Friday, November 15, 2019

Brian Dunning was not just a fraudster, he was ineffective;
now, who will tell Sharon Hill and others?

That's one of several findings from this long piece about online advertising

It's snark-heavy, with a headline of "The new dot-com bubble is here: It's called online advertising."

A key early point of Jesse Frederick and Maurits Martijn is that here, in the most dismal of the social sciences (advertising as part of economics), as in other sciences, correlation is not causation.

From there, we dive into some actual research, which the hand-wavers didn't.

Finding one? Paid company brand name keyword links? Bupkis.

We then move beyond that to:
The benchmarks that advertising companies use – intended to measure the number of clicks, sales and downloads that occur after an ad is viewed – are fundamentally misleading. None of these benchmarks distinguish between the selection effect (clicks, purchases and downloads that are happening anyway) and the advertising effect (clicks, purchases and downloads that would not have happened without ads).
Interesting.

Now, that's all true of old-fashioned ads as well.

From there, the authors talk further about "selection effects" (i.e., selection bias) vs "advertising effects." And they apply this to the Hucksterman Empire.
In seven of the 15 Facebook experiments, advertising effects without selection effects were so small as to be statistically indistinguishable from zero.
Well, that's pretty serious.

From here, the authors note that this also shows advertising can't manipulate people as much as digital advertisers claim.

The information above is as true on affiliate marketing as on search, the authors show.

And, given that much of the research on both search and affiliate that the authors cite is about eBay, the specific company that unrepentant pseudoskeptic fraudster Brian Dunning pled guilty to defrauding? Schadenfreude is sometimes a semi-Nietzschean recurring bitch.

Now, who will tell Sharon Hill?

No comments: