+44 (0)20 8744 4350

Attribution: finding your way through to the answers



It’s seemingly impossible to create a model that acts as a ‘one size fits all’ and is objectively the best method to measure ROI, leaving many marketers scratching their heads on how to proceed. This holy grail of accurately assigning credit to all the touchpoints in the conversion path has proven to be elusive for a while with many marketers choosing to adopt the last-click attribution model, famed for its ability to make us PPC workers look awesome. The problem with this is that it does not show the full picture.


Sure, you might know what landing page a customer converted on ... but what about all the social posts, content, and site pages that also influenced their decision to buy? Are these not valuable and significant interactions that deserve recognition? And how do you decide whether the first touch (like clicking a Facebook ad) is more important than the webinar that generated the sign-up?


The answer is generally a lot of shoulder shrugging and confused faces.


But you wouldn’t be alone. A recent study by HubSpot found that 40% of marketers stated that proving the ROI of their campaigns and online activities is their company’s top marketing challenge. No wonder this is such a concern for businesses due to the fact that with no dependable ROI measurements, clients are less likely to want to increase their budgets. The two go hand in hand.


Luckily, there are a handful of different attribution models out there to choose from, and depending on the product you sell and the length of your buying cycle, one might make more sense than the other.


Let’s go over a few of the most common types:


The Last interaction model attributes 100% of the value to the last channel before a sale or conversion. It’s a great baseline for comparison with other models and is the popular pick amongst marketers.


The First Interaction model can help you understand which campaigns create initial awareness for your brand or product.


The Time Decay model assigns the most credit to touchpoints that occurred nearest to the time of the conversion, with less weight distributed proportionally the further back you go. This approach can be very useful for campaigns with short sales cycles, such as promotions. Think of it like Usain Bolt on the last leg of a relay race – if the team wins, it’s likely due to him.


The Position Based model can be used to adjust credit for different parts of the customer journey, such as early interactions that create awareness and late interactions that close sales.


The Linear model might be used if your campaigns are designed to maintain contact and awareness with the customer throughout the entire sales cycle.


With the basic introductions out the way, here are some tips to guide you in the right direction to picking your attribution model of choice:


Consider the length of time it takes for a lead to convert – The more interactions a user has in their buying cycle, the more sophisticated and refined the model will need to be. When buying a house, for example, the cycle could last anywhere between 6 months and 2 years plus. In this situation it would be damn right ludicrous to use a last click model that doesn’t acknowledge all the research and house viewings before heading direct to the estate agent to close the deal. A position based or time decay method would be a more appropriate tool to measure success of different channels.  


How many channels are active - The fewer channels involved in a conversion, the simpler the model you can use. Email campaigns are a great example of this. Mainly used for brand awareness and for delivering personalised content email marketing, a last click/first click attribution model could be equally effective in judging success.


Compare, compare, compare – You wouldn’t go and buy a microwave from the first store you shop in, you would shop around a bit and compare the different products, the pros and cons of each. Attribution models are no different. Find the best fit for your campaigns by evaluating each model against one another and even build your own if none are a perfect fit. Ultimately, a correctly-configured custom or algorithmic model gives the best insights, and the more you test and optimize the closer you get to attribution utopia.


So to conclude, which model should you choose? We speak to many clients who have struggles with accurate attribution and work with them to understand their business. We help them build a model that gives an accurate picture and allows you to make informed marketing decisions. If this sounds like something you or your company are interested in, give us a call and we’ll do the rest!