Call 0845 485 1219
We love digital - Call and say hello - Mon - Fri, 9am - 5.30pm
by Stephen Logan on 28th July 2010
Yahoo suffer artificial loss of traffic in comScore’s June statistics due to a ‘processing error’ from the online analytics firm.
If ever you wanted proof that all online statistics need to be taken with a pinch of salt, comScore have announced that in their latest search engine traffic figures Yahoo had over a billion page views wiped off their monthly tally. Usually one of the most reliable analysts of online data, this slip shows fallibility in almost all statistics gathering.
Yahoo were understandably a little miffed by this sudden drop in their supposed traffic levels. With online commentators reporting the results verbatim, the trailing off of their traffic and duration spent on the site across all markets created some unwanted and, as it turned out, unfounded headlines.
comScore put it down to a ‘processing error’, which of course is possible in any online analytics. Whilst a 3% drop-off in visitors might not be a major issue for a low traffic site, for Yahoo those 1.1 billion page views made a rather sizeable dent in what was already a pretty woeful month in the U.S. for them (see chart below for amended statistics).
So rather than the actual 4.7% drop in traffic, comScore had instead suggested that it was nearer 7.4% – oh dear. Of course this has all been cleared up now with some carefully worded press releases from both parties [see TechCrunch | Yahoo: comScore Underreported Our U.S. Page Views by 1 Billion Last June], but the damage may already have been done.
In a competitive industry like the search engine market, analysts pour over statistics and any major change is reported instantly. So whilst comScore have admitted their mistake, Yahoo will have to pick up the pieces of any loss of market confidence and other associated issues.
Needless to say, Yahoo’s problems weren’t confined to the U.S. It turned out that both the duration of visits and page views were underreported in markets right across the world. A big slap on the wrist then for comScore.
What’s the lesson in all of this? Well, don’t assume that all statistics given are 100% accurate. There can be shifts in the reported data caused by any number of reasons. Whilst comScore have admitted the problem was within their own calculations, programmes can be affected by all kinds of issues.
Take Google Analytics for instance. There are plenty of ways that visitors can be hidden from the software’s gaze. Users are afforded the opportunity to opt-out of having their movements tracked when logged into Google for instance. This shouldn’t have a massive impact, but it does mean that reporting can sometimes be skewed.
It isn’t the end of the world though and Yahoo will recover from this blip. comScore are still a highly reputable source of online data, along with others like Experian Hitwise, but this shows how a small error can have a significant impact on statistical accuracy – particularly if it isn’t picked up quickly.