Never mind ‘cookiepocalypse’ – the phasing out of third party cookies in Chrome. This is ‘GArmageddon’ – the impact of GDPR consent on Google Analytics (GA), in which website traffic, perfectly healthy one day, appears to fall off a cliff the next. We knew it was coming but publishers were still taken aback by falls of up to 30 or 40%.
“Half the money I spend on advertising is wasted, the trouble is I don’t know which half,” is a century-old adage from a department store owner. Website analytics seemed to change all that but it was never quite as solid as we liked to think. Anyone who’s looked under the GA hood will start to see numbers that don’t seem to add up. Why is the average time on page that appears in one area of GA so vastly different to what ought to be the same figure that appears in another area? How do we know which is right – or do we just pick the one that looks better?
Why and how on earth do so many fly-bys visit for zero seconds? In reality, they don’t but GA, depending on how you have it set up, may count the last page someone was on, even if they read the article, as 0s. We still quote unique users yet a few years ago, publishers saw numbers mysteriously float upwards and stay there. A cause for celebration? No, it was just GDPR rules taking effect, users refusing tracking cookies, thereby turning one user into several users overnight. GA doesn’t even call it unique users ever more.
Measuring paid search is straightforward enough though, isn’t it? You pay your money, people see the paid for result, click on it, land on the site and then we can see how many are converted to paying customers. That’s what eBay long thought too, as it was investing an awful lot on paid search but getting what looked like a great return from all those converted customers. Then it decided to run an experiment, reported in Harvard Business Review at the time. First it turned off its brand name advertising and… nothing happened. People still kept coming because those that used to click the paid for result just clicked the unpaid one instead. Then eBay started turning off other paid for search terms in some regions and comparing the results. The conclusion was that paid search on brand was worthless and elsewhere it was just not worth the cost.
There are some caveats. Your situation may be quite different if you’re a start-up for example, rather than the market leader. And plenty of marketers say paid search could work for eBay if they just implemented it in the right way – in fact, it has crept back in a more selective way. But the point is that we need to be more sceptical about success stories in data that appear self-evident. Especially, from those who really, really want to believe. And the analytics hasn’t been quite as good as their advocates made it out to be – especially the basic way in which many sites implemented GA.
Now it’s all gone really awry with publishers seeing all their GA metrics suddenly plumet. Our own site has been hit hard – partly because it’s technology news and technologists tend to be very concerned about privacy. We know they are thanks to the numbers of visitors to our online privacy coverage we get – or used to.
Anyone who’s looked under the GA hood will start to see numbers that don’t seem to add up.
What can be done?
What can publishers do about it apart from panic? There’s no silver bullet out there but some mitigations worth exploring. Other analytics packages may go further than GA to allow you to sell a more targeted advertising proposition, but they still require cookies or other tracking techniques and therefore consent.
You can tweak the wording, design or defaults in the consent form. The worst of these try to fool readers into giving away their data. We’ve all seen those online and it might sometimes work for some audiences but not ours nor most other publications’. You risk alienating readers. I believe that transparency pays.
Moreover, tweaks won’t bring back those missing millions. Google has delayed its cookiepocalypse but trust and truth is a growing issue and the need for consent is here to stay. We’ll look back in decades to come and see the period we’ve just been through as a strange time when most people gave away everything about themselves just to get more cat memes.
GA can still give us a rough idea of growth, or insights into which articles are more popular than others. So, it will remain but not as the headline sell it once was. It may seem like a small consolation but all those missing visitors aren’t really lost – most of them are still there, if you have something worth reading, viewing or hearing. You just can’t see them.
It reminds me of… print! The readership is hard to measure but we get a sense through stand sales, surveys or even just the mail bag. Just as we reinvented our print products, now it’s time to reinvent online models: look again at logins, memberships, subscriptions and all the other possibilities. It’s not the end of the world.
Just as we reinvented our print products, now it’s time to reinvent online models.
This article was first published in InPublishing magazine. If you would like to be added to the free mailing list, please register here.
