By Sean Burton, Senior Consultant, Foviance, a customer experience and digital research consultancy.
The way web audiences are measured came under the spotlight last week with the publication of a report by net measurement firm comScore, which claimed that cookies used to track user behaviour could be being over-counted.
The results caused a stir in the web analytics industry, and triggered concern in the marketing world about whether the validity of web analytics data can be trusted.
At Foviance we were less surprised by the findings – ‘cookie churn’ has been an issue that we and other analytics experts have been aware of for a number of years.
The root cause can be traced back to users who either delete cookies from their machines or log onto a site through a number of different machines.
In that instance, unless you are able to use registration data, then your analytics tool will likely count each visit as a different user. According to the comScore report, users can be over-counted by as many as 20 times.
But does that really matter?
The people who will be affected by these findings the most are those in online advertising. If content owners have fewer eyeballs, then the cost of advertising will have to drop as the market adjusts.
But the loss of revenue will only be temporary. Clients will be able to measure the value and effectiveness of their activity more accurately, while greater transparency, honesty and accuracy can only help the industry grow.
One interesting thing within the report was the fact that first-party cookies are deleted with the same frequency as third-party cookies. This suggests that users aren’t taking the time to separate the cookies of their favourite sites from unknown cookies.
This is an issue for the industry as a whole, rather than simply analytics vendors.
If valuable site personalisation isn’t taking place because of a lack of cookies, then the customer experience is affected.
It would indicate that there is further work to be done with the browser and virus companies to help users distinguish between them.
Ignore long-term data
The report found that the cookie churn had a greater impact over time, and our own research has shown that visitor figures can be over-estimated by as much as 300 per cent over a year
This is likely to vary from site to site, and what this really highlights is the need for robust interpretation of the data, rather than simply accepting the data at face value.
However, what requires further study is whether the figures are inflated exponentially or linearly over time. If it’s linear, then allowances can be made. If it’s exponential, then quarterly and yearly analysis of user numbers become increasingly meaningless.
The key lesson from all of this is not to throw the baby out with the bathwater.
The fact that a base-line statistic is frequently inflated does not make analytics any less essential to a business. It just makes you evaluate how you use your software, and the level of information it can provide.
The report acts as a reminder of the reason we bought analytics packages in the first place – not to be the sole measure of success but to help us spot trends, answer questions and provoke further thought.
Check out 12ahead, our brand new platform
covering the latest in cutting-edge digital marketing and creative technology from around the globe.
12ahead identifies emerging trends and helps
you to understand how they can apply to modern-day companies.
We believe 12ahead can put you and your
business 12 months ahead of the competition. Sign up for a free trial today.