No Peaking.

Thanks to Brian Morrissey's Digiday article yesterday, I was introduced to the concept of Peak Advertising, as espoused in a working paper by Tim Hwang and Adi Kamdar of the Nesson Center for Internet Geophysics*. The core concept here is that online advertising is at a point of diminishing returns; like "peak oil," the paper contends, "peak advertising" portends a cratering of the economic viability of the internet itself. "Worryingly," the authors intone, "advertising is not well."

Yes, advertising and marketing are the economic engines of a free web. And yes, there are important questions of sustainability and value that need resolution. And they deserve deeper consideration than the pseudo-scholarship offered here.

This week's Drift is proudly underwritten by PubMatic. With PubMatic's platform, publishers have the ability to offer their inventory to over 400 global Demand Partners - ad networks, demand side platforms, ad exchanges, and agency trading desks - and have on demand access to all the software, tools and services they need to realize the full potential of their digital assets.

The working paper leads off with a whopper. "Consider the long-term trend: when the first banner advertisement emerged online in 1994, it reported a (now) staggering clickthrough rate of 78%." We're in trouble already. I was on the team at HotWired that booked and managed that ad and I can tell you this is pure fiction. Dynamic ad serving and analytics programs would not be invented for another year, so even if anyone at Wired or HotWired was counting clicks, there would have been no way to determine the denominator of the clickthrough fraction. The authors uncritically took this bogus stat from "The Law of Shitty Clickthroughs," a blog post by Andrew Chen, who had uncritically accepted it from someone named @ottotimmons.

I won't spend time here refuting the wildly inflated claims about ad blocking that the authors also cite, uncritically, as a pillar of their premise. (If you want more on that, leave a comment to that effect below.) The most damning thing about this paper and the conclusions it draws is the assumptions that ads -like oil - are a commodity and that there is a single global marketplace in which objective value can be determined. Those of us paying attention to more than the latest heavy breathing of investors can see that the state of programmatic, exchange-based advertising is already getting far more nuanced and subjective. Automation does not a marketplace make, and basing assumptions on such a premise is dangerous.

Digital advertising is not, after all, a pit of tradable commodities. It's as broad as an entire world of marketing options. In the actual bricks and mortar world, the price of ads in the Pennysaver or the effective CPM of a billboard on the Long Island Expressway don't drag down the value of a TV spot on Monday Night Football. The valuation of digital advertising is about a thousand different models across a dozen channels.

So why do Hwang and Kamdar insist on the false logic of market connectivity and the false value of clicks as a proxy for effectiveness? They anticipated these challenges on pages nine and ten of the paper, but their rationale is - in my opinion - soft. The real answer is that simple conclusions garner outsized attention. Simplistic arguments like these are the Kardashians of research. A sophisticated business full of smart, dedicated people warrants more thoughtful analysis.

(*Nesson Center for Internet Geophysics? Please search for it.)