Dead people don't engage - as engagement measurement shows invalid traffic

TLDR: Dead people don't get involved. This sentence is so logical as to be boring. But in the Matrix, the digital illusory world of click fraud and feigned performance data, the sentence is exciting. Because there, in the Hades of the black boxes, who can distinguish the living (humans) from the dead (robots)? We can't either. But still, nonclick-based engagement measurement can provide clues that reach purchasing is generating "visitors" who are obviously dead on arrival. Here's a recent case that points to a problem few know the extent of. In the aggregate, we're not alone in estimating, organized ad fraud affects 60-90% of all display budgets.

"Traffic" vs. "Engagement", an interplay that must be mastered

When a company spends money to build up reach for a landing page, you can expect something to happen there. But what is happening there?

Of course, the engagement rate drops when more visitors come. This is the eternal media interaction between mass and class. If you want to have a top engagement rate, leave the page to its developer and never publish it. Because even the first visitor shows less engagement (interaction) than the programmer. So if you want to lead thousands of visitors to a website with advertising, you won't be surprised if the engagement rate drops sharply. That's how it is: the bigger the reach, the lower the engagement (the attention, response, contact quality...).

No "traffic" without engagement measurement! How website performance must be evaluated

"Traffic" vs. "engagement" are two sides of the same coin. When advertising budgets come into play that are supposed to generate reach, you should take a very close look at what the arrivals on a website are doing. If purchased visitors bounce right back, it was all for nothing, of course. The bounce rate, the standard key figure of every analysis tool for inactivity of the user, is easy to read.

Unfortunately, however, the ad fraud experts, the admins of the "walled gardens" and the "metrics" shell players also know this. They run simple programs on their AdTech server farms that trigger billions of clicks (click bots) just as they were booked. The click has taken place, the advertiser pays on the basis of reports that "prove" this, but a media service has never actually taken place. (For the digitally naive: media performance is communication contact with real people, fully analog).

Now, any computer science student can put together click bots in the first semester. And that's why it's no problem to have these scripts simulate not only clicks on the advertising medium, but also clicks on the advertised websites (landing pages). This means that even the bounce rates measured by the advertiser's own systems can be manipulated with simple means.a bot mimics the visitor. This means that the bounce rate cannot be used as a control for the quality of the purchased "traffic". But what is the control value then?

The answer: a web analytics metric that makes a statement about engagement, but is not based on counting easily simulated clicks. That's exactly how our engagement rate works, as we apply it to EVERYTHING according to our Content.ONE standard, including websites. The engagement measurement on the website also enables a reliable statement there about how attractive the content of an advertised target page was for the visitors, no matter whether text, app or video.

If this key figure is continuously collected, it is easy to derive empirical values as to how high the average engagement on advertised pages is. And then one of our clients runs a larger campaign and sees the following based on his own benchmarks:

A real case from July : Pay six figures at Doubleclick (Google) and get almost nothing.