One common premise put forth by critics of digital advertising is that it simply doesn’t work. This argument is usually encapsulated in some variation of: digital ad platforms traffic exclusively in junk inventory that doesn’t produce marginal value to the advertiser, and these platforms merely claim credit for sales, revenue, engagement, etc. from which advertisers would have benefited, anyway, without buying the ad impressions.
The crux of this position is that digital ad platforms with extensive enough reach can paper the surface areas of their products with ads such that the proponderence of subsequent conversions is merely a coincidence: expose enough ads to a user and sooner or later that user is going to purchase one of the products advertised to them. This is a sort of adaptation of the “a stopped clock is right twice per day” idiom.
If it is true that digital advertising is a shell game, then by definition the $600BN in annual digital advertiser spend must be at risk of collapse. Proponents of this position have been claiming as much for years in frantically-titled articles like The End of The Online Advertising Bubble (2016), The new dot com bubble is here: it’s called online advertising (2019), The Ad-Based Internet Is About to Collapse. What Comes Next? (2020), and Is online advertising about to crash, just like the property market did in 2008? (2021). These articles — and a search for “online advertising bubble” surfaces many more like them — all propose that digital advertising, broadly, doesn’t produce what advertisers would call incremental yield for the companies that purchase it, and the market for digital advertising is akin to a Ponzi scheme that will ultimately collapse violently. Obviously, that has not yet happened.
Most large advertisers would contend that they are capable of measuring the incremental value of their ad spend, and that their armies of data scientists can quickly identify the ad campaigns, ad creatives, and targeted audience segments that produce real marginal yield for them. I was once asked to give a talk about a tool I had built to the marketing data science team at a large consumer technology company — which deployed a nine-figure annual digital advertising budget. I expected an audience of a handful of people; I was surprised when I was ushered into an auditorium into which over 100 people streamed. Would all of these smart, credentialed people really be unable to determine if their ad spend was wasted on utter fraud? Were they being completely bamboozled?
My rebuttal to the “most digital advertising is fraud” contention is straightforward. Advertisers with massive budgets can employ large teams to assess the efficacy of their ad spend, but small advertisers can’t, and Facebook’s ad revenue is comprised predominantly of spend from small advertisers. But similarly, these small advertisers can’t afford to spend indefinitely on non-productive advertising campaigns: they’d go out of business. And there’s some minimal level of scale — in terms of brand awareness, user base size, ad spend, etc. — where the threat of cannibalization of revenues from advertising even becomes relevant.
If a start-up with a minuscule customer base spends money on advertising and experiences positive results, then that ad spend is unlikely to have cannibalized revenue the advertiser could have expected anyway. And if that same advertiser sees no results, they will stop spending their precious cash on advertising. How does a company achieve early scale through advertising — which many do, to the point that VCs started grousing about the colossal share of their investment cash that portfolio companies spent on Facebook ads — if that ad spend, in the main, is wasted on fraud? If a start-up scales ad spend from $0 to $500k or more per month and DAUs or engagement grow at a similar scale with profitable unit economics from a similar Y-intercept, it seems unlikely that the ad spend was entirely absorbed into a black hole of fraud (the fact that some of it was lost to fraud becomes incidental).
Occam’s razor applies here. Could a marketplace on a trajectory to reach $1TN in global annual spend in this decade, in which nearly every consumer tech and retail company on the planet participates, really peddle nothing but fraud? If so, how could any company without existing brand recognition or organic traction survive? Does there exist a conspiracy that keeps money flowing into this opaque vortex of imaginary outcomes despite widespread knowledge that it produces no economic value?
It seems impossible to defend the position that digital advertising spend is wholly or even substantially subject to fraud across delivery or reporting. Even while acknowledging that fraud of course does exist in the ecosystem.
Recent