One of the hottest marking topics is attribution — i.e., which marketing channel (or channels) gets credit for a single sale. Attention to attribution is skyrocketing because in many companies reporting has been set up so that every channel claims 100 percent of the final sale, even when their efforts were only partially responsible. This "double counting" of converted traffic has led to inflated marketing channel reports and, more importantly, has companies overpaying for sales and leads.
In an attempt to eliminate double counting, the majority of retailers choose to only value the "last in" marketing channel, giving that channel full credit for the sale. This is particularly prevalent, but also very problematic, in the affiliate world, as last-in attribution also determines who gets paid when there are multiple affiliates in the same click stream.
So what's wrong with last in? Five years to 10 years ago, there were far fewer marketing channels online and it made sense to value the channel closest to the sale. The logic was if a customer came through a given affiliate a few months ago, then came through a second affiliate at a later time, the second affiliate probably had more responsibility for the transaction. So last in became the default industry standard since it seemed that there could be more abuse with first-in affiliates who tried to set a lot of clicks in hopes that someone might buy later.
However, today's online shopper may come through up to five online marketing channels before completing a single transaction, with many industry studies putting the average at about 3.5. And while the last-in attribution model grew from the best of intentions, it's led to a very large number of affiliates that explicitly develop and deploy business models to be that very last click in a purchase cycle. Unfortunately, the value of the majority of these methods are questionable, as they inherently target consumers who fall into one of the following categories:
- consumers already looking for the brand (e.g., company name vs. generic product search);
- consumers already on the shopping cart page who pop out to look for a discount code;
- consumers who have just been on the website and are returning; and
- consumers trying to go to the site directly before being intercepted.
Examples of these types of affiliates include loyalty toolbar sites, coupon sites, cookie stuffers, domain typos, retargeting sites and others that usually do nothing to introduce new customers to a brand. Since these affiliates are swooping in at the last click and taking full credit for the sale, their payout is often undeserved and outsized.
Counting on these kinds of affiliates for the majority of a program's affiliate sales has two major downfalls:
- these affiliates are overwriting the cookies of affiliates earlier in the funnel who actually introduce new customers to a brand; and
- these affiliates are being paid a new customer premium for targeting existing customers.
The impact of No. 1 is what often ultimately encourages quality affiliates to leave the program due to poor performance, and No. 2 can dramatically impact program costs and lead to high channel overlap.
These aren't the only issues, however. There's also the problem of companies applying last-in rules across multiple marketing channels, which includes deduping affiliates without full disclosure.
For example, let's say a customer first learns about a company's product via an affiliate who has a popular blog, visits the retailer but doesn't buy, and then goes back again later through branded pay per click or an email campaign. Some companies don't pay affiliates for this sale despite the important role the affiliate played in creating a new customer. The affiliate creating the demand is paid nothing, yet another channel is credited for the revenue and conversion while doing far less work. The activity that should be valued the most — bringing new customers to the brand — is being discouraged under this model. This is how we know that the last in logic no longer makes sense.
So what can a merchant do? Here are two examples of how top e-commerce brands have taken a smarter approach.
Tiny Prints’ (part of Shutterfly) award-winning affiliate program has a multiattribution model thanks to technology developed with its affiliate network and our company. Using its model, sites that show up at the very end of the purchase funnel (e.g., within a few minutes of purchase) may have their commissions reduced for those transactions if another content-based affiliate was overwritten and credit is given to the first affiliate. This approach has the dual benefit of making sure the right affiliates get paid and ensuring Tiny Prints isn't overpaying for orders that have a high channel overlap.
One Kings Lane, a flash-sales site for high-end home goods, has taken a different approach to building a high-quality affiliate program. Its solution is to block coupon, loyalty or any other nonbrand relevant sites from its affiliate program. This puts all affiliates on a level playing field, making the program very attractive to those affiliates who drive top-of-funnel orders and brand awareness. At the same time, it preserves the One Kings Lane brand.
It's clear that the last-in attribution model needs a serious revisit as companies take a closer look at multitouch attribution. Not only does appropriate attribution allow companies to value and incent their marketing partners appropriately, it can have huge impact on the bottom line.
Bob Glazer is the founder and managing director of Acceleration Partners, a digital strategy and marketing firm focused on driving profitable new customer acquisition.