In December (2016) Facebook announced yet more miscalculations with its ad metrics – for the fourth time since September. The social giant is creating all kinds of headlines about inaccurate metrics right now but it’s not the only tech firm experiencing problems.
Ad metrics are a tricky business, even if the likes of Facebook genuinely try to be accurate and transparent about things. And now Twitter has had to come out and admit it has been overestimating its video ad metrics for the end of 2016.
What’s going on with all these metrics blunders?
Twitter says its recent miscalculations were due to a bug in its Android app. The network was quick to reveal the problem, which overestimated video metrics by up to 35% between November 7 and December 12. Customers have reportedly been refunded for any overcharges during that period but the talking point remains.
This probably wouldn’t be making the news in the same way if it wasn’t for Facebook’s recent problems with metrics. The reliability of metrics is a big concern for advertisers where spend, performance and ROI are so dependent on those vital numbers.
However, this isn’t a new concern by any means. Any seasoned AdWords advertiser will admit to having questioned the accuracy of Google metrics. In fact, it’s only as far back as July since we were complaining about its keywords tool becoming less accurate.
Facebook and Twitter’s problems are a different thing altogether though. These aren’t feature changes we’re talking about here but actual miscalculations – bugs in the system. They’ve both been very open about their troubles but the media response has been pretty big.
So why are advertisers less worried than the journalists? Well, first of all, less than 1% of publishers were actually affected by Facebook’s miscalculations but there are a number of other reasons, too.
Metrics taste better with a pinch of salt
In the advertising game it’s pretty easy to take a long list of metrics to the boardroom and make it sound like good things are happening. ‘X’ amount of video views, clicks, comments and other engagement stats sound great but the only one that really matters is profit. Not turnover, not ROI and certainly not the number of people who saw your video.
That’s not to say these numbers aren’t important or that miscalculations don’t matter but advertisers know they have to work with this potential for error. The metrics are a means to gauge performance and improve the only number that matters – profit.
Metrics are also something a flawed concept to begin with. It’s still difficult to separate first-time visitors from return visitors or even know which traffic is actually human (roughly half of it it bots). Things are more complicated than ever due to people moving between multiple devices too. And this is before you even consider click fraud and other factors that can drastically influence metrics.
Advertisers know they have to deal this is room for inaccurate figures. We focus on the things you can measure (eg: spend, return and profit), using metrics to give us clues about what the right next step will be.
So it’s all a bunch of fuss about nothing?
Well, no – let’s not play it down. Any issue with metrics is significant but most advertisers will be glad to see how quickly Facebook and Twitter have been to admit their mistakes. You have to wonder if Google has/would be as honest as those two about miscalculations – and we’ve all had our doubts over the years.
The fact is, you shouldn’t rely only on the numbers provided by the platform you’re funding – that’s just common sense. But, even when using third-party analytics platforms, it’s important to understand the limitations of metrics to begin with.
Metrics are getting even more difficult to calculate
We’ve already mentioned a number of reason why it’s hard to accurately measure metrics but the worrying thing is it’s getting even more difficult. Take the introduction of AMP and Instant Articles for Google and Facebook. These users click to see your content but you don’t actually get the traffic – so what kind of click-through does that even count as?
Or how about predictive/personalised search in Google – are users really searching for what they want or whichever query Google shoves in front of their faces?
The big challenge for the tech giants is keeping up with their own developments. Things are moving so fast now it’s hard to keep track of individual user movements between different channels, devices and the online/offline divide.
The days of search engines and websites being the only path to track are long behind us and this complex journey is making analytics even more challenging.