Forget what you know

Living in the moment, programmatically

Back in the day, web pages were ranked by Google using keywords. Nowadays they are ranked by ‘entity’, which describes the content of the page and its relevance. The best programmatic agencies now use a blended system of keywords and entities and then qualify each page by context. When it comes to a campaign, these pages are then filtered by relevance. Again, until this was automated, checking these pages for relevance was labour–intensive, and meant opening individual URLs and reading the content. As these pages are now read and subsequently ranked by relevance in an automated fashion, machine-learning comes into play. But, rather than narrow-down and target audiences which makes the pool smaller, these machines can heuristically do the opposite, and search out like-minded people to grow your audience.

At a recent event run by the boffins at Illuma, it became clear that people’s propensities to buy stuff on the internet can rarely be categorised by their historical data. It’s about catching them at that moment in time when they are ready to make a decision, however impulsive that may be. But how? You can see why serving them ads from a site they’ve just browsed can become so overt and annoying. John’s just been looking at suitcases, so he’s then served ads for the cases he’s viewed on every site he then goes to. These subsequent pages he’s viewing can be so irrelevant contextually to buying a suitcase that the ads stick out a mile. Most punters find this creepy if not downright intrusive. Plus there are the privacy issues around the use of cookies and GDPR which will only become stronger in time. Browsers like iOS and Chrome are already ahead of the curve in adopting this compliance.

Illuma (other programmatic agencies are available) use what they call a ‘nearest neighbours’ system to scale reach. Counter-intuitively, by ignoring all historic data and trends, which may well be out of date and can only serve to narrow audiences, they look at which pages individuals are actively looking at in real time. For instance, Rachel could be buying insurance but she’s been distracted and is looking at a yoga site, so that’s where they serve the insurance ad. By then targeting other individuals simultaneously reading similar yoga pages, they can then scale the campaign in real time. Illuma have proved that when individuals are reading similar pages at the very same time, there is a propensity for them to act in a very similar way.

This really turns things on their head. Instead of going after your existing base and trying to create referrals you can find like-minded people under the radar. Not only that but also you can be seen on pages not cluttered by competitors or others that may be distressful to your brand.

It makes so much sense when you think about it. Rachel, in a moment of concentration, on a yoga site that she trusts, is then presented with an ad for the very thing she has tasked herself to buy…insurance. And many of her peer group are too. Instantaneous and action-oriented, it avoids all the pitfalls of intent-focused search and content-based approaches.

Fascinating, don’t you think?