Digital display advertising has come under the spotlight quite a bit in 2017, for all the wrong reasons. There was a headline earlier this year, following the Times expose on brands appearing next to terrorist content, that proclaimed brands were having to respond to the "programmatic terror dilemma."
The narrative was of clueless computer algorithms placing display ads whenever and wherever, no matter the content they were appearing next to, so long as there was money to be made.
The perils of programmatic advertising are real, but despite all the negative press, the amount of digital display advertising bought and sold via this method continues to rise.
According to eMarketer’s latest forecast, a 23.5% increase this year will see programmatic digital display ad spending in the UK reach £3.39 billion in 2017. That’s three-quarters of all digital display advertising. And growth will continue, so that by 2019, it will reach £4.52 billion, accounting for 84.5% of the display total.
So, what’s going on? Why is money continuing to flow into a method of trading digital display ad inventory that is purportedly so badly broken?
The first thing that you need to understand is that programmatic isn’t straightforward, which is part of the problem. eMarketer attempts to simplify things by defining it as simply and broadly as possible, as "an automated, technology-driven method of buying, selling or fulfilling advertising."
Even then, though, you need to consider that there are different ways to buy and sell programmatically. There are three broad ways to do programmatic: The first is on an open exchange, where anybody can bid for ad inventory (think eBay), and the best price wins the placement. The second is via a private marketplace. That’s much like an open exchange (you still bid for inventory), but the environment is closed to just a few preapproved buyers. And the third is programmatic direct, whereby buyers and sellers have a one-to-one relationship, it’s just automated. There’s more to it than that, but you get the idea.
The thing to note is that only one of those methods sounds particularly risky. While private marketplaces (PMPs) and programmatic direct are fairly safe and controlled, open exchanges often aren’t. In an open exchange, you can’t always be entirely sure about the buyer or seller in the transaction.
Why doesn’t everybody just stop trading on open exchanges and move over to PMPs or direct deals?
Again, to (over)simplify, it comes down to cost and scale. PMPs are gaining a lot of traction, but they remain time-consuming to set up and difficult to scale, while programmatic direct offers little advantage over the traditional methods of directly trading ad inventory (ads bought and sold on Facebook make up the vast majority of programmatic direct revenue). Open exchanges, meanwhile, offer a truly competitive environment.
But what about those algorithms going awry and placing ads in brand-unsafe environments? Well, this where the assumption that programmatic is broken appears to be misplaced. The problem has been overstated at times, but even so, most all players in the programmatic chain are taking steps to get better at heading off brand safety issues as well as that other perennial concern, ad fraud.
These aren’t just fanciful words, there’s actual evidence out there of real programmatic progress. Integral Ad Science, a specialist in measuring the value of digital ad placements, found that the proportion of brand safety infractions arising through programmatic buys stood at just 3.8% as of first half of 2017, down from 6.9% in H2 2016. And the rates are getting smaller.
Bill Fisher is a senior analyst at eMarketer