For two years Facebook has been reporting an inflated figure for the amount of time people spend watching videos on their site. Average time spent watching videos was being overestimated by 60-80%. This is more jaw-dropping when you hear they will make over £1bn this year from advertising when their biggest push has been for video.
This isn’t the first fly in the online ointment. The last 12 months have seen revelation after revelation and advertisers like Procter & Gamble and McDonald’s have been making U-turns.
No wonder that Yahoo’s lovely Nigel Clarkson recently bemoaned the lack of time spent celebrating the good things about online advertising. The IAB too came out asking for people to give it a break and remember the good points.
Of course there are plenty of good points, not least online’s magical, symbiotic relationship with TV. But most of the bad points point towards one thing, and I’m afraid it is quite a dull thing – a thing that gets banged on about too much. The thing you’re dreading me saying. Yes, it’s data.
An important part of the fallout from Facebook’s error has been an increasing clamour for impartial, audited data, not the secretive, walled-garden approach that some online giants have taken so far.
With perfect timing, another example of why we need data from glass boxes, not black boxes, presented itself this week.
Google claimed that YouTube generates twice as many Google and YouTube searches per impression as TV.
Leaving aside Google’s now pathological desire to aim at TV (this is the latest in a succession of disputable or meaningless claims comparing YouTube with TV), what is this claim based on?
I haven’t seen a detailed methodology, but it sounds like it is based solely on Google’s analysis of its own data. They are marking their own homework, as Sir Martin Sorrell has put it.
Everyone should be sceptical about every piece of media research, but especially when it isn’t impartial and we are left partially sighted.
As well as the origin of the data [new book idea: On the Origin of Specious], Google’s latest claim misses the wider context. We know that their attribution systems tend to over-credit the last click so as a result they under-credit what went in to making it happen.
But advertisers need to know about every media exposure that made that final click take place. Looking at searches per impression is an interesting but crude measure. What matters is creating profit and building a brand in the long term, not short-term clicks.
But I digress from data.
The point is that if Facebook and Google gave the market impartial, audited data, at a stroke they could shut down many of the criticisms currently being levelled at them – and find themselves wearing snugly fitting, reliable pants.