Thus far in our attribution series, we’ve covered how you can consider giving credit to your email channel and how database marketers gain attribution credit. In today’s post, we’ll be covering ecommerce attribution – or – how to make sense of your dollars and cents.

Ecommerce businesses live and die by their attribution model. If you nail it, you can effectively allocate your budget and scale your marketing channels proportionally and effectively. If you misattribute successes to the wrong channels, you will be placing bets based on bad intel, and your ecom sales will reflect that.

Think back to the last few ecommerce purchases you made. Specifically - think about the brands you recently tried for the first time. How many touch points did you have with that brand? Chances are it was not just 1 interaction. You may have seen a friend post about them on social media (with a surreptitious referral link that you clicked), then been retargeted by the brand as you browsed on and off social media, then signed up for their email list to receive 10% off your first order, and finally arrived at their site to purchase, but not before a quick Google for any coupons better than your current 10% off. While your experience was pretty straightforward in terms of modern ecommerce, you created a pretty nasty path for an attribution team to make sense of. Let’s say you purchased $50 worth of merchandise. Your $50 sale will likely pop up on at least 5 channels’ reporting – of course the affiliate team, the email team, the retargeting team, the social team, and the organic search team will want to take credit for this new customer – not only your $50, but your extensive future lifetime value. However, if all 5 teams are taking equal credit, future budgeting decisions are going to be murky at best, and way off base at worst.

If you’ve made it this far and are waiting for the perfect ecommerce attribution model – sorry, you’re in the wrong place. There are dozens of ways to attribute online sales – first click, last click, equal weighing of all touchpoints, giving more credit to first and last click, and so on. In truth, there is no perfect attribution model, just like there is no perfect tracking technology. Some sales will be straightforward, with one interaction that leads to a nice tidy sale. Others, like the example above, will be linear paths in which multiple channels come into play. And yet others will be a mess, where there are huge lags of time, multiple unique identifiers matching a customer, and every type of marketing popping up in a users’ journey. This is all to say – do your research, try out a few attribution models, and keep tweaking your methodology over time. Just as your customers and marketing strategies are changing, your attribution model should be evolving over time.