How to fix programmatic advertising: the whole story

The ISBA/PwC report into programmatic supply chain transparency, vital as it is, only tells half of the story. Here, Nick Manning looks at the other side of the equation

The dogs bark but the caravan moves on.

Certain stories may lose their immediate impact but shouldn’t be left to die. The Dominic Cummings saga is one and the recent ground-breaking ISBA/PwC study into programmatic transparency is another, even if both have been covered extensively before being eclipsed by other news.

The reaction to the ISBA study has been both significant and international, and rightly so. Yes, some of the findings matched previous exercises dating back several years but that doesn’t make them any less startling. It just shows that very little progress has been made.

The danger is that the repetition of bad news can induce complacency. It shouldn’t be considered ‘normal’ that 51% of an advertiser’s budget is used up before an ad appears even if it’s one we’ve heard before.

Some commentators have tried to use other supply-chain cost analogies but these don’t take account of the fact that online advertising is invisible to the advertiser (and often to the public), so the client often doesn’t know what they’re buying, why it costs what it does, whether their ads were seen and by whom and whether they had any effect. There are no ‘voucher copies’ in online.

Yes, identifying the 15% delta in unidentified costs is important but that isn’t going to happen overnight and meanwhile the ship sails on.

No other channel costs as much to sustain as programmatic online display, and as automation becomes the de facto way of trading, already including out-of-home, surely we don’t want it becoming the norm, do we?

Connected TV is arriving in the UK and there is a danger that the programmatic trading mechanisms of online display are reproduced, but with bigger budgets.

The ISBA/PwC report shows that only 12% of the sample impressions could be fully tracked; this was across the 12 premium publishers in the study, accounting for some 15-20% of the advertisers’ total activity. The 31 million impressions tracked therefore represent only some 2% of the total impressions bought by the sample advertisers during this period. A market worth billions of pounds is not being adequately monitored.

And this is hardly surprising when the average advertiser was using 40,000 websites (with one reportedly amassing 150,000 sites), across over 1,000 distinct supply chains (for 15 advertisers only). PwC could only match impressions for 290 of them.

In an open market environment optimisation engines should progressively narrow the number of sites as evidence of success emerges. This enormous number of sites suggests a ‘spray and pray’ approach that does not remotely justify the high costs of data-led targeting. It looks like the targeting and trading mechanisms that cost 49% of the budget are not doing what they are supposed to do.

Some serious questions arise; can a campaign of 40,000 sites be efficiently optimised, measured and reported on, including all of the key metrics for evaluation? Can advertisers be expected to know where their advertising appeared with such a long list of sites? How can anyone know whether these sites are worth anything or are even real? Isn’t there an inherent brand safety issue when using this number of sites?

How can the user experience be managed with this volume?

The ISBA/PwC study is remarkable in that it shows that just getting the right data is a real barrier; 40% of the impressions tracked could not even be ascribed as being static, video or rich media.

It’s not too hard to see why much of this happens. The media agencies lose control of the activity as it passes down the system and lose sight of the process; the publishers likewise in the opposite direction

The 15% delta identified by PwC seems to sit between the ad exchanges and the DSPs which agency contracts and audit rights don’t cover.

Clients whose contracts require full transparency should not experience this lack of visibility, but clearly even the UK’s biggest advertisers have not achieved this and may look in vain at their media agencies to do so. The media agencies are asked to police the entire supply-chain when they cannot control the whole process; their fees only represent one sixth of the total 49% costs (although other revenue lines may arise if the client contract permits) and they have limited control over the 42% balance. While the agencies sometimes have some control over the DSP and tech fees , their visibility is also limited thereafter.

The supply-chain is too lengthy and unmanageable, even for the UK’s savviest advertisers who took part in the study. The automated systems have been allowed to drive the market without sufficient control; automation has bought cost for advertisers, but not the precision targeting nor efficiency that is both possible and the point of programmatic. This is not how it was supposed to be.

[advert position=”left”]

Some re-selling and arbitrage have added extra complexity to the market, reducing transparency.

On the plus side the publishers have reduced cost through automation but have seen low yields as the volume of traded impressions has grown (partially fuelled by synthetic traffic). However some of their cost-efficiencies have been nullified through the low ‘take’ from the programmatic market after all costs are accounted for, including the average 14% commission paid to SSPs referred to in the ISBA/PwC report (which translates into 8% of advertiser budgets in the ‘waterfall’).

This isn’t just an arcane discussion about supply-chain costs but part of a much bigger picture. Advertising effectiveness relies on public exposure, and the loss of effectiveness observed in recent times may be a function of the dilution of budgets as more money shifts into programmatic and the proportion of ‘working’ budget lost to costs rises.

The shift to short-term tactics compared to brand-building may also be partly driven by this factor, given that performance-led advertising (especially in social channels) generally has shorter supply-chains, lower intermediary costs and focusses more on cost-effectiveness and ‘working’ money. The majority of the advertisers in the ISBA study are brand-focussed or a mix of brand and acquisition; pure-play direct advertisers have shortened their processes and reduced costs.

We shouldn’t lose sight of the fact that this loss of advertiser spend also means that publishers are under-funded precisely when we need the opposite. The closures and job losses are mounting in the paid-for publishing sector even as the audiences hit new heights. The loss of proper reporting helps fuel the user-generated media that allow the political extremists and conspiracy theorists to thrive. BuzzFeed, for example,  will be a loss for the UK, and some publishers, such as Bauer, have made significant cut-backs in their printed editions.

Programmatic is unquestionably the way of the future, with creative and media combinations providing stand-out and precision targeting, leading to demonstrably better results. But it requires a high degree of control throughout the planning and execution cycle, and this means more human involvement over the automation; less ‘set and forget’.

Programmatic does cost extra but the data-led and tech-enabled systems have to work super-hard in effectiveness to compensate for the costs. There is a reason why online display often performs badly in media mix modelling based on spend; the gross spend gets cut in half.

Imagine how much more effective online display could be if more of the money was seen on-screen. And, as I have previously written, it is highly debatable that much programmatic does what it says on the tin, but still incurs the costs as if it did. I see female fashion ads every day, several times a day and hardly any that seem to know anything about me.

Transparency isn’t just financial – it’s much more fundamental. The ISBA/PwC report, vital as it is, only tells half of the story. It did not aspire to look further downstream, at the sharp end where ads appear (or don’t). Its scope, which was hard enough, was to look at the financial aspects of the supply-chain and where the advertiser’s costs reside. The other half of the equation (based on 51% reaching the publisher) is another kettle of fish.

The supply-chain structures explored by ISBA/PwC exist because of the way that the programmatic market has evolved organically; it has been supply-led, with an explosion in publisher numbers and vast amounts of advertising inventory being offered to the market. The present systems and currencies have been created to manage this tsunami of apparent opportunity, with the demand-side creating its own structures to manage the increase in supply. The ‘upstream’ supply-chains reflect the chaotic nature of the internet.

It’s hardly an original observation that outside of reputable, well-managed publishers (for example the AOP members in the ISBA study) the web can be a murky place. The democratisation of content has opened a Pandora’s Box where legitimate websites and apps jostle for money with the nefarious. Legitimate publishers have the hard task of attracting people’s attention; bot-manufactured artificial traffic and clicks don’t have this problem.

True, valuable impressions and invalid traffic co-mingle with adtech systems struggling to identify the difference when trillions of impressions are flooding the market.

The vast majority of web and app content is perfectly kosher (whatever people may think of it) but there are parts which are just there to raid the advertising market. Spotting the real from the fake is achievable if the right systems are employed pre- and post-bid, but this doesn’t always happen.

Ad fraud is nothing new, but the extent of it and its various types continue to proliferate. The bad guys are always one step ahead of the good ones, and adtech has been perverted to provide new ways for grifters to ply their trade. The size of the prize is huge; the meteoric rise in online adspend has led to an increase in the types of fraud and the amount of money being siphoned off from advertisers. It’s drug money without the guns and white-collar crime usually goes unpunished. The fraudsters come in and go out quickly and cover their traces well.

Content piracy and synthetic traffic is a case in point. Megan Graham of CNBC in the US recently showed just how easy it is to raid legitimate content providers and set up websites that can attract advertising from reputable brands at incredible speed even though the content has been illegally purloined and there is no legitimate human traffic.

The IAB has done sterling work in attempting to head off the ‘bad actors’ with their ads.txt and sellers.json accreditation, but the scamsters are nothing if not creative in finding ways to game these initiatives. They benefit from the scale of campaigns as described above, where keeping track of 40,000 sites to ensure legitimacy is a big task, even when automated. Tricks such as domain spoofing can fall through the cracks given the vast amount of traffic flowing through the system.

Sometimes the fraudsters don’t even need to create artificial websites, as evidenced in the ‘404 Bot’ which took the money through an error page. Even a reputable publisher like Newsweek fell into the trap of allowing rogue employees to enable fraud on their sites.

The early indications are that the connected TV market is heading in the same direction.

The coronavirus crisis and the Black Lives Matter outcry has potentially given further ammunition to the wrong people. While legitimate publishers have lost out on revenue owing to keyword blocking (anything crisis-related), the fraudsters have no such qualms and will happily feed upon the public hunger for crisis news-the more sensational the better. Conspiracy theories are good ways to attract audiences, even if its for prurience, and the sites that carry the most outlandish ‘stories’ are rich pickings for fraudsters.

The Global Disinformation Index has estimated that European fake news sites earn around $75 million per year, and a number of household name brands have appeared on websites in multiple geographies (including the UK) that aim to stoke tensions and misinformation about the current crisis. These brands would claim they have robust brand safety measures in place.

The recent HBO documentary (After Truth: Disinformation and cost of Fake News) should be compulsory viewing for anyone in doubt about how public opinion is being manipulated. It’s easy to get from #pizzagate to #obamagate, especially if the US President’s 80 million Twitter followers retweet it repeatedly, even if they don’t know what it means.

The combination of a global virus, economic downturn, racial disharmony, Brexit and the US election has created a seething hotbed where a mix of legitimate (but often partisan) news and rampant misinformation drives record audiences and vast engagement scores as locked-up audiences vent their spleen or share the latest #NWO theories.

While advertising doesn’t in itself cause this, it feeds off it and can act as a source of funds for malicious behaviour. Ad fraud is a virus that thrives on confusion, and brand safety is compromised when the real and the fake become indistinguishable.

Meanwhile, viewability remains a challenge. The highly valuable WFA report (with data aggregated by Digital Decisions from the ad verification players) shows that the ‘in-view’ rate for UK display ads (based on the low bar of 50% of pixels for one second) remains stubbornly rooted at 61%, with only marginal improvements in both Desktop and Mobile since early 2018 . The fraudsters have even found ways to game viewability to make standard measures less reliable.

We can’t expect the public to search webpages for ads but we have to acknowledge that an above-average  viewability score is anything above 60%, so 51% of an advertiser’s budget reaches a screen but maybe some 40% of the ads that get through are not being seen in any meaningful way (generously taking one second for half the ad as ‘meaningful’). And with 40,000 sites to track, it’s not an easy task to know which sites have performed well in viewability and real, human traffic and optimise accordingly.

A reduced number of sites, better creative content and more considered placement can improve viewability and reduce ad fraud considerably.

And, just to stress the point, good impressions attract the sort of costs identified by PwC but so do the bad ones. Fake news, fake traffic, fake ads…it’s a jungle out there if the right control isn’t exercised.

Now, much of this is not new even if some of the techniques have evolved. What has changed, though, is the amount of money at risk. The UK is a world-leading market for digital spend with its share of total approaching 70% and ad fraud alone may now be now globally worth some $23 billion, funded by advertisers and no-one else. Data varies, but up to 5% of the total global advertising market may be lost to ad fraud. Somebody somewhere is getting very rich with little prospect of indictment.

Meanwhile the money siphoned off is eroding ad budgets and damaging effectiveness.

The good news is that the industry is co-operating to address these issues. The ground-breaking collaboration between the advertisers, agencies and publishers that enabled the ISBA/PwC report is evidence of real progress, supported by the IAB.

This level of scrutiny cannot come soon enough; the industry cannot afford to become complacent about the status quo. Some of the commentary post-ISBA/PwC suggested that the issues contained in this article are ‘priced in’ and advertisers have to accept that their CPMs and CPCs are as low as they are because there is so little visibility of the true composition of the inventory they buy. It’s not OK to flood the market with cheap goods and justify that with low prices.

If the industry doesn’t take control of the market there is a danger of a regulatory approach. The DCMS and the ICO are well aware of the vagaries of the market; while much of their scrutiny will be privacy-led, there is a growing drumbeat that could lead to a top-to-toe investigation. While some further regulation may be welcome if it rebalances the market, the scope of any enquiry would be unpredictable.

Cleaning up the industry is necessary, and the ISBA/PwC report shows that the extraordinary number of supply-chains will need to be rationalised if data and money transparency is to be achieved. Yet advertisers have the task of breaking down the costs of the supply-chain and then making sure that their ads go to the right places with the best possible chance of being seen, engaging the audience and/or being acted upon. They have to consider 100% of the process, not just the upstream supply-costs.

Programmatic can be simplified and clarified with the right people, process and technology. The first step is to create in-house expertise in programmatic media to complement your external resources and to achieve the necessary control over data and executional capability. This is a technical subject; one of the best post-study articles provides illumination while demonstrating how complex the architecture and vocabulary is.

If in-house experience isn’t possible, strong independent resources also exist.

The advertiser framework should include these elements:

• Treat the creative and channel options as indivisible, with optimisation of both against audience, context and device
• Place the user experience front and centre, with sensitive re-targetting and frequency settings constantly refreshed
• Select the most expert agency partners in programmatic media, separately from other media channels if necessary
• Interrogate all of the data used in the process and work to ‘data sufficiency’, not overload. Consider alternatives to log-level data collection if this proves too hard to achieve (as demonstrated by ISBA/PwC)
• Appoint an independent programmatic tracking specialist company to provide a transparent data provision service aggregating all platforms
• Use the most in-depth research and data tools to understand the composition of the inventory you acquire. Focus on audiences, not just impressions, and consider new currencies
• Take control over the data and technology infrastructure while providing log-in access to your internal and external teams such that everyone sees the same data. Be involved in the optimisation process
• Be highly involved in the selection and use of DSPs and adserving partners, with due diligence over operating practices. Consider alternatives to the ‘usual suspects’ with DSPs that offer a more curated and transparent process suited to your business and audiences
• Work directly with reputable and accredited publishers to ensure the right inventory is bought, however transacted. Limit the number of sites you use to those where you can verify appearance and delivery. Favour private market-places but across a carefully selected range of publishers in the open market who provide a variety of content suited to your audience and their interests
• Ensure understanding of how your publisher partners work with SSPs and Ad Networks and how they manage their processes and data
• Pay the right CPM or CPC, not just the lowest. Track your pricing through time and optimise accordingly
• Deploy the full suite of content verification tools in real-time throughout the trading process and avoid any sites that do not meet all the right criteria. Tighten your whitelists and blocklists, reduce your site numbers
• Construct contractual frameworks with external parties that maximise financial and data control and verification rights
• Measure, analyse and optimise throughout and only across a manageable range of sites; test, test and test again
• Work closely with the IAB and observe their protocols on best-practice in content, inventory and execution
• Employ measurement and verification partners who can guide you through your contractual needs as well as back-end audits

This is a lengthy list of actions to take but it’s immensely do-able with the right focus.

In summary, the ISBA/PwC report was about programmatic financial transparency through the supply-chain. This is important and comprises one half of the equation. The other half is, if anything, even more important in that it is where the consumer enters the equation and advertising exposure and effect can be managed well or badly, depending on the governance applied. Advertisers should take an all-encompassing view throughout the advertising life-cycle.

This subject must not be allowed to become overshadowed by other pressing issues. We must not get snowblind on unacceptable numbers and treat them as normal. As we ‘Build Back Better’, let’s make this a top priority.

—
Nick Manning is the co-founder of Manning Gottlieb OMD and was CSO at Ebiquity for over a decade. He now owns a mentoring business, Encyclomedia, offering strategic advice to companies in the media and advertising industry. He writes for Mediatel each month.

He’s just joined Twitter: @NickManning64