The death of last click attribution

4 min read

The average person uses three devices on a daily basis, browsing a multitude of different websites. While this has increased the opportunities available to marketers, it has made the analysing of marketing campaigns somewhat tricky. Google offers a helping hand with its attribution models, but none have been up to scratch until now – but more on that later.

Attribution plays a critical role as it credits the campaign which led to the conversion, but the issue nowadays is that there are often multiple aspects of numerous campaigns that play a role in converting a user. We’ve taken a look at current attribution models, those that are dying, and a potential holy grail for marketers.

Understanding Google’s attribution

To understand Google’s attribution model, we’ve used an example from their Attribution Modeling page.

“A customer finds your site by clicking one of your AdWords ads. She returns one week later by clicking over from a social network. That same day, she comes back a third time via one of your email campaigns, and a few hours later, she returns directly and makes a purchase.”

Last click – The final ad which the user clicked-through before purchasing. In the above case, the direct channel would receive all the credit.

First click – The first ad which the user clicked through before purchasing. In this case, Paid Search would receive the attribution

Linear attribution – Equally distributes to each touchpoint within the conversion path. All channels involved above would receive 25% of the credit.

Time Decay attribution – The touchpoint closest to the time of conversion gets the majority of the credit. In this example, the Direct and Email channels would get most credit given their activity just hours before the conversion. The Social Network would receive less, with the Paid Search interaction, which happened a week earlier, receive the least amount of credit.

Position based attribution – This assigns 40% of the credit to both the first and the last interaction, with the remaining 20% being distributed evenly to all other touchpoints involved.

The majority of these all have a role to play, but the issue that they don’t correctly assign credit is still very much present. Meanwhile, last-click attribution is barely keeping its head above water, exposing itself for its lack of clarity for analysis of campaigns.

RIP last-click attribution

The issue with last click attribution is it’s crediting the touchpoint which has done the least amount of work on the entire conversion journey. If we use the example from above, only the direct visit to the website will receive credit for this whole interaction. The irony behind this is that the last click is usually the smallest contributor to the entire campaign.

It’s imperative for marketers to have full sight of a campaign, as well as the contributing factors to conversions. Given the advertising platforms available to marketers nowadays, it’s inconceivable to credit just one channel for the conversion of a user. Luckily for said Marketers, Google has recently released a new model that fixes this attribution headache.

Data-driven attribution

In the late summer of 2016, Google released a new attribution model known as ‘Data-driven attribution’. They understood the likes of last-click attribution all had their biases and wanted to create a more dynamic attribution tool that would credit accordingly.

Data-driven attribution uses machine learning to understand which ads, keywords and campaigns are best performing towards converting users. The below example is from Google’s definition of data-driven attribution.

You own a tour company in New York City, and you use conversion tracking to track when customers purchase tickets on your website. In particular, you have one conversion action to track purchases of a bike tour in Brighton. Customers often click a few of your ads before deciding to purchase a ticket.

Your data-driven attribution model finds that customers who click your “Bike tour East Sussex” ad first, and then later click “Bike tour Brighton seafront”, are more likely to purchase a ticket than users who only click on “Bike tour Brighton seafront”. So the model redistributes credit in favour of the “Bike tour East Sussex” ad and its associated keywords, ad groups and campaigns.

Now, when you look at your reports, you have complete information about which ads are most valuable to your business.

Given the fact this attribution model makes use of machine learning, there are some specifications required to take advantage of it. An account must have over 15,000 clicks and 600 conversions within a 30 day period – upon meeting these conditions, data-driven attribution will become available.

Data-driven attribution for smaller marketers

If your business only does a few hundred sales per month, is there any way to leverage data-driven attribution? Fortunately, the answer is ‘yes’.

If you analyse the funnel and your customer decision journey, you’ll notice that there are critical stages where the potential customer has shortlisted your product for consideration, which can be any one of the following:

  • Visited some key pages on your website
  • Expressed interest by signing up for a newsletter
  • Started a check-out process

If you map any of those to a conversion, higher up in the funnel, more people would be more likely to trigger it.

Google’s data-driven model will then help you optimise for that stage in the journey, which will primarily assist you in sending more qualified people towards your final conversion.

Google and Facebook considerably support the lives of Marketers, and this continues to be the case as they get better at leveraging applied Artificial Intelligence. With there already being an array of tools assisting the work of marketers, who knows what the future holds.