When CEO Alicia Navarro started Skimlinks 10 years ago, affiliate marketing was a transactional game. Drive the last click before a sale, scoop up the credit.
But that scenario didn’t reflect that content producers tend to bring value higher up the funnel.
“Affiliate is very good at rewarding and incentivizing last-click behavior,” she said. “But if you’re a marketer trying to properly allocate budget across all channels, it can be really deceiving.”
Giving retailers information about which publishers were actually responsible for the awareness that ultimately leads to a conversion, however, means that they can appropriately reward quality editorial content, Navarro noted.
In its early days, Skimlinks automated affiliate marketing processes like inserting tracking links into content and then keeping those links up to date.
But, increasingly, publishers use the service to capitalize on the purchase intent data generated by browsing and click behavior as readers engage with content. Skimlinks hosts a data co-op for its publishers to share that information, which in turn can be used to create anonymized segments based on predictions of future shopping behavior. Publishers get paid when their intent signals contribute to a segment used by a retailer in a programmatic exchange.
“With retargeting, you’re only getting people who have already gone to a retailer’s site,” Navarro said. “With what we’re calling ‘pre-targeting,’ you’re getting people who have been researching particular products or brands – the people who want to buy, they just haven’t clicked on a link yet.”
AdExchanger caught up with Navarro to talk all things affiliate.
AdExchanger: Skimlinks recently struck a deal with Tapad. How is cross-device behavior changing the affiliate game?
ALICIA NAVARRO: One, it allows us to match up cookies that look like different users and combine them into a single profile across mobile and desktop, which helps a lot when we’re trying to predict shopping behavior and preferences. Two, it enables advertisers to target individuals in-app. And three – this is something we’re still working on – we’ll be able to know if someone looks at something on mobile and later converts on desktop so we can properly attribute the affiliate commission.
Some publishers, like BuzzFeed, king of the listicle, were born for affiliate marketing. How do you convince news sites to try it?
We don’t run on news sections, only appropriate sections, like a holiday gift guide, for example. But when we first began, we did have to spend a lot of time dealing with objections around church and state and never the two shall meet.
But over the years, publishers have realized that their users are pretty savvy. As long as publishers write honestly and authentically about products they genuinely like, the fact that they make a little commission if someone buys something is not a problem – as long as it’s disclosed.
They’re starting to see that if they weave “comtent” – commercialized content – into their editorial, that it can be a great source of revenue. It yields a ton of data and users that engage with it tend to be higher engaging users overall.
Look at the success of publishers like Wirecutter, owned by The New York Times, or Refinery29 – it’s almost entirely comtent-related. Others, like New York magazine with The Strategist or the Mail Online with Femail Fashion Finder, create commerce-oriented subsections of their main site with the same voice and feel. They’re thinking strategically about how to authentically weave and intermingle content and commerce.
What kind of data is being generated when readers interact with publisher content?
Publishers can get rich insights on the kinds of products and brands that users like, and that can inform who they should be approaching for ad sales deals.
But outbound click behavior and engagement with publisher content is [also] the best measure of intent beyond paid search.
But how do you differentiate between a shopper and someone who’s just reading content?
It can be tricky. We categorize data by looking at the states of the purchase process.
At a high level, the first group is made up of people with a general affinity for a brand or retailer. They haven’t necessarily started actively researching and they probably haven’t clicked on any links, but they are looking at content about a certain brand or product.
The next stage is “wants to buy,’ which is made up of people who are probably still a few weeks away from buying, but they’re researching and probably have clicked a link.
Lastly, the “about to buy” people show a greater concentration of browsing and clicking behavior and will probably make a purchase in the next few days.
We use the data we collect from the affiliate side to predict shopping behavior down to the product, brand or category and to predict what stage in the purchase process an anonymous cookie is at.
Affiliate marketing aside, why do some publishers seem to fail so spectacularly at combining commerce and content?
We always breathe a heavy resigned sigh when publishers think that the only way they can or should make money from commerce is by doing the whole thing themselves. Some think it’s what they need to do for the largest share of the pie.
But doing both content and commerce rarely works out, because being a retailer is really hard. Look at Condé Nast and Style.com. It was too distracting and too costly to do both. In the end, they realized that they needed to return their focus to the curation side of commerce.
