Dark art or deep science?

Unless you purposefully seek it out you won’t necessarily come across anything untoward on the internet. Like slow-moving fish lurking on the riverbed, the stuff we don’t want to see – the bad stuff – remains hidden beneath a thin layer of iridescent algae, barely visible to the naked eye. But lurking there all the same.

Programmatic advertising has changed all that: with its clever algorithms and technological know-how, it’s brought the horrors of the deep right to the surface, giving big-budget brands around the world the fright of their lives.

The situation came to a head in February, when The Times revealed that household brands had been inadvertently bankrolling Islamic extremists, white supremacists and pornographers by advertising on their websites and videos. Then, in March, Coca-Cola, PepsiCo and Walmart pulled non-search ads from Google after The Wall Street Journal discovered they were being shown alongside racist or anti-Semitic YouTube videos. Others soon followed, including HSBC, Havas Group, Lloyds, Marks & Spencer, Thomson Reuters and Dropbox.

Later that month Google held its hands up and took some of the flack by admitting that brand safety is a global problem and that “we need to do better”. And in April they announced plans to introduce an ad-blocking feature in the Chrome web browser, which would filter out certain online ad types deemed to provide bad experiences for users. Serious stuff. All of a sudden, the internet – purveyor of clever, shiny things like social media, ecommerce and video – was suddenly looking a whole lot less enticing.

A moral maze – the ethics of programmatic

Questions over extremist content are only part of the conversation: what about B2C brands whose adverts for alcohol or gambling sites pop up on the screens of addicts? What’s the ethical stance there? How can brands be sure that this supposedly highly targeted, highly measurable approach to advertising meets ROI goals as well as moral standards? And how do customers feel about being pursued across the internet by brands that really don’t interest them?

According to Rory Sutherland, vice chairman of Ogilvy & Mather Group, much of the murkiness stems from the algorithms themselves. “Algorithms are very good at optimising a specific thing, but they usually have no idea of the wider context in which they operate,” he explains. “They thus might pursue something to a ridiculous degree where a human agent would give up. What concerns me is how programmatic might affect people with addictions. If a human being said: ‘Let’s advertise booze at AA meetings’, any sane and decent person would view them with horror. But an algorithm doesn’t have that sense.”

A vision of dystopian society emerges, where our vices – however trivial – are preyed upon by algorithms looking to deliver financial results. “In ethical terms, it’s fairly nasty,” says Rory. “You can envisage a world where everyone feels exposed and tempted by the thing they’re most trying to avoid. There’s a risk, I’d say, that algorithms become too good, and that’s a worry.”

The sceptics – what do marketers really think of programmatic?

So what do marketers feel about recent headlines? “Using programmatic is something we’ll consider for a few of our clients with really broad audiences, but I’m not a fan of it and it’s not our go-to strategy,” says Ebony Grimsley-Vaz, owner and creative director of Above Promotions. “You only have to watch videos on YouTube to see the lack of control in terms of where your products end up. Not enough technology has been created or implemented to give brands peace of mind, and in this day and age, a setback to a company’s reputation means a quarter’s loss or even more. You really have to evaluate if it’s worth it in the long haul.”

So the supposed lures of programmatic (wide reach, advanced targeting and increased measurability) haven’t won over Ebony. And Justin Hartzman, co-founder and CEO of social media ad agency Needls.com, shares the same opinion. To him, the whole practice is pretty shady: if it’s not a lack of clarity over cost (are you paying per click or per sign up?) it’s unreliable analytics and – in some cases – disappointing outcomes. “We used to do programmatic but we stopped recently because of the cloudy nature of what’s going on in the market,” he explains. “When it comes to the major demand-side platforms, the world of programmatic is so close-knit that it becomes tough as a buyer to figure out how it actually works.”

For Justin, the decision to pull budget from programmatic was a no-brainer: thousands of dollars were being eaten up in a matter of hours and the return just wasn’t high enough. “For SMEs without decent budgets, programmatic isn’t really an option. If you’ve got hundreds of thousands – or even millions – to play with, it’s a different story,” he says. “And then there’s the question of click fraud – is the provider in question delivering real clicks? What’s the measurability around that? And is your ad best suited to high profile websites or more targeted and less noisy websites where there’s more opportunity to be seen (and which cost significantly less)?”

For companies like Justin’s, social has proven a more fruitful channel for lead generation and engagement. Concerns over click fraud – coupled with the dawn of a new age of ad and pop-up blockers – mean he remains unconvinced of programmatic’s potential.

So it seems that the programmatic vision doesn’t quite marry up with the reality: while some see it as synonymous with low-cost, laser-sharp targeting, others might take a little more convincing. “Ad space sold programmatically is usually the remnant, unsold inventory that doesn’t command high prices,” explains Craig Gooding, co-founder of Vibrant Media, an independent consultancy for martech and ad tech. “This is because it’s not the same quality of that sold through the traditional inventory order process, where the most optimum advertising environments are achieved.”

The proponents

Someone who takes umbrage with Justin’s views on programmatic’s limitations for smaller brands is Neil Eatson, co-founder and CEO of Appraise Digital. “It’s not about the money,” he says. ‘It’s actually about the partner, tool, technology and data. If you understand the data you’re working with, programmatic’s a natural bolt-on, but if you’re just throwing money at it in the hope that it’ll work it’s a different story.”

Neil does admit that a lack of transparency has had a negative impact – something that isn’t helped by a limited understanding of programmatic on the part of some clients (which in turn isn’t helped by the over-complication of the programmatic process by vendors and agencies). “Programmatic’s currently seen as a dark art in the same way that search was 20 years ago, but the process itself isn’t as complex as it first appears,” he explains. “The fact there are so many moving parts – and it’s so fragmented – can be confusing, and what we need to do is demystify it.”

Another fully paid-up programmatic proponent is Craig. For him, the main benefit is cost – and data, of course – which allows marketers to hone their ad buying around particular objectives. “Ad buyers and planners now need to assess more factors when selecting ad space, but they’re able to predict the likely performance of ad space for certain campaigns, as well as evaluating it more effectively. The data is also starting to show that ‘premium’ providers of ad space don’t necessarily deliver premium ad inventory for every single impression: some ads will have poor viewability, be served within irrelevant or even brand damaging content, or be displayed to a totally irrelevant user. That’s not ‘premium’. Likewise, what a marketing manager may not initially perceive as a ‘premium publisher’ – maybe because it’s a niche site – could well be delivering premium inventory.”

Increased understanding of what a good ad inventory comprises has shifted perceptions around the purchasing decision: saying that an ad impression is ‘premium’ merely because it’s been bought from a particular publisher no longer holds much weight.

Putting the customer front of mind

And what about the consumers in all of this? Are we considering their needs? Or is this another case of ‘channel first, customer second’? We all know, for instance, that re-targeting is a risky game: while on the one hand it can gently nudge customers down the funnel, it can also really piss them off. “As consumers, we don’t want to be ‘stalked’ by advertisers who track and monitor our digital footprint: it’s creepy,” says Katy Halewood, head of Maxus for Business at Maxus. “Particularly as it’s often misplaced – after all, being served an ad for a pair of shoes you’ve just purchased doesn’t quite hit the mark.”

Rory Sutherland concurs, pointing out that programmatic approaches that are effective in the short term can ruffle some serious feathers in the long term. “There’s definitely a limit with re-targeting,” he argues. “It obviously works – if someone looks for hotel rooms in Torquay, and then are served up an advert for a hotel in Torquay it can encourage them to book. But if you’ve bought a children’s book on Amazon for someone and then find that Amazon infantilises your recommendations for the next month, you’ll start to go deranged. There’s a question of degree here: extrapolating short-term effect into long-term desirable.”

But re-targeting – or the test and learn approach, as it might also be known – does allow dynamic messaging to be used to great effect, as Katy points out. “You can target by geo-location, IP or demographic, and then reap the benefits of personalisation. There’s thus great opportunities for better and more effective ads, so long as the data is used carefully and adds value to the user.”

In the second of our two-part series on programmatic, we’ll explore what marketers should consider before plunging in headfirst and making an investment.