Is last-click finally on its last legs?

2nd March 2018 - 4 minutes read  Data & Analytics
James Wolman - Data Scientist

Google’s closing the book on last-click with a new attribution model to rule them all.

In May 2017, Google announced to its partners that the last-click attribution model was officially over and now 2018 looks to be the year that it makes good on its promise.

It’s tough news for a model that’s been marketers’ go-to for years, and many still agree that last-click offers a satisfying picture of where your conversions are coming from. Be it a digital ad, an email campaign or a piece of SEO content, buyers will have been in contact with at least one of a brand’s outputs before they make a purchase, so it’s useful to have credit allocated to that touchpoint – it paints a clear portrait of which channel is performing best.

It might not be the most sophisticated or accurate way of assigning credit, but it’s simple, easy to use and works well with most kinds of campaigns.

Other rules-based models are available too, including Linear, Time Decay and Position-based, which improve on the sophistication to varying degrees, and are varyingly popular. And although many of them incorporate multi-touchpoints, none of them really stand up next to the rather more glamorous Google Attribution.

We all know that the average transaction these days has multiple touchpoints – possibly as many as 30. Buyers customarily enter multiple searches, visit several to a dozen websites and browse reviews and feedback before making a purchase. That means ever increasing numbers of touchpoints before the conversion.

This being the case, the data last-click can provide is just not detailed enough for marketers to reliably analyse their ROI and re-allocate resources to the best performing channels. If it seems wasteful to be spending on a campaign with a poor ROI ratio, you need to know which campaign it would be prudent to invest in as an alternative.

This is where machine learning arrives on the scene. Google’s product uses machine learning to assign a weighted value to every to every single touchpoint along a consumer’s path to buying.

Senior Director of Product Management for Analytics Management at Google explained how it works and why it’s such a game changer: “It creates a prediction model that learns by weighting a set of touchpoints on how likely a user is to purchase something. The presence and absence of marketing touchpoints across channels and across campaigns will either decrease or increase the likelihood of a conversion.”

The model considers:

  • Ad creative
  • Number of ad interactions
  • Order of exposure
  • Other variables to determine which clicks and keywords are the most effective at delivering results.

Pretty thorough, then. Crucially, Google Attribution doesn’t just assign attribution accurately to all the touchpoints in play. It also identifies patterns for marketers, comparing the click paths of customers who convert to the click paths of customers who don’t, and in so doing examining a wide array of search behaviour.

It’s meant to provide a much faster service than existing models and, enticingly, it’s also free in the version designed for small and medium sized businesses.

So loved you as we have, our old friend last-click, it seems to be time to say goodbye. We hope Google Attribution is about to usher in a newly efficient golden age, providing unprecedented levels of accuracy and the chance to optimise our campaigns in a big, bad way.

If you want to talk more about why last click probably shouldn’t be part of your marketing plan anymore, give us a shout and we’ll talk you through it.