Two weeks ago I read Simon Holland’s piece entitled “’Flawed’ Last Click Attribution Dominates Display Retargeting”. We can all agree last click attribution is flawed but I was surprised to discover that in the programmatic ad space finding a ‘proper attribution model’ is considered a bigger concern than the lack of transparency.

This surprised me because transparency and insight are key to moving to any new attribution model, including a multi-touch attribution model that would aim to assign credit where credit is due to all touch-points across a conversion path. To do this you need to have a deep understanding of the roles of each of your vendors and the components that contribute to the delivery of traffic and sales.

As an example, with retargeting, you have publisher traffic sources across mobile, tablet and desktop, variables such as media prices, dynamic ad designs, creative messaging and product selections – as well as a host of other attributes. Understanding the actual contribution of all these things is no easy task! Mapping them into an attribution model without first understanding what the key influencing factors really are will lead to a broken model and potentially lower performance.

Test, learn, take action and test again

In order to learn what the key factors are in driving performance and understand your vendors role you need to have strategic conversations with them, this is where the necessity for transparency across the space becomes imperative.

At Struq, we use a transparent cost plus business model with all our clients. This involves charging a fixed percentage on top of ad spend so we are able to advertise on the best available media to find the best users, whilst achieving the client’s ROI goal. Algorithms and tests continuously run to verify the best combinations of ad design, creative messaging and products. What is important is that we present data to the client meaning they can see the publishers contributing to success, the impact of cross-device retargeting on different devices and operating systems , the most effective creative messaging, and much more. The client gets the complete picture and understands the factors that contribute to success.


So, rather than analysing attribution models, what action can the client take with this approach? Take mobile where, although some clients now see 50% of traffic to their mobile website, they do not see similar sales performance compared to their desktop website. Through cross-device retargeting we can show the effect that the role of browsing on your mobile site has on subsequent laptop conversions. This is an actionable insight because now the client can control investment in areas that are proven to drive performance. Easily missed if your attribution model isn’t considering a multi-device world.


In his article Simon cited a lack of confidence as the reason advertisers have not moved to a new attribution model. This is perfectly correct, what was not said is where we can gain this confidence from. Those insights and strategic conversations that so easily stem from a transparent relationship are the perfect starting point. By showcasing the value of the components that work together to drive performance, through statistically based testing and clear reporting, providers can start to supply their clients with the knowledge necessary to begin solving the attribution puzzle.

It is easy to see why the last click model has been the de facto standard. If nothing else, in the past it’s not been easy to understand the effect of the advertising impressions and click-throughs that happen in the middle of the conversion path (and that’s before this multi-device world emerged!). However we now know through testing, reporting and working closely with clients that these not only make a difference but also how much of a difference they make. Forward thinking advertisers are rightly considering how to solve the challenge of attribution within the programmatic ad space. The benefits of transparency are all there for the taking, first we need to lift the lid off performance