Programmatic, meet Mood

It’s time to take account of the mood and the moment a person is in at the other end of the ad server, writes Jacqui Wallis

There’s a powerful subtext to programmatic. One that definitely merits repeated viewing. Not least as it now makes up 79% of online display spend. As well as 99% of conversations around how it’s the scourge/saviour of advertising.

But first, let’s get the most obvious secret of its success out of the way. Programmatic sold itself as hyper-targeted, efficient, scientific even. In short, no one ever got fired for buying it. Algorithmic buying promised the end of wastage, even if, as Rory Sutherland has argued, it had a tendency to pick the path of least resistance, playing up its influence. Even as it conveniently ignored the knottier (and more important) questions.

What has been less well documented is programmatic’s unavoidable psychological pull. Of offering security – a data safety net if you like – to reassure an increasingly insecure market. Even if we don’t completely understand artificial intelligence or machine learning, it still sounds a more reassuring insurance policy than something apparently so vague as creativity.

I was reminded of all of this by William Hanmer-Lloyd’s piece in Mediatel, setting out the many potential benefits of applying behavioural science to advertising. The intersection between academia, psychology and the marketing world has been there

for a while, of course – as Dominic Mills has documented. Since influencing behaviour is our stock in trade, it’s a wonder it still has doubters at all.  In the words of Richard Shotton, psychology is “far more robust than what sometimes goes for rationale in advertising.”

But as sociologist William Davies has observed, from the anti-vaxxer movement to Olly Murs in Selfridges, we’re seeing a societal shift to feeling over reflection. Is advertising suffering from its own shift away from experts and evidence, in favour of a craze for novelty and short-term emotional reassurance?

Programmatic may give us the warm glow of diminished risk, even security. No mean feat, in the face of shortening CMO tenures and perpetual pitchapalooza. But how does it stack up as actually executed, versus how we think and act as humans? What I’m trying to say – you guessed it – what happens when programmatic meets behavioural science?

Retargeting & Radio in the ‘20s

Let’s consider first the most overused tactic in the whole programmatic oeuvre. Retargeting, in its simplest form, is the serving of an ad based on a site visit. While it may not be all that intelligent, artificial it certainly is, at least when experienced in the wild. Especially in the absence of opt out, or even checks in place you haven’t already bought the product. On the scale of nuance, it’s more Dalek than Cylon. And little wonder then it has been fingered as playing a major role in the rise of ad blocking.

Besides retargeting, the vast majority of machine-driven buying is built on the bones of techniques we’ve been honing since the emergence of radio in the 20s. Demographic segmentation and targeting have stood the test of time – but crucially, context as audience proxy versus targeting and reaching audience directly are not the same thing. In fact, it turns out they’re as different as Dynatron and DAB. There are three reasons why.[advert position=”left”]

– One is the accuracy of third-party data online, the backbone of many a programmatic campaign. For starters, it’s largely deterministic, and not always reliable. For example, you have only 50 per cent chance of correctly identifying whether a user is male or female.

– Second – the sheer omnipresence of personal targeting. What’s impact on effectiveness when most advertisers are targeting people the same way as you, 24:7? For competing brands, equal access to a shared market for data and inventory, coupled with a single execution tactic invariably risks a tragedy of the commons.

– And third, as Hanmer-Lloyd points out:Demographic or attitudinal audiences are not reflective of how audiences buy or think, and usually don’t explain who is most likely to be influenced.

This is especially true in our online lives versus other media. With the shift from content as audience proxy, to audience itself, what’s emerging is our ignorance of the sheer number of moods, and even identities we adopt online as we go about our daily lives.

From work research, to shopping, to passive viewing, active reading or talking and interacting in multiple other ways – all in the same browser session. All things considered, does this combination of behaviours lend itself well to a set of permanent, static identifiers? If anything, it’s an argument for real-time advertising to restructure around the way we experience it instead – on present mood.

Channel 4 is already moving towards doing this for TV, and they’re not the only ones looking to take online advertising in new directions – away from personal data and micro-targeting. Early tests suggest their matching of ads with “positive contextual moments” within programmes is onto something – with brand recall doubling in testing.

For all the talk about big data, processing power, machine learning – our industry is just as much a reflection of our individual, understandable human weakness – a need for reassurance, stability and security amid the uncertainty.

Here’s an idea: let’s try applying a bit of hard science to back up those techniques. Or failing that, just take account of the mood and the moment the person is in, at the other end of the ad server.

Jacqui Wallis is Managing Director of Illuma