analytics tool tells the entire story. Great insights require great people. A social media analyst should know there is a major flaw in the following Google Doubleclick Ad Planner view of Twitter statistics: statistics from Google's Doubleclick Ad Planner
Figure 2: statistics from Google’s Doubleclick Ad Planner

The problem? Namely most of twitter usage is via twitter’s API and third party apps (many of which twitter now owns, but that is another story). At the Chirp developer’s conference in 2010, Twitter stated that only 25% of their traffic was reflected on A great analyst will know this. Tools won’t.

Clouds On The Horizon For Google’s Ad Planner

Google's August 22, 2012 email to AdWords advertisers on the Future of Ad Planner
Figure 3: From Google, with love (click to enlarge)

In an August 22, 2012 email to AdWords advertisers, Google outlined several planned changes to Ad Planner. Google says that as of September 5th, 2012,

  • data will only be available for sites carrying Google advertising, known as the Google Display Network
  • less demographic data will be available, i.e. Keywords Searched For, Videos Also Watched, HouseHold Income and Education will be removed
  • Unspecified “adjustments” to a few of the traffic columns like Unique Users and Reach, removal of Page Views
  • Publisher Center, the feature that allows publishers to claim their own sites, list relevant ad placement availability & pricing, will be deprecated, which I understand to mean removed
  • The tool name will become Google Display Network Ad Planner

One view might be that Google giveth and Google taketh away. While I understand that Larry’s chat with Steve has led to lots of product pruning, some of it very understandable, this is one of those cases that I find rather unfortunate (Google Browser Size is another favorite tool headed for the dustbin although limited functionality has been ported to Google Analytics). The new restrictions placed on Ad Planner do seem to fly in the face of Google’s stated mission: to organize the world’s information and make it universally accessible and useful.If you agree, let your Google contacts know how you feel – maybe someone with the appropriate authority will rethink the current plans. The continued availability of Google Trends for Websites should somewhat mitigate the planned reductions in Ad Planner data availability. Unless Google decides to remove it too.

Similar Posts:

Registration is now open for the next SEO Course and Google Analytics Course in Milan. Don’t miss the opportunity!

About Sean Carlos

Sean Carlos is a digital marketing consultant & teacher, assisting companies with their Search (SEO + SEA = SEM), Social Media & Digital Media Analytics strategies. Sean first worked with text indexing in 1990 in a project for the Los Angeles County Museum of Art. Since then he worked for Hewlett-Packard Consulting and later as IT Manager of a real estate website before founding Antezeta in 2006. Sean is an official instructor of the Digital Analytics Association and collaborates with the Bocconi University. He is Chairman of the SMX Search and Social Media Conference, 12 & 13 November in Milan. He is also a co-author of the Treccani encyclopedic dictionary of computer science, ICT & digital media. Born in Providence, RI, USA, Sean received Honors in Physics from Bates College, Maine. He speaks English, Italian and German.

3 Responses to "Can Google Be Beat On Competitive Web Intelligence Data?"