×

WireColumn: Content & Persuasion, the Ad Tech End-Game

Ezra Pierce is co-founder at PubSquared.

Advertising is about one thing: persuasion. You put an ad in front of a person, they see it, a desire is created and an action engendered. What’s odd is that this isn’t what our corner of the advertising industry is focused on these days. The industry is currently obsessed, not with great content and persuasive creatives, but instead with the latest and greatest software. That’s perfectly sensible, but I can’t see that lasting. Here’s the thesis: trading technology is having it a moment, but we’re going to reach a new equilibrium and top-notch tech will be table stakes, not something to brag about.

Advertising technology is at a rolling boil at the moment. Venture capital firms have shown a willingness to invest huge quantities of funding to anyone with half a bright idea and a marginal ability to execute. Simultaneously, the barriers to entry have been progressively lowered by the arrival of RTB ad serving, cloud computing and big data infrastructure as commodities. These factors have converged and accelerated the pace of progress in display advertising. The question is, does the innovation continue forever, or is there a natural end point to this progression?

The answer is a bit of both. The problems that are being addressed currently – let's take three of the big ones – performance optimisation, scale for rich media implementations and creating an integrated and safe demand environment for publishers, are fundamentally solvable. Problems that can be solved, and solved profitably, get solved. The bit of both comes from the factor which profitability plays. There will always be more to do, but at some point it ceases to be profitable to do so.

Let’s just take one of these problems and pull it apart – performance optimisation. The state of the art today is the Demand Side Platform – Invite, MediaMath, Turn, AppNexus, etc. Fundamentally, their approaches to optimisation come down to surprisingly simple factors: how has this tag performed for this creative in the past, how many times has the user seen the advertiser’s creative, perhaps the time of day and day of week, etc. Clearly, there’s more work to be done here in at least two areas: 1) going to more granular levels of analysis of performance and 2) bringing some sanity to the ways in which performance itself is measured. Going granular means finding additional ways of dividing up users – male, female, age, purchasing behaviour (there are people who buy things on the internet and people who never do) and myriad other factors. Dividing allows the buyer to move closer to the correct price for every ad impression. It also has it’s own inherent limits, eventually, when cutting up users more doesn’t yield enough data to generate a statistical significance.

A further catch is that there are diminishing returns to those divisions – knowing the difference between a user in London E1 and W1, great; the difference between age:31 and age:32… not as interesting. There is a threshold at which the increasing granularity stops being profitable. By extension, when marginal improvements decrease in added value, so does the attractiveness of funding new DSPs.

To put it simply – performance optimisation will become a commodity and there’s a reason Venture Capitalists don’t go looking for really innovative potato farmers.

Moving on to measurement, it’s incredible that we’re (almost) all still looking to clicks, post-click conversion and post-view conversions. Take clicks – when did you last click on an ad? Do you think people that click on display ad units are your ideal customers? Take post-view – how are we still measuring conversions on ads that the user probably didn’t see? (Let’s talk about the fantastic PV performance of social networks while we’re on the subject) – and one event get’s all the credit?

There are problems, but again, there are technological solutions. AB testing as standard practice would be a start (and a disaster for ecommerce retargeters) and time in-view would be most of the way to a robust solution to some of the measurement issues. Advertisers will start demanding these things, and they’ll get them. Back to my main point – there are problems in ad tech, but they will be solved and commoditised.

What we’re left with when that happens is the work that media companies and advertising agencies have been at for decades. Creating good content and coming up with gorgeous, persuasive ads. Today’s innovation will be assumed and the focus will return to the irreducible core of advertising… ads and somewhere to put them.