Most people don’t like display ads. Is that a good argument for ditching them? In this case, yes.
In advertising, we can’t get caught up in popularity contests; we have to be rational and honest with ourselves and do what’s best for the brand.
People have always been critical of different advertising mediums throughout history, but the advertising industry always felt the business results were worth the criticism.
If the general consensus is that something really sucks – or more importantly, the data actually supports that notion – maybe it should be dropped. It seems like the digital economy was built on 300x250 banners, but do they have a place in its future? In my view, no.
This notion isn’t new. P&G, Uber, and Unilever have all made headlines for drastically reducing or cutting display spend altogether.
Here are four reasons why I think it’s time for everyone else to ditch display ads:
advertisement
advertisement
Performance. I suppose it’s cliché to reference John Wannamaker in an industry-targeted article, but despite astounding advances in tracking technology, we still don’t seem to know which half of our advertising actually works and which half doesn’t.
Last-click attribution is clean and simple from a tracking standpoint, but it’s lucky if a display campaign gets a small fraction of a percentage point click-through rate (CTR). I’ve been told a campaign with a .05% CTR was a high performer, which is frankly delusional.
Some websites push view-through metrics because they pump up numbers and help justify the “display ads are reserved for the top of the funnel” narrative, but this viewpoint lacks a significant amount of context. Did a view-through conversion come from someone who was going to buy anyway, or did the system count an impression that loaded but was not actually seen?
I’m not even going to get into the broader discussion on fraudulent clicks, I’d just like to note that display ads are so ineffective that spam bots can’t even produce fake results that look good.
User Experience. If bots aren’t interacting, are people? We’ve all tried to load pages that are completely bogged down with ad units, exiting before getting a chance to do what we came there for. Imagine this on a scalable ad network: millions of publishers with free access to programmatic ad units can place banners of varying sizes on their site with little restriction.
More ad units mean more revenue per page load, but more ad units also mean slower (and therefore, less) page loads. The ideal balance is difficult to achieve, and there is no standard method for doing so.
While Facebook, Google, YouTube, and other standardized platforms are not without their problems, they do offer one thing that display units do not: predictability. If I click on a YouTube video, I know I’m going to have to sit through at least 5 seconds of pre-roll and I’ll put up with it since I know what’s coming. Marketers need some say over the user experience to know how to adapt their ad strategy, but display networks offer no such control.
Audience Targeting. Researcher Nico Neumann analyzed programmatic targeting data against verifiable information about individuals in the targeting pools and found anywhere from a 10-to-85 percent discrepancy across different sets. A 10% discrepancy isn’t bad, but an 85% discrepancy is unusable. This really comes down to algorithmic transparency. With display ads, there is none. Buyers cannot verify audience matching. We have to trust the sellers. Performance, user experience, and bad targeting all diminish the value of advertising, which leads to the last point.
The Economics Are Broken. Websites rely on revenue from display advertising, a product whose value keeps decreasing. This incentivizes creators to write for eyeballs, not exclusively for integrity. When the value of content decreases, the fundamentals of advertising effectiveness – targeting, engagement, the ability to incite action – decrease as well. The stats are weak, we lack control over the experience, the audience targeting is inconsistent, and the value is dropping.
It’s time to finally ditch display ads in our marketing efforts.
There are absolutely horrid websites that bombard you with ads (yes, USAToday and CNN.com I am looking at you); however, more personal-interest based publishers tend to see better results thanks to contextual relevance. On one hand, there are certain agencies and brands that recognize the importance of contextual targeting to reach specific audiences; however, during my 20 years of selling and buying online media, I can count on one hand how many times the kind of creative thinking was put into display ads as agencies put into video ads.
The discussion always seems to revolve around CTRs and attribution - which is absurd. If the creative is relevant (and repetitive), how many times does someone have to click on the ad to know it's effective - if they have to click on it at all.
It would be great if these analysis would focus on creative executions instead of dated metrics. The most important element of the ad (creative) seems to always be an afterthought when it should be the first thing we analyze.
Agreed, Dan. The metrics most commonly used to gauge ad effectiveness for all kinds of digital ads ---CTRs, for example----represent a tiny fraction of the audience--like less than 1%. What about the others? How many of them are even present? How many looked at the ad? What went through their minds when they saw---or read it?To what extent did it influence them---along with prior and future exposures to the same ad campaign---to eventually take the desired action and buy the advertised product?
The same questions apply to traditional media and, as with digital, they are generally not dealt with because it's easier for the media ad sellers and the buyers to deal in potential audience not ad impact metrics. And they are probably right about the latter as it's not the medium's job to convince the consumer that an advertiser's ad message should be heeded. But more and more people are realizing that we need something more than audience potential. What's needed is an indication of attentiveness---that people saw or heard---the ad message. Whenever a media seller cloggs up the ad page with lots of ads---or presents ten TV commercials in a row in a commercial break---that has a negative bearing on ad attentiveness. In like manner, boring, ho hum content generates low atttentiveness levels for ads while more relevant or engaging content generates higher attentiveness rates. Why aren't such factors given more consideration in media buying and selling?
Thank you Dan - reading your comment at least calmed my blood pressure. I've read so many of these articles like this that have clearly only skimmed the headlines of the trades and never actually built out an attribution model or dug deeper into the issues at hand. Patrick, no disrespect, but when you say "Ive been told..." - it tells me that you're not writing from experience. Rather than highlight what I have issues with in this article - I just want to add to Dan's note... Media buyers have to rely on volume, because most creatives can't be bothered to design a banner ad. It might be budget constraints - but let's be realistic - we can't expect to stand out with an ad that looks the same and speaks the same as every other ad. As a media buyer - you can target until your heart is content utilizing contextual and marrying it up with 1st party data and speak almost one to one with your audience, but if you continue to not have the creative speak to the audience, you'll continue to see horrible results.
I do want to address view through data... In the early days of affiliate networks, the view through was a compromised metric. It stood alone, and it took credit for everyone's sale. In the days of attribution, the view through of a banner ad (which - you should be buying based on 100% viewability, so below the fold arguments should hold no water) should be used as a pathway to a sale. Just like saying watching a TV spot or seeing a billboard along the highway should not be said they aren't effective, nor should they get full credit in your funnel. It's about building the attribution model and understanding your campaigns data well enough to be able to pull the right levers in your optimization pattern.
(part 1)
Dan, Michael, Ed, this is a good discussion, and a lot of the questions you're bringing up are frankly things I could not address in my 750 word limit for this article :)
Additionally, these are complex subjects, and ones I really like talking about and picking apart. If anyone here is interested in a discussion outside of an internet comment section, I can be reached at pbrown@bamideas.com and would be happy to meet for a call.
So to address a few points, when I said "I have been told...", I was referring to a conversation with a publisher rep during a campaign I was managing. After our campaign had been running on their site for a few weeks, they reported "the click-through rate is .05%, which tells us that people are really responding positively to this campaign!" I can certainly understand how my wording may have sounded like I was just ranting about hearsay, but I intended for that to be a recount of one particular experience I've had in buying media. So this more "I have been told" in the sense that "somebody was trying to convince me that" rather than "my friend's cousin's father's best friend from college's chiropractor said..."
Now that particular experience echoes several I've had with sellers of display and banner ads, be it networks or individual publishers. Click-through rates are an important first-step metric, but my attribution models still focus on conversions. When discussing view-through conversions, the data is way too noisy to be useful without marginality. Let's take the basic definition of a view-through, where somebody who has seen an ad (but not clicked) converts within the allotted window of time. That metric immediately needs to be discounted since the targeting itself could bias towards people who would have purchased anyways. What's often not clear is the marginal benefit of a campaign: how many ADDITIONAL buyers did it generate against the existing trajectory? There are platforms that help dissect this, but without this dissection I really do see view-throughs as a worthless metric.
(part 2)
I do tend to focus on the click-through rate because it's such a direct path to sale (though I understand that method has some weaknesses). When somebody searches for a product on Google and peruses the results for the one they want, they click on what they want. This is simple and direct and a strong measure of intent -- people compare creative (SEO metadata) and make a decision for the one that resonates most closely with their query. In the display network world, we're talking about a completely different funnel step. Somebody is not reading an article with intent to shop (most likely). There is a spectrum of reasonable expectation here where on one end we have the possibility that we can convince someone to remember a brand message delivered in a small, unstandardized, and potentially annoying format while they're browsing for something unrelated...and the other side where these aspects of the medium are completely working against even the best creative work.
If you use raw conversions as the dependent variable in that problem, the answer lies somewhere in the middle of that spectrum: yes display ads can drive some conversions. However -- and this is getting out of the current realm of responsibilites for media buyers and agencies, but it shouldn't be -- profitability matters. The agency I work for has incredible creatives with amazing ideas and tremendous executions, but all evidence points to the medium of a banner ad working against them. In other words, it's a borderline impossible task to raise the creative potential of a 300x250 so high that it can make up for .05% click-throughs, over-inflated view-throughs, and unbalanced cost/benefits.
I hope this discussion continues outside of the realm of internet comments -- seriously, please feel free to contact me!
I'll end this with a list of research, articles, and other resources that I think are helpful in understanding my thought processes:
*Book - Subprime Attention Crisis (Tim Hwang)
-Note the research by Nico Neumann referenced here
*Podcast - Freakonomics explaining and citing several studies about advertising effectiveness
-Pt 1: https://freakonomics.com/podcast/advertising-part-1/
-Pt 2: https://freakonomics.com/podcast/advertising-part-2/
*Anything by Zeynep Tufecki about the nature of attention
*The Law of Shitty Clickthroughs - https://andrewchen.com/the-law-of-shitty-clickthroughs/
Patrick, good points. However the real gorilla in the room is the distribution of online advertising. There is one party that the vast majority of advertisers would say. The same party that I am the lead plaintiff against in a class action lawsuit. Why? P&G, Unilever and others advertisers are asking how much of our advertising dollars are being spent on distribution? The answer is a ton. Where else can you can send your ads too for distribution? I advertise the advertising through sweepstakes. Why does the gorilla dislike me? I am competition. A sweepstakes entry will collect maybe 5 times more information per individual. Yet, the gorilla has tried to tear me off the net with horrible actions including call my company a horrible name in writting and lately added this: “quality guidelines with little or no added value” in referrance to my contents and advertisers. The gorilla clearly has been fearful of competition and acted to hurt or eliminate any threat to them. To solve nearly all of the ad problems, there needs better rules and improvements in the ad distribution channel.