Imagine Don Draper sitting back in his chair, looking out over Madison Avenue. His team of copywriters sit behind him, awaiting his next iconic ad idea.
“It’s a medium-sized square. Inside it a squiggle of a woman. One where she’s clutching her belly fat and looking down at it in distress. Some copy above it reads: ‘One Trick.’
‘One TIP to a Flat Belly.’
And it needs to be a Weird, Old Tip!”
That never happened. But the ad did:
If you’ve spent any time on the internet you’ve seen that ad or something like it. You’ve also seen ads like these:
And, if you’re old enough, you remember this gem:
As the internet has vastly democratized content publishing, so has it the ability to make and distribute ads. And, as ad networks and publishers accrue more and more data, ads can become more accurate. So, with all this increased sophistication, why have ads gotten so weird?
First, consider the modus operandi of advertising: to get your damn attention. If no one is looking the ad has effectively failed. The Mad Men of Madison Avenue certainly made it all look very good; it all had a certain sheen to it, a certain sophistication. But where advertising may have once been more about art, it is now almost exclusively about utility.
Clicks. Click-throughs. Opens. Impressions. Plays. Interactions. These metrics didn’t exist in the late 50s. Now the ad agencies live and die by them. Every ad must produce as much revenue as possible while maintaining a conservative bottom line, usually amidst an increasingly shrinking budget.
Another factor has also played into this uglification of advertising: A/B testing. One ad in particular got me thinking about this topic. Its an ad I saw for the Robinhood financial/brokerage app on Instagram.
The comments to this post were understandably “wtf” and “lol who screwed this ad up.” But one comment stood out to me:
“ah this must be the pic that won the ab test.”
Robinhood could have screwed up, sure. But I’ve seen this ad numerous times over a series of weeks so that’s not really feasible. What is feasible is that a company that makes it’s money providing a service that has been around for centuries just works. There is no need for prose, beautiful imagery, or striking fonts, it just needs you to read that one line of text below the photo and, hopefully, click the Download button.
A/B testing has led to incredibly productive advertising, web layout, and design. For example, the homepage for Barack Obama’s 2009 presidential campaign experienced a 40% increase in donations after the team A/B tested a couple different photos, slogans, and button commands. Without A/B testing, the campaign would have chosen, on instinct, a photo that woefully under performed testing.
A/B testing allows Netflix and Amazon to constantly experiment with new page layouts, optimizing for those that get us to pay – or play – the most. There is a good chance that when you and I log into Netflix this evening that we will be served very different layouts. A/B testing allows companies to test display ads across a vast network and quickly whittle down their options to the most effective. Another, albeit less sophisticated, example of A/B testing is the recent crop of companies creating several sub brands on Instagram and watching to see which accrues the most likes. Based on the winner, a brand may create a new product line or introduce a new sub brand.
Obviously, the gains to be had are quite clear. But how did we end up with advertising that’s so weird? Well, they result from techniques practiced by the likes of The New York Timeswhen it comes to crafting eye-catching headlines as much as they do the techniques that allow Amazon to get you closer and closer to the Buy Now button. They result from A/B testing.
Combine this A/B approach with the proliferation of do-it-yourself social media advertising platforms and increasingly money-hungry online publishers and you’re left with Facebook feeds filled with bizarre ad copy and online articles that dedicate over a 1/3 of their pages to just inane display ads:
While personally this doesn’t bother me, I have to wonder if this is the logical conclusion of advertising. Have we reached the quantifiable limits of how effective a picture and some copy can be? Will this ultimately result in a return to a more antiquated, stylized form of advertising?
Maybe this is the end of advertising as we know it. Commodities, things like paper towels and soap, will jockey for your attention with discounts and promotions while more elastic items, like shoes or cars, will be promoted exclusively by influencers.
Whenever I catch myself maligning an ad like the aforementioned Robinhood one, I have to remind myself that my commentary is actually evidence of the ad’s efficacy. I’m still talking about the brand paragraphs later. Maybe tomorrow you’ll bring up the app to a friend, subconsciously reminded by this blog post. That friend may in turn download Robinhood, place a trade, and incur a brokerage fee. What kicked off this chain of events? An A/B test told someone in marketing that the blonde girl performed better than an alternative photo that I’ll never see.