Google Analytics Premium has existed for about two years; Data Driven Attribution Model, one of the most awesome features of Premium version, was added to the tool just couple of months ago and it is already making waves in an online analytics/advertising/marketing community. Creating and building an attribution model is a very complicated and sophisticated task/project, not everyone can do it, and even if someone can, the model can be incorrect and not accurate in many ways. For many years analysts and marketers were baffled by finding a way to better understand how customers reach the destination goals/pages/conversions and which advertising campaigns/channels (touch points) should get most credits and which do not. There are many attribution models, such as last touch, first touch, customized, U shape and others, but the problem is that all of them are not 100% right or wrong.
Well, firstly let’s have a brief explanation of what an attribution model is: Attribution model is a numerical model which provides information (in percentage) to the marketer about each marketing campaign that was on the customer’s path to the destination point which can be a conversion, landing page, application, specific webpage on the website and etc. Attribution model gives credit to each touch point depending on its influence in the path.
Ok, so, what is DDA? Well, Data Driven Attribution feature is a part of Google Analytics Premium (paid version) tool. It builds the attribution model based on the historical data (90 days) which was collected and stored in Google Analytics tool. After analyzing historical data, DDA attributes credit to touch points according to the data provided. The other cool part is that you can also submit to Google (Measurement Protocol feature) the data from offline campaigns and it will also be analyzed, considered in the process and accredited accordingly. The other very useful task that DDA feature can do is to revise the attribution model according to the changes that happened historically. So, for instance, if one of the campaigns (touch point) is underperforming comparing to the previous month, attribution model will be automatically revised and credits will be redistributed differently. The users actually have the option to choose how often the attribution model should be revised/recalculated, monthly or weekly. Additionally to all of this, Model Explorer feature of Google Analytics Premium, allows customers to see the interior of attribution model which was created by DDA, it lets them to better understand how the model was calculated and how credits were distributed among the touch points In addition this feature allows the customers to redesign the model manually if they want to make some changes to it.
So with all these tools and features, now you don’t have to be a statistician to be able to write algorithms to build an attribution model but you should still understand an online analytics industry to know how to use it, implement it and better interpret results.