Towards an End to Retargeting

Want to see a key difference between how our industry uses data and how direct marketers and brands have been using data for decades? Try this experiment: buy something from a Frontgate, Victoria’s Secret or Land’s End catalogue, then wait 30 days and watch the magic unfold.

You’ll probably receive a new catalogue from the merchant with whom you did business within your product shipment, as well as a new one in a few weeks. You’ll also get other catalogues from different – even competing — merchants. And you’ll almost certainly be able to ascertain what the relationship is to these other catalogue’s products and what you bought. For example, if you purchase five big planting urns from Frontgate, you’ll probably get catalogues from companies that sell other home-and-garden items, like benches, hoses and other gardening equipment.  It’s not terribly complicated.

Meanwhile, if you make the same purchase online, you might see dozens of ads in the next few days, let alone the next months, featuring the same exact product you just purchased, or similar products from the same merchant. When Michael Learmonth wrote “The Pants that Stalked Me” in Advertising Age last year – a column that brought him interviews with many other outlets, including the New York Times – he struck a nerve that still resonates in our segment. The question is, when does retargeting become stalking?
The offline model of cross-targeting is used so seldom online because of our affection for retargeting technology and also because of how bad most of the data is. It’s no wonder that consumers are creeped out by our industry. But let’s face it, those pants should not have been chasing Learmonth. And while consumers are creeped out, marketers should be upset about this too, since most users who get retargeted will return anyway. These are wasted impressions – and they may be doing more harm than good. Do you know what the source of your data is? If you did know more about the source of your data, think about how much more you could do with it. Instead of treating it in aggregate, what if you could optimize against it by its source with an understanding of the brand behind the data? This is the kind of transparency that enables cross targeting.
Catalog companies rent each other’s lists to enable cross targeting, as illustrated above. Why don’t we do more of this online? All the sexy, algorithmic data delivers very little value if the data is only used by buyers in silos, with no transparency and no ability to optimize by the source. Also, if the algorithms are so smart, why not apply them more broadly against the simplicity of cross targeting? Shouldn’t these algorithms be able to tell us more? They can if you apply them more strategically.
Those of you who have been in the research world for years all know the corollary about Saab drivers and Apple products, where more than 80 percent of Saab drivers used Macs instead of PCs at some point in the 1990s. How many corollaries can our intelligent algorithms find online if we spread them across campaigns?  Would that Frontgate buyer be more interested in clothing from Lands End than in something from L.L. Bean?  You won’t know unless you examine this – which requires that different level of transparency.
Instead of chasing the same user or chasing some new user created by an algorithm, why not attract new users from other sites that are similar to your own? Why not deliver more new customers from similarly situated brands, instead of the same ones who plan on returning anyway?  If it’s new customers that you want, cross targeting makes a lot more sense than simple retargeting, and it won’t create bad will with your customers or their consumers.