“Data-Driven Thinking" is written by members of the media community and contains fresh ideas on the digital revolution in media.
Today’s column is by Kevin Jennison, chief technology officer at Gladly.
A well-targeted online ad can be a delightful moment – that instant you find the exact gift or concert you were looking for.
Sadly, these moments are rare. Despite a decade of innovation, ad tech still fails to consistently deliver delightfully targeted ads. Worse, it seems impossible to fix: Too many companies are sucking up data and spitting out ads to imagine it could ever be a coherent and pleasant user experience.
But we must try. There are several approaches the industry could take.
User Control And Shared Preferences
One approach is to let users tell advertisers what they want, but this requires users to trust in whomever is managing their data and notice improvements in the ads they see. Notable attempts include Facebook’s ad preferences, InMobi’s Miip and Rubicon Project’s Project Awesome, though the idea is far from proven.
To provide a coherent ad experience across the web, ad tech companies could share a database of anonymized user preferences. The database might look something like a data management platform, but with entries for brands and interest categories in addition to users. Then, the fragmented ad world could speak the same language: If a user tells one ad network that he or she never wants to see ads from a particular advertiser again, the network could log that preference so that other ad tech companies can respect the user’s preferences.
Cooperation on this scale between ad tech companies might be a pipe dream. Yet there’s some precedent in the somewhat inert AdChoices, a shared self-regulatory mechanism for the industry and agreement on standards for ad categories in OpenRTB. If better user experiences grow overall online ad spend, cooperation would make financial sense.
The worst part of this approach is that it’s a disaster for user privacy. A variant could be to store user preferences solely on the client, with regulation to ensure that servers use the data only for a single ad transaction. Mozilla has explored targeted but privacy-conscious advertising of the sort.
A Throwback To Contextual Targeting
Let’s forget about people-based marketing for a second and remember that we used to target ads by their surrounding content, the user’s location and time of day. There are huge upsides to this approach: reduced technical overhead, far fewer user privacy concerns and probably a greater likelihood that the ad is interesting to the user in that moment. The world of machine learning opens doors to targeting ads based on many other correlates to ad performance that aren’t user data. And interestingly, the rise of native ads starts to feel a little like the second coming of content-based targeting.
The major drawback of contextual targeting is that it doesn’t include the capabilities that marketers now expect, including demographic and cross-device targeting. For this alone, I’d wager the industry is unwilling to take it seriously.
Or Let’s Wait For Technology To Get Better
Some ad tech optimists quickly dismiss the claim that we need an alternative approach to solve the user experience problem of bad ad targeting. They claim technology will solve it: More user data and better algorithms will eventually make ad targeting so good that users will love the ads they see.
It’s possible. There’s no doubt that ad tech is improving, especially under the pressure of ad blocking. However, with a rising rejection of all online ads, increasing user awareness of their online privacy and perhaps increased regulation, there may not be time to wait.