Display Advertising: Why it Doesn’t Work

When’s the last time you bought something because you saw a banner ad, like this one?

Assuming you’re not one of the 45 million Americans using an ad blocker that is. I mean really bought because of the ad, like you didn’t know about the brand, then you saw it in a banner ad, so you went to the website or walked into a store and bought it.

What percentage of the time, when you click on banner ads, is it on purpose? How much time do you spend trying to get them to stop blocking the article you’re reading?

When’s the last time you actually looked at a banner ad and digested what it said? Not even clicked on it — just looked? If you do look, do you give the product some thought, or do you get annoyed?

So given all of that, why does display advertising exist? Why do US companies spend $25 billion per year on it?

Display advertising companies lie with numbers really well.

In traditional advertising (TV, print, etc.) there is no way to directly know whether someone buys a product after seeing an ad. The only way to evaluate ad effectiveness is to look at the change in sales during and after ads were run: an inexact science at best.

The internet brings much more data into the picture: in theory a good thing for marketers. But data can be used to obscure as well as illuminate. The primary way digital ads are evaluated is by tracking users across the internet. As you probably know by now, scary internet companies know what you’re doing online, down to the last click. Because of that, it’s relatively easy to tell that you saw an ad for a product, then went to the brand’s website and bought the product.

Great news right? Yes and no. There are a lot of ads on the internet. You might see digital ads in multiple places before buying a product. If that’s the case, which ad should get credit for the purchase? Plus, what if you heard about the product somewhere else, and just happened to see a bunch of ads before buying it?

This is a really difficult problem for marketers. Several considerations go into measuring which digital ads truly generate incremental value for advertisers. Marketers can use an attribution system that assigns value to ads based on when they were shown in relation to the purchase and whether the user clicked on them or not. Digital ad attribution alone is a multi-million dollar industry, as many technical hurdles need to be cleared to pull together data from different sources. Even with complex attribution systems, marketers have at best an inexact picture of what’s working.

But those are good faith efforts. The nature of online attribution data makes it vulnerable to distortion and abuse. The most egregious case of this is that of the display retargeting companies: AdRoll, Criteo, and Steelhouse. More than $1 billion is spent on ads served by those three companies every year.

Here’s how it works:

• You go to an e-commerce website
• If the website works with Criteo, AdRoll, or Steelhouse, a piece of code on the website fires, letting the retargeter know you’ve been to the site
• Criteo, AdRoll, or Steelhouse shows you ads (these are the ads following you around the internet that people find so creepy)
• If you return to the website later and buy something, the retargeting company knows that you saw an ad, and therefore takes credit for the purchase

The results in an extreme overestimate of the effectiveness of display retargeting. It’s common for people to visit a website before making a purchase, and it’s easy and cheap to show people display ads, so retargeting partners often end up taking credit for a large portion of a retailers’ revenue and surfacing a very efficient return on ad spend (revenue/ad spend, the most common metric used to evaluate ad effectiveness)

So display remarketers are able to “prove” that their ads are getting people to purchase product and justify more ad spend, even if common sense reflection on user behavior makes it clear that — more than anything else — display ads just annoy people.

It’s not fraud (although Criteo and Steelhouse have certainly been accused of that too) but it is certainly misleading. In many cases, display retargeting companies will display their return on ad spend without any contextual information about how it’s calculated. (Unlike the real ad giants — Facebook and Google — Criteo, Steelhouse, and AdRoll take credit for a purchase even if the user did not click on an ad, and even if the purchase takes place several days after the ad exposure.)

So, what to do? If you ever start and brand and want to advertise it, don’t bother with display retargeting. If your customers want to hear from you they’ll give you their email; use that to keep in touch with them. They’ll appreciate it more than banner ads.