Digital Advertising doesn’t work.

This is the conclusion you could get from recent Freakonomics podcasts and the case studies of P&G, Uber and Chase that Dr. Augustine Fou has summarized in his Forbes article here:

Of course, it would be a wrong conclusion – in general. And for those who didn't click, also Dr. Fou doesn't draw it ;-). Our client work at Analyx has shown that #SEM for example has a 1.5-3.5x higher short-term(!) effectiveness than #TVAdvertising, depending on the category (display ads do worse). And for recent scientific proof, consider the fantastic research of Prof. Evert de Haan, Prof. Thorsten Wiesel & Prof. Koen Pauwels here:

Yes, there are many pitfalls in digital advertising. Just to name a few:

• There are #bots: https://bit.ly/3p1gXKC
• There is #AdFraud: https://bit.ly/39MTWor
• And: it is difficult to prove that your digital ad convinced an individual consumer.

But this doesn’t mean, digital advertising doesn’t drive sales. The lessons for #CMOs are rather different in my view:

(1) Don’t get seduced by giant impression and click numbers: Bots can click. Sales impact is what counts. As one of my CMO clients always says: "The goods have to go!"

(2) Actual modeling is a must: #Lastclick is NOT a model! It is a rule and actually often wrong. Without a proper marketing mix model you cannot isolate effects of channels.

(3) Forget individual attribution and focus on aggregate response: In other words, do my weekly sales go up following a marketing activity? This is true at least for the big questions like strategic budget allocation:

“Chasing individuals around the internet produces more data than can be managed, yet not enough to solve the problem.” https://bit.ly/390uehf

(4) Enrich your data with controlled experiments: By ramping your digital spending up and down and shutting sites on and off you not only get smarter on what works or not (https://bit.ly/3nZ06H5). You also enrich your next round of modeling by generating variance which your model can learn from and get better. This is an essential part of what I call #AgileBudgeting, for example here:

(5) Long-term effect can outrun short-term effect: According to our most recent client work at Analyx, TV is still the channel with the strongest long-term sales effect relative to its short-term effect. In other words: TV is still your channel of choice for brand-building. Interestingly, #YouTube is closely trailing TV by now (Maybe no wonder given that in the US more than a third of views are on #BigScreens as Shane O'Leary has pointed out here:

So at least in brand building, digital ads are well behind despite their strong short-term impact on sales. More on that in one of the next #WednesdayWhizz posts…