Sneaky Little Bastards

For years, the Truth and Reconciliation Commission here at Ad Contrarian World Headquarters has been determined to expose the unreliable nature of most of what you read about digital advertising. There are three reasons for this:

1. The digital industry is full of sneaky little bastards whose "facts" and "data" usually turn out to be either intentionally misleading, willfully incomplete, or stone cold bullshit.

2. The research industry, heavily dependent on the digital ad industry for business, is complicit and almost always spins its findings about digital in the most positive light. 

3. The pitiful trade press -- devoid of perspective or skepticism -- swallows this garbage whole and publishes it with a tone of gee-whiz boosterism that would embarrass a high school newspaper editor.

Which brings me to last week. My life is so empty and pathetic that I was actually reading Ad Age one night.

I came upon an article concerning the remarkable effectiveness of digital ads in magazine tablet editions. Here's what I read:

Starch Digital used online surveys to measure consumer recall of more than 13,000 ads in magazine tablet editions during the second half of 2012. We found that nearly 9,500 of those ads offered interactive features -- and that half of the people surveyed who read or noted those ads went ahead and interacted in some way.

I was amazed. I know that fewer than one person in a thousand interacts with display advertising in general. And yet, here we have a report of half the people interacting.

Soon I turned out the light and went to sleep. Except I couldn't sleep. My experience in advertising told me this report was highly unlikely. My experience as a former science teacher told me it was completely impossible.

I laid there for at least twenty minutes thinking. And then it struck me. I turned the light back on and re-read the piece. And there it was. To a casual, unsuspecting reader the impression was that half the people interacted with the ads. But what it actually said was that "half of the people surveyed who read or noted those ads went ahead and interacted in some way."

In other words, you had to either read the ad or "note" it (I think that means remember seeing it) to be counted. So if only 3% of people read or noted it, then the true level of interaction was not 50%, but 50% of 3%, or 1.5%.

Of course, neither Starch nor Ad Age told us what percent of people "read or noted" the ads because that would be way too open and honest. So we have no way to know the true level of interaction.

All we know is that, according to Starch...

...large numbers of consumers...interacted with ads for gas and oil drilling investments, a book about Proust, a store catering solely to runners and walkers, and exercises to improve one's golf game. 

Really? If that's the case, why don't you give us all the information so we can decide for ourselves whether the nembers are "large" or not?

But here's the thing. Even if they did give us the full picture it would still be meaningless. Their definitions were so absurdly biased in favor of "interaction" that the numbers mean nothing .

For example, some of the ads were video ads that started up automatically. By sitting there and doing nothing, you were counted as interacting when the video started on its own.

Some of the ads were very small and were essentially unreadable unless you expanded the ad. And once you expanded it, you were automatically counted as interacting. So the act of reading became de facto interaction. For people who expanded the ads to read them, there was no difference between reading the ad and interacting with it. When Ad Age said...

...half of the people surveyed who read or noted those ads went ahead and interacted in some way...

they had not done their homework. A great many of the people who expanded the ad to read it didn't "go ahead and interact in some way."  But they were counted as interacting. And get this --  of the people who were counted as interacting, the largest group was the group that just clicked or touched the ad to expand it.

Hey, Starch, here's an idea. Why not count clicking the "X" to get rid of the ad as an interaction? That should make your masters happy.

But wait, there's even more. Who was sampled to determine these numbers? According to the article, it was an "online" sample. And we know how online samples skew, don't we?

But wait, there's even morer.

Being fascinated by this exercise in deceptive nonsense, I went to Starch's PR release about this study to see what they had to say. Here's what I found...

More than half of consumers who read a magazine ad on their tablet or e-reader interacted with the ad, according to new research covering more than 30,000 digital ads across 1,000 magazine issues from GfK MRI Starch Digital.

The research shows that 55% of digital magazine readers “noted,” or read, a digital advertisement on their tablet or e-reader...

What's your take-away from this? If you are not thinking like a sneaky little bastard, one of your take-aways is probably that people read 55% of the digital magazine ads they were exposed to.

But that's not what it means at all.

What the language actually means is that during the 6-month test period, 55% of people ever read or noted a digital magazine ad. In other words, if you were exposed to 1,000 digital magazine ads during this 6-month period, it doesn't mean you read 550 (55%) of them, it means there's a 55% chance you read one of them!  

In my hundreds of years in the ad business I have witnessed more bullshit than I care to admit. But I've never experienced a more persistent and unrelenting effort to distort and mislead than I have seen from the digital advertising industry and its unprincipled lap dogs.