Data-driven attribution – can we finally put ‘last click’ to bed?
Attribution models are tough for most people to get their heads around – and the level of detail available to marketers for analysis is peculiar to the digital age, but will Google’s latest attribution push help the uninitiated choose the right model?
Forgive me an extended metaphor, but I promise to keep it reasonably short.
You own a shop at which consumers will potentially arrive in many ways – they may take a taxi, then a bus, then a train followed by a short walk.
Another consumer, however, may arrive solely following a journey via any one of these methods, while still another may use any other combination of them before they eventually arrive at the shop and make a purchase.
In this situation, it would be natural (and it is common) to see the final mode of transport as the most important and, therefore, miss the importance of all the others. While this may not cause trouble for some, if you don’t see the bus delivering any visits to the store it can seem reasonable to assume that the bus journey is not important to your business.
What if, however – and unknown to you – the bus plays an integral role in 3 out of every 5 visits to the shop? While you may see no direct results, visits from the passengers on the busses could theoretically reduce sales by 60%. It is this that is the potential downfall of last click attribution (and, to lesser extents, other standard attribution varieties which suffer the same oversight to greater or lesser degrees).
Data-driven attribution is different from the other attribution models, in that it uses your conversion data to calculate the actual contribution of each keyword across the conversion path. Each data-driven model is specific to each advertiser.
Google AdWords Help
While the new data-driven attribution model requires a minimum of 15,000 clicks and at least 600 conversions within 30 days to be used with an account, it promises to offer the most precise attribution model available so far (though only on google.com search ads, and only for Google Analytics conversion actions).
Obviously this level of insight into the conversion path of your consumers offers a tremendous opportunity for increasing ROI and optimising campaigns to ensure the best possible performance and, while some agencies have run their own modelling exercises (see the below slide from a Benchmark Conference talk by Click’s Head of Paid Search Dave Karellen), this feature will offer some new opportunities for those agencies and in-house PPC teams that have not previously had them available.
From Benchmark 2016 talk ‘Maximise Returns with Mathematical Modelling
While clearly there is no substitute for experience and data and Google’s modelling methods can only improve understanding so much, there is no doubt a an opportunity for improvement in attribution with this new option and, while it may not be the final nail in the coffin of last-click attribution, it makes the alternatives far more accessible.