Conversion attribution modelling is one of those things that every marketer is aware of and understands its importance. However, it is usually black-boxed and delivered with no details of how it is calculated. This article will show you our approach towards attribution and why we believe it could be just about what you need to help you distribute marketing budgets in the best possible way.
Attribution modelling is all about learning which online marketing channels (e.g. Google Ads or Display) or, to be more precise, which types of campaigns make users become your customers. This in turn, lets you decide where to put your marketing budgets to increase sales.
It would be so much easier to read Google Analytics traffic sources reports to see the ROI on every channel and each campaign. However, the standard traffic sources reports are based on the last click data, that is on sources that brought users to the website just before the conversion. But as we all know, customers don’t make their decisions the first time they enter the website. More often their path is longer and more complicated. Let’s take a look at the following example:
It shows a path of a hypothetical user that starts its customer journey on the website with SEO visit and after 4 days makes a conversion – purchases products from the store. In that situation the “traditional” – last-click attribution model will attribute the conversion to the last session source, that is to SEO – branded search. However, as we can see, the path is more complicated and covers more steps than just the last one. Assigning 100% of this conversion just to the last step would be unfair in regards to all the other sources.
What if the path was more complicated? For instance it might involve more just than one purchase, like here:
How to assign sources to conversions then? How to calculate the costs of acquiring the first conversion, the next ones and how to pair it with customer lifetime value?
This is where the custom data-driven attribution models come into play. The role of data scientists is here to devise such methods to assign their share of conversion to every step on customer journey according to its calculated influence on the customer’s decision.
Ready to know more about our custom models? Proceed to the next article.
