Commentary

The Affiliate Ecosystem -- And How to Attribute

Many vendors provide cross-channel attribution services. Not so many offer them in the context of affiliate marketing. Rakuten Marketing, an affiliate network that helps e-commerce clients with tasks ranging from display buying to attribution, does.

Affiliate marketing, in which advertisers partner with publishers to drive both sales and traffic, is expected to top $6.8 billion in 2020, according to a survey released last year by Rakuten. For insights into this growing area, we spoke with Danny Kourianos, senior vice president of product strategy for Rakuten.

What’s your approach to cross-channel attribution?
Everyone wants to have a logical and sound attribution methodology. But how does that equate to sales? It’s not just being able to say that you want to assign credit to touchpoints: It moves way beyond assigning credit to providing insights that will have a meaningful business impact.

Do third-party affiliates complicate this task?
That piece is not that complex. Some solutions are probably going to deemphasize the value of affiliate marketing: They will probably look at the last click that drove the person to the site, and see affiliate marketing as one big bucket. You’d be surprised, though, at how much discovery is at the top of the funnel. People assume it’s focused on cash back, coupons, vouchers, but it’s not the case. There’s upwards of a million publishers in the network, from bloggers and influencers to content marketers. Affiliate marketing is a whole separate eco system unto itself.  

advertisement

advertisement

Where do you see marketers having trouble?
We found when talking to retailers that the majority are overwhelmed with taking on attribution. Even today, you find very little adoption in the pure-play space of cross-channel attribution. Few have the analytics teams in-house who can dive in and affect change across the organization.

And once they do, how do they end up in jams?
A lot of ecommerce sites might have site analytics integrated into their tools. But understanding where my traffic is coming from is different from understanding my marketing mix on sales. Some firms have relied too heavily on site analytics tools that only give them the last click. So they over-invest their spend on last-click touchpoints and under-invest on areas that are going to have just as much impact.

What’s the purpose of the models — to find lookalikes?
The model is a standard, out-of-the-gate model that divvies up the touchpoints so there’s a weighting for every touchpoint. You start with the sale, look over a predefined period, add up the touchpoints, then divide those by the total. We’re not just looking at those that resulted in a sale: Equally important are the journeys that did not result in a sale. What was unique about this segment of users? It takes looking at a lot of data.

Do the individual channels have much of an impact?
When one of the luxury brands added social media into the mix, we saw a 567% lift with those who were exposed to social over the non-exposed group. That means I obviously need to increase my spend in social. And we’ve found that converting sales with multiple touchpoints brings dramatic increases in the average order rate.

What about offline media?
We can take offline data and email addresses and link those up with cookies to tie them back to the digital touchpoints, and see when digital has an impact on offline.

How has the company evolved?
When we acquired LinkShare, our affiliate piece acquired a display ad solution and a retargeting solution operating at the lower end of the funnel. Then we took a step back to maximize the synergies between the two entities. Did we need to acquire more media assets, horizontally, to fulfill across the CMO’s entire budget? [The answer apparently was yes: Rakuten recently acquired Manifest Commerce and NextPerf.]

Do you measure non-Rakuten properties?
We can also measure touchpoints outside of the media we manage — it’s a non-biased and agnostic view.
Next story loading loading..