Informed Guesswork

23 February 2026

The great Bob Hoffman referred in his AdContrarian blog to ‘informed guessing’. His point being that despite wrapping ourselves in scientific-sounding words and phrases there is precious little in the advertising world that a qualified scientist would recognise as ‘science’.

What there is, what we do, is informed guesswork.

How could it be otherwise? Circumstances change; people act unpredictably; moods swing; ad qualities vary; ad contexts fluctuate from the deeply inappropriate to the highly relevant. Advertising is not an industry conducting regular scientific experiments under laboratory conditions.

The more information you have, and the clearer your objectives the better the chances that your guesses will pay off.

 Here’s a simple choice. Would you prefer to focus on your ad being seen by masses of people, or to take actions that contribute to you selling more stuff?

It’s a choice because these two things are not inextricably linked.

The ‘masses of people’ who ‘saw’ your ad may not actually have done so. They might have had an opportunity to do so, but that’s quite a different thing.

And when we say ‘saw’, or even ‘opportunity to see’ we don’t mean ‘notice’ let alone ‘remember’. What we mean is that a pair of eyes were (possibly) on a screen for a second, maybe two during which an ad may have played.

Did it play? If we’re talking online, we think so, but we’re not sure. Nobody will guarantee anything beyond saying the likelihood is the ad was served. Was it all played? Shrug.

Can we say where and within what context the ad appeared (if it appeared)? No. We can guess but we don’t know.

Furthermore, when we say ‘masses…’ we’re not sure how many masses. And when we say ‘people’ we more accurately mean devices. And when we say ‘devices’ many of those may be phones racked up in their thousands and programmed to click away automatically.

So, the first part of your choice might more accurately read: … for your ad to have been served (not necessarily at all, let alone in its entirety), but has not necessarily appeared on many (we’re not sure how many) sites (we’re not sure which), on an unvalidated number of devices some of which (but not all) may or may not involve a pair of eyes (we don’t know how many) open (we hope) in front of the screen for a second or two.

I once wrote a chart for an MRG Conference paper. Goodness knows where it is now, but from memory it contained a pyramid the base of which read: ‘How many people saw any edition of a magazine used by my client?’

The next step up changed ‘any edition’ to ‘an average edition’.

Then moving up: ‘How many people saw the actual edition of a magazine in which my client’s ad appeared?’; and on to ‘How many people saw the actual page on which my client’s ad appeared?’.

There were several other stages until at the top of the pyramid we got to ‘How many people saw my client’s ad and acted upon it?’

My point was we were stuck on the lowest segments of the pyramid.

This was in 1980. 45+ years ago.

Now replace ‘magazine’ with ‘website’ and change some of the obvious descriptors.

At least back then we knew that our ads had appeared, where, and the surrounding context. Now we don’t.

Back then there was one unquestioned source of audience data. It may have been guesswork but everyone had the same base. No longer.

Remember the digital evangelists sneering how we have so much more audience knowledge today than back in the ‘dark ages’?

On 23^rd January, ‘Private Eye’ reported, in the context of several states considering banning under 16s from social media: “Both META and Snapchat have said there is a significant margin of error when seeking to determine whether a user is in fact under 16.”

And yet the platforms sell on the basis that they can tell advertisers exactly who their ads reach, on a one-to-one, minute-by-minute basis.

The foundation of advertising is to be Legal, Decent, Honest and Truthful. Those words are important; over time they’ve shielded advertising from external regulation.

The ASA builds on these foundations:

• Truthful: substantiate all objective claims.
• Non-misleading: must not deceive, omit or mislead through exaggeration
• Social responsibility: must not encourage illegal, unsafe or anti-social behaviour.

Shouldn’t these foundations apply to the collection and use of online audience data as much as to any other advertising discipline?

There is plenty of evidence on which media forms deliver consistent business results for advertisers.

Guesswork informed by evidence beats buying the salesman’s patter. Without question.