Mobile

Ground Truth debuts highly accurate way to measure traffic on mobile sites

GroundTruthA new mobile measurement firm, Ground Truth, today announced it has launched the first such service to use census-based measures of actual Mobile Internet usage. The data, which covers millions of mobile subscribers in the United States, will come directly from mobile operators and other data providers, allowing the company to aggregate figures on mobile data usage on any visited mobile site.

As a result, Ground Truth data will likely show a more realistic picture of the mobile web than the survey methods of its competitors, who include ComScore and Nielsen. Ground Truth’s launch may also be very good timing, given that ComScore has recently drawn some ire for under-counting traffic numbers and then offering more accurate counts for a steep fee, a move entrepreneur/blogger Jason Calacanis called “extortion.

A data set Ground Truth released suggests that mobile-only web properties have not been run over by traditional web companies as has been suggested by reports on the issue in recent months. According to the data, traditional mobile properties, such as mobile-only social networks like MocoSpace, AirG and mobile-web centric companies like Cellufun or Myxer continue to be among the most visited, despite reports to the contrary within last 12 months. That means mobile web companies will have stronger traffic numbers to pitch to advertisers and could attract more advertisers to mobile.

Also, according to the data, MySpace has more traction on mobile in the US than Facebook. So Ground Truth’s numbers vastly contradict the numbers we’ve all been relying on heavily to date.

Ground Truth ranking of Top 10* mobile web sites for the first week of 2010 (January 4-10, 2010)
Ranked by total page views
1. MySpace
2. Facebook
3. Google
4. Mocospace
5. FunForMobile
6. AirG
7. Yahoo
8. Cellufun
9. Mbuzzy
10. Myxer

*Not including pages served by operator portals

In a call with VentureBeat, CTO Michael “Luni” Libes and CEO Stirling Wilson also revealed another interesting tidibit on the size of the mobile web. They say 97% of mobile web traffic in the US happens on around 12,000 websites.

Accurate measurement of traffic to mobile sites has been very difficult until now. Frequently, when I double-check the traction mobile companies report, I find other data suggesting their numbers are off by factors of 2 to 3. It’s rare that companies manipulate the data on purpose, they just rely on bad estimates given to them by research houses.

I remember asking the mobile VP of a larger research house on the ratio between operator (“on”-) portal and off-portal traffic just days after the company released a report on the growth of the mobile web. The VP was not able to give me any specifics.

As the data reported is so bad, rankings like the mobile advertising one I cited last week are frequently contested by the companies listed in them. And people like me who watch the market have learned to rely on the work of very few analysts over the last few years or have relied on their own calculations.

Ground Truth improves the quality of the data we can access on the mobile web in three main ways. For one, the company is reporting on actual data, not consumer surveys. That makes the data more factual. Surveys depend on users remembering which properties they have been accessing. As the desktop web is still dominant, many users have the tendency to over-report web properties.

Secondly, the company reports recent data. So far much of the data from surveys has been three months old by the time it got published. That meant, for example, that a data company could produce a December 2009 mobile web report, based on data from October 2009.

Finally, Ground Truth uses a very large sample size. It includes data from 2.5 million subscribers across all networks, while the largest sample size I recall from a competitor is 30,000.

Competitors will likely argue that the demographics of Ground Truth’s sample size may not be representative of the whole US. For example, if the sample data comes mainly from Texas, it would be hard to argue that the conclusions drawn on that sample hold true for the whole of the US.

I asked CTO Libes about that, he admitted the sample size does not copy the geography of the US. However, he believes the sample is “broad, varied” and includes “mixed geographies and various operators,” although under the terms of the deals with its partners, the company is not able to disclose which operators it has partnerships with.

The key to Ground Truth’s success will be that the data has always been there with the telecommunications infrastructure companies, it was just not reported on so far. Since the company received its $2,6m funding from Steamboat Ventures and Voyager Capital in June 2009, the founders have had “many drinks and dinners with telco executives” to give the company anonymous access to the data. Telecommunication companies have a lot of information on subscribers so far, but they have used little of it to help their own business. They were also not eager to share the data with their competitors. By positioning itself as a trusted third party, Ground Truth is able to access some of it.

The Seattle-based company is based on the traditional market search model, it expects to secure paid annual subscriptions from its clients. It hopes to have a superior product for a wide range of case studies for publishers and advertisers, including benchmarking or detailed ad network reporting.


Mobile developer or publisher? VentureBeat is studying mobile marketing automation. Fill out our 5-minute survey, and we'll share the data with you.