Attribution Modeling for Dummies
First of all, what is attribution modeling?
It is the model in which you decide what channel (where your traffic comes from) gets credit for a conversion whether that’s a sale, a lead, or an action taken on your website.
At first glance, it seems simple. A user visits your site through a PPC ad and buys something. Therefore, 1 user through 1 PPC ad = 1 sale. Done. Your ad works, your marketing is successful. Congrats!
Whoa Nelly. Not so fast. What if that person came to your site through Facebook first, then saw a retargeted ad, then searched Google organically and finally clicked on your PPC ad to make that purchase. Is your formula still the same? Does that last interaction with the PPC ad get all the credit? Or does Facebook get some credit for attracting that customer in the first place? Without Facebook, would you have that customer at all? Maybe Facebook should get all the credit in that case.
Here’s where attribution modeling comes into play. Google Analytics gives you the option to use the following default models or you can make your own custom one specific to your business:
|Linear: This means that every interaction contributes equally in the final sale (or conversion)|
|First interaction: This attributes the conversion to the first interaction|
|Last Interaction: This attributes the conversion to the last interaction|
|Time Decay: This means that the closer to the last action, the more gets attributed to that action.
In other words, the closest action to the conversion means the most, second closest means the second most and so on.
|Position based: Simply, you chose how much gets attributed and where (for instance, first touch could get 40%, middle gets 20% and last gets 40%)|
So what’s the best option to judge your success on? That is a very hard question to answer and will vary alot depending on your users and your marketing strategies.
If you do digital marketing all over the place using many methods (multi channel or omni channel), you will likely have more touch points on your users whereas if you do no digital marketing and all of your traffic comes through organic search, it will be much clearer where the conversion can be attributed to.
Attribution Model Analysis
Under Conversions in Google Analytics there is an “Attribution” drop down. Here you can compare the different models.
Here is a quick example of what this can look like. For this site, there is a big push through paid search. In this case more people find the company through paid search first but far fewer buy directly (last interaction) after coming through an ad. This would suggest that people are finding this company through paid search but then doing research elsewhere before coming back to make a purchase through another channel.
This is just one quick example of an insight that you can pull out of this report, but this is an incredibly powerful tool that can really help you better understand how your marketing efforts affect your bottom line as well as help your marketing team judge what's working, what’s not, and where improvements can be made.