6 min read Data Attribution Models: Get the Full Story

A lot of people switch off when I start talking about data attribution, it’s the gripe that no one wants to talk about but one of the most important elements to understand if you are really going to make strides in your performance marketing efforts. Making decisions based on data is the tagline for a lot of marketing professionals, but if you don’t understand your data from every angle, you’re doing yourself and even worse, clients a disservice.

#1 What is Data Attribution?

Attribution models in performance marketing are there to help figure out which marketing efforts & channels actually lead to a conversion—whether it’s a sale, a sign-up, or any other goal (or key event, shout out GA4) . Since customer journeys can often be a lot longer and convoluted than the ‘ideal’ scenario, attribution models help track their journey across different touch points.

#2 Different types of attribution models:

A first-click model rewards the first interaction.

A last-click model rewards the last interaction.

A linear model spreads attribution evenly across all touch-points in the journey.

A Time decay model spreads attribution across all touch-points but rewards the points later in the journey with more value. 

A U-shaped model spreads attribution across all touch-points but rewards the first and last touching point with 40% value, spreading the remainder across the other points, evenly.

A Data Driven Model (GA4 Default) spreads credit across touch-points based on how users interact. Giving more credit to touch-points with higher engagement. This model also uses keyword data to provide an understanding of what really drives a conversion. 

#3 Why is it important?

I’ve gone a bit overboard there with the models. In the real world generally speaking businesses use one of either first click, last click or data-driven but selecting an attribution model based on your business needs and understanding of your unique customer journey is crucial. For example: If you only look at last-click, you might ignore the role of brand awareness or if you credit the first touch too much, you might miss the importance of retargeting.

Ultimately it makes your life as a marketer much easier if you have a handle on attribution:

ROI: Optimise your marketing spend via understanding which efforts drive the most value. Particularly important when it comes to annual re-budgeting – this helps improve your ROI as a team and department by knowing where to spend the money and get the most back.

Customer Journey Improvements & Personalisation: If you know how your users and potential customers are arriving at your website, you can improve these touch points to better cater for that stage of the journey, even altering messaging and testing different content to appeal to customer behaviour.

Competitive advantage: You may not be doing it, but the competition may be, and in an ever growing market of digitalisation, any upper hand you can get you should take.

Decision Making: Ever been in a situation where you’re scrambling to prove the point or ROI on a campaign or piece of content? Something that isn’t necessarily commercially driven – well attribution could be your saving grace.

#4 Data adds justification

What separates a successful year from a not so successful year is the decisions you make along the way, budgeting for the following year is all well and good but if you don’t routinely revisit those budgets you’re likely going to end up being disappointed come year end. Attribution allows you to have a full understanding of what is working and what isn’t working allowing you to ramp up and down where necessary.

The most common question we get asked is around the justification of informational content or brand awareness activities – generally these are harder tactics to justify from an ROI perspective. When was the last time you read a blog on a website and then instantly converted or purchased a product? Probably never. Well, having different attribution models allows you to justify those less commercially focused channels and validate why they are part of the overall strategy.

#5 A real world example:

This was back in 2022(ish) when GA4 was being thrust upon everyone and accounts automatically switched from last non-direct click to data-driven attribution, handily without telling anyone so people were adjusting to the new normal.

A business approached us in the beauty industry who came to us for paid media management, they were semi well known and had a loyal customer base with a solid offering. Customers often turned into advocates, their main challenge was getting them through the door as they knew after that they would stay and the LTV was significant.

We set about our business as usual, developing a strategy, auditing the existing accounts & implementing the growth strategy while managing, monitoring and tweaking as we go to reach optimal account management.

Fast forward a couple of months and we’re really happy with the results, we could see ROAS had improved by over 350% in a short period of time, we hop on a call ready to discuss the performance of the account to be greeted with a screen shot from their internal reporting which was showing a very different story, we weren’t in fact 350% up we were actually flat and as far as they could see it wasn’t profitable to continue with paid media – I pulled up our reporting and showed them our 350% improvement in performance in direct contrast to their report.

We discovered over the course of the call that they were using a last click attribution model which in fact missed off over 60% of the paid media value. In short, due to a high price point, users were clicking through via a paid ad – but often then leaving to consider the purchase and returning back to the site, often through direct or other means.

A bit more digging and there was a huge spike in direct revenue in the period between us starting our work and this conversation, while there was a large increase in new users from paid search the revenue hadn’t increased in line. Using first click attribution modelling the story switched on its head and it was clear that paid had a significant impact on performance in the past couple of months.

That is 2 different attribution models showing 2 completely different stories, without looking at multiple models you could make a decision that significantly impacts performance without even knowing it.