Advertisers and affiliates can get a little heated when the conversation of attribution models gets brought up.
Since it's dealing with revenue - made and loss - it's rightfully a touchy subject. With first-click attribution, for instance, the entire conversion is credited to the first "touch." Let's say a shopper searches for "women's boots" and is served some ads within the search environment and she clicks on one from Retailer A. She leaves the site, and goes to her favorite social media network. A day later, she is served an email campaign from Retailer A and clicks through to the website but gets distracted and doesn't buy. The very next day, she's had enough information about the retailer and what she wants to buy, so she goes straight to the website and purchases the pair of boots. With first click, that initial paid search ad gets credit for the sale, because it was the first interaction in this conversion path that the shopper had.
In this same scenario, the last interaction attribution model would credit the direct visit to Retailer A's website regardless of so many interactions happening before it. As you can see, there are many "touches" happening between first click and last click - lots of content consumed, channels visited and influence being made - not even accounting for any possible in-store visits or outside influences. Regardless, this is why mult-touch attribution models are being used more and more - although not as much as single touch mostly because of the sophisticated data mining required.
Google, however, now offers advertisers the ability to select from many multi-touch attribution models, including linear which distributes the credit for conversion across all clicks. The three others are time decay (gives more credit to the clicks closer to the time of conversion), position based (gives 40 percent credit to both the first- and last-clicked keyword, with the remaining 20 percent spread out across the other clicks in the path) and data driven (gives credits to clicked keywords based on how imperative they were in the conversion process).
As Facebook challenges Google for ad dollars, it too has sophisticated attribution modeling for advertisers of all sizes - even crediting conversions to "views" of ads and not even clicks. That would, however, benefit the social network to count those as the more ROI provided, the more likely advertisers are to continue spending money there. Facebook, however, does have a similar advantage to Google in that it has data from signed-in users, which allows them to track users across devices too - in order to best credit a sale. Affiliate Window, however, is one network that can also track conversions across devices so it's not just limited to Google and Facebook, but this is not the industry norm (yet).
There's a lot to learn about attribution modeling but fortunately for the industry, the giants like Google are making it easier to pay up across touchpoints so that credit is more reflective of modern consumers' fragmented path toward conversion.
