Is it time to re-think programmatic advertising?

JP Morgan Chase found it was advertising on 388,000 sites to no effect. Brands using programmatic advertising need to sit up and take notice.

Over the past several years we've seen programmatic ad spend saw, Icarus-like, to dizzying heights. At the start of this year, UK programmatic spend was predicted to increase by over 30% in 2017, whilst US programmatic ad spend has tripled since 2014, becoming the single fastest growing ad medium. So-called gurus predicted this trend would continue, by doing what so-called gurus always do and assuming the future will resemble an extrapolated version of the past. Hence the predictions that programmatic spend will double by 2020, reaching an incredible $42 billion dollars globally. But the present has got in the way.

Ever since The Times uncovered programmatic ads for major brands appearing on the websites of far-right hate groups or terrorists (See Mercedes Advert on an IS video below), brands have begun to look again at programmatic, and we're starting to see the evidence piling up that programmatic is failing to deliver many of its promised gains.

The problem with programmatic isn't just that advertisers may be trashing their brand by appearing on questionable websites - this can fairly easily be avoided, as Google provide tools to stop your content appearing on sites associated with certain keywords. However, only 15% of marketers are using these tools, begging the question, who's responsibility is 'brand safety', the client's, the agency's, or the ad provider?

The key problem as I see it is that programmatic just isn't doing what ads have to - engage their audience and drive business goals. JP Morgan Chase revealed recently how cutting 97% of sites from their programmatic ad campaigns had no negative impact whatsoever, a finding that for me is far more scandalous and indicative of the failure of programmatic advertising than the revelations published in the Times.

JP Morgan Chase was, until a few weeks ago, using programmatic to run ads on 400,000 different websites. The idea being that automated bidding, tracking and clever algorithms can let you target the perfect people at the lowest possible cost per click. Sounds great in theory, but in practice, it's not working. Of the 400,000 sites JP Morgan Chase was programmatically running ads on in a given month, 388,000 were not generating any kind of action from potential customers whatsoever. That is, they were seeing the ads, but not clicking on them.

But that damning set of performance statistics is not the worse of it. Of the 12,000 sites that had generated clicks, 7,000 were deemed to not be the kind of site that JP Morgan Chase wanted to advertise on when checked manually. So even of those ads on sites that were 'working' (in the loosest possible sense of the term) over half were ones the brand didn't want to be associated with. This is disturbing, to say the least. But it also shows programmatic is systematically failing to live up to its promises.

If the theory behind programmatic was correct, stopping running ads on 97% of the sites it advertised on should have a big negative impact on the campaign's effectiveness. Since programmatic should let you target the most valuable people on the lowest cost sites, removing 395,000 sites from your list and only using a whitelist of 5,000 manually approved sites should mean increased costs and reduced effectiveness. Yet JP Morgan chase found nothing of the sort. Their CMO told the New York Times that they hadn't seen any deterioration in their performance metrics. This to me is proof that the benefits of programmatic have been over-sold, and it's certainly not delivering the promised benefits. On top of that, you have the issue of endemic ad fraud, with bots wrongly inflating view counts and clicks.

Brands would be wise to review their current rates of programmatic ad spend, and consider using greater human oversight and whitelisting, both in the interest of brand safety and in getting the greatest return for your ad spend.

The future of programmatic

For me, it's clear that programmatic advertising has been over-hyped, in large part by ad tech providers. It's soaring flight has brought it too close to the sun, and now it is likely to fall back to earth. In the short to medium term, advertisers are likely to require more human oversight and this will constrain programmatic ad spend growth. The future of programmatic ad will depend on new technology developments. It will either stagnate as a medium as more and more advertisers realize its major flaws, or it will re-invent itself. It could do this by leveraging the power of new artificial intelligence tech to make ad placements far smarter and less risky for brands. But until someone builds a program capable of detecting threats to brand safety, advertisers will continue to find programmatic problematic.