(2/3) Previously, we have shown how Google Analytics flaws put marketers’ knowledge at risk. Fraudsters have always been able to create “ghost traffic, ghost clicks” and even “ghost sales” – that is, falsely claiming credit for sales that have already occurred. Collectively, these have at best cast suspicion on the accuracy of marketers’ information; and unfortunately, they have consistently tricked marketers into allocating more ad dollars to scammers through programmatic buying channels, due to the illusions of “better performance” (more clicks), “better profitability” (price lower) and “more reach” (larger quantities). If you thought it was bad; you should definitely sit down for the rest of this article.

What is euphemistically called the “murky programmatic advertising supply chain” is a hundred times worse than the “problematic” analysis above. The Association of National Advertisers (ANA) sums up the problem well in its recent public tender, “The programmatic market has been riddled with hardware issues, including a lack of transparency (data and dollars), fractured accountability, and staggering complexity. These issues hamper the critical decision-making of marketers, leading to unnecessary and unproductive media buying decisions and crippled and meaningless analytical abilities. However, despite the additional details sought and the expanded size and scope of the proposed study, the amount of hidden trade-off, waste and fraud plaguing the programmatic supply chain will not be properly captured in simple graphs. cascade of what happened where. Let me explain.

Previous studies have been summarized in simple waterfall charts

The above slide shows three previous industry studies – ANA (2015), WFA (2016), and ISBA (2020) – all neatly summarized in simple waterfall charts. The totals and averages seem to show very similar conclusions – that about half of every dollar spent by advertisers goes to non-functional costs; and only 50 cents are spent on serving ads. the most recent ISBA study also revealed that an average of 15% of the money invested in the programmatic supply chain could not be counted. This euphemistically named “unknown delta” accounted for almost a third of those non-functional costs – 1/3 of the money that went to programmatic ad technology providers is gone. In addition, the average of “15% missing” masked the more frightening truth that the amount of money missing was 86%, “with the majority from 2% to 23%”. As a reminder, in a programmatic campaign, 86% of the money could not be counted. Many other key details of the study are often not cited. For example, 1) only 12% “of the 267 million scanned impressions could be matched successfully [through the supply chain from advertiser to publisher], 2) “only 19% of impressions were generated at the study publishers; advertiser ads appeared on average on 40,524 websites, most of which were not premium ”, and 3) even though the study only involved 15 advertisers and 12 publishers, data access was found to be extremely difficile, in addition to poor data quality, different types of data (aggregate vs. log level) and inconsistencies in data formatting.

Important details obscured by averages

Much like quoting a “missing 15% average” obscuring important details, waterfall charts do not adequately describe the diversity of risk, cost and fraud that pervades the programmatic supply chain; they will also hide more important details that are needed to understand the issues and inform concrete actions.

The right slide shows how the overall average click-through rate (9.4%) masks click fraud that would have been evident if line item details had been used instead of the average. The high average might even make marketers think that this channel works significantly better and inadvertently allocate more budget to it. The programmatic channel has remained “non-transparent” by design, because the parties that benefit the most have a vested interest in keeping it that way, so that they can continue to make money. This is one of the reasons why previous initiatives like blockchain, designed to increase transparency, have failed due to lack of adoption – that is, parties that needed to adopt it in order for it to be adopted. it works preferred the current, less transparent state.

In addition, while every conclusion of the ISBA report is correct, some of the recommended actions require significant retooling (new analysis and monitoring), the cooperation of many ad technology providers in the supply chain (standardization of types and formats data), and re-negotiation of agency and supplier contracts (e.g. addition of audit rights, data access, access to the dashboard, etc.). All these current obstacles; any delay in adopting them will mean that non-transparent practices will continue and money will continue to be wasted. Are there concrete actions marketers can take without retooling, standardizing, and renegotiating large contracts? Yes. And these will seem so simple in hindsight.

Even before the money reaches the supply chain

Before we get to the waste and fraud in the supply chain itself, there are some hidden issues that are hard to depict on a waterfall chart. For example, how can we show “undisclosed principal trades” on the waterfall chart? The main undisclosed trading is where the agencies, through their holding company trading desks, take possession of a $ 1 CPM ad inventory and sell it to their own clients for a CPM of $ 10, without disclosing the mark-up. Instead of acting as an agent, the agency acted as a principal – like a grocery store that buys oranges wholesale and sells them to retail customers where margins are not disclosed. But in the case of digital media, agencies are supposed to act as agents, buying media on behalf of their clients, not arbitrating their own clients with undisclosed margins. Without the right clauses in agency relationship contracts, nothing prevents agencies from doing this or obliges them to disclose the practice. This throws one of the many keys in the waterfall chart. This is one of the elements contributing to the “unknown delta”.

Also, how do you account for “unspent media dollars” – the difference between planned and actual spending? It’s pretty easy when it’s only a small percentage. Most agency contracts provide for amounts to be carried forward to be spent the following year. But what happens when 450 million monthly impressions were planned, but only 150 million impressions were actually purchased and served (not a theoretical case, but the numbers have been changed to protect the identity of the parties involved)? Where does 2/3 of the unspent media money go? Advertisers can’t get that much money back and figure out where else to spend it; and media agencies cannot afford to generate that much lower income. It has been documented repeatedly over the years that simple accounting tricks in Excel spreadsheets have allowed agencies to keep unspent amounts discreetly, with or without the knowledge of their clients. While the extent of this practice is not known, it is certainly convenient for the parties involved – advertisers do not have to take back large sums of money and agencies can easily integrate it into greater profitability of businesses. trading desk operations.

In the next article in this three-part series, we’ll cover more examples of fraud and waste in the programmatic supply chain itself – some of them are illustrated in the following slide presentation. If you are a marketer waiting for the next article to be published, in the meantime ask yourself if this is more than 1/3 of the money that has gone to the middlemen of “disappeared” ad technology or what 1/2 of every dollar you spent went to those ad tech middlemen and not to serving your ads.