TUESDAY JANUARY 17th 2023

Last week, I had a conversation with Robert Regular—a friend and colleague of mine who I have known for over 25 years. We talked about many things throughout our discussion, but one topic kept coming up: does Programmatic advertising even work?

The media loves to seize on new technologies and turn them into must-haves before people have even learned how to use them. And as artificial intelligence improves, you can be sure that the hype cycle around it will only get worse—especially since we're in a post-Turing world where anything seems possible!

Because of this, you see large companies and small businesses alike investing nearly all their advertising money in programmatic ads.

Just as the Wild West was filled with settlers looking for a better future, so too is your industry today—with hordes of brands like yours vying for success.

Both Bob and myself have heard from many different businesses of all sizes, who have tried programmatic advertising over and over again—sometimes with no success at all.

At events we've both experience the same, where people have often come up and asked us if programmatic advertising "even works?" Story after story we've been told from random strangers to potential clients-- that they have been extremely burned by programmatic advertising with little or no proven results given for often budgets wasted in days.

How can this be? We've been told for almost a decade that programmatic is the key to online advertising, that it doesn't just make things easier, but it is highly effective.

With programmatic, advertisers no longer have to do deals with publishers directly and speak to those audiences—the audiences themselves become decoupled from context.

The assumption is that the real audience you get through programmatic isn't just vague like customer profiles were in the past; it's a true representation of who people are as consumers today. That through "data" programmatic targets people based on real interests, and "science" proves it.

However, despite all the claims of immensely great targeting, the main reason large budgets are drawn to programmatic is because of the promise of scale. That using these complex algorythms, they promise to scale your campaigns almost infinitely, finding large audiences and inventory that was never known before.

Agencies especially have gobbled this idea down, especially when it decreases CPMs and they can show their clients they reached "millions more people."

Some adtech vendors and agencies have convinced advertisers that the lower price of programmatic advertising is evidence of greater cost efficiency, when in fact it simply represents inferior targeting capabilities.

"Advertisers have been addicted to programmatic media for the last ten years, because of the enormous scale, cost efficiency, and high performance. The large quantities of ads comes from bot activity, fake sites, and fake mobile apps. The low CPMs are due to fake sites selling ads at low prices because they have no costs for making content. The high clicks are from bots, programmed to click on the ads to create the appearance of performance." - Augustine Fou

While adtech companies have made a killing off programmatic models, networks and "magic," it's the brands—not them—who suffer most. This isn't a realistic business model, relying on advertisers to ignore that stuff just doesn't work.

So what is the real issue of programmatic besides just obscene amounts of fraud?

1) These are ads no one wants to see. For example, a consumer is surfing the web looking for recipes for gingerbread cookies to make with her daughter. She visits a recipe website but finds that many of its pages are covered in pop-ups and interstitial ads. The assumption is that "based on data" she will be interested in all the ads popping up about "Spanx" underwear. Why? Because someone claims that since she visited Oprah.com and then her phone was used to buy a calendar, she is "more likely" to want to buy Spanx products.

She may not even be able to find the recipes she needs, as food websites have deteriorated particularly badly. Did she want to see any of that sponsored content? Of course not, she is there to find a recipe and at the time would be most interested in products related to cooking and those things that interest her right there. She is prime for buying cooking products, maybe a new pan, but instead of targeted site-specific ads, she's being shown "programmatic" targeted ads that are irrelevant.

When you hassle people in this way, they tend not to like it. In practice, the success rate of programmatic advertising is not much better than simply placing those ads completely randomly sites because minimal attempts are made to target them based on the content of their pages—the advertiser seeks no information from people who visit. This flies in the face of everything we've been taught about how marketing and advertising works. Throwing "stuff" at the wall, and then paying attention to what sticks may not actually be a real model.

2) It's bad for Publishers, Destroying Targeted CPMS. Anyone remember site-specific buys and sponsorships? This used to be the core of most media buys, back "in the day." Where brands would buy ads on a specific site, integrate their brand into the site through skins, giveaways, content and the such. This stopped when media buyers, frankly, became lazy and overpaid. It was too hard to "scale" they said, and despite being highly effective, it doesn't peak the interest of the tech gods.

The race to the bottom in terms of pricing has meant that independent publishers have seen their ad revenue steadily decline. Indeed, it is easy to place advertisements on your site and start earning money from them—but only if you have a massive number of visitors. People's work on a website, when there were site-specific buys and brands "worked" with publishers, used to make a living. No longer.

When you’re buying directly from publishers on Hulu, Pluto and other similar sites, it’s much less risky than through third parties. “The process is cleaner and fraud is far less likely to occur since content owners are more willing sellers who don't need the additional revenue stream provided by advertisers or media agencies' trading desks,' said Dave Campanelli, Executive VP/Chief Investment Officer for Horizon Media

Researchers from the University of Minnesota, University of California, Irvine and Carnegie Mellon University found that data targeting through the ecosystem contributes only 4% more revenue for publishers than standard online advertising.

The “ad tech tax,” which refers to the portion of each dollar spent on programmatic ads that goes toward middlemen's fees rather than directly paying publishers for their content, eats up 60 cents out every dollar.

The proliferation of fake news and clickbait has helped to drive the Internet towards a state where programmatic impressions reign supreme, since that generates traffic for sites by whatever means necessary.  

The $200 billion programmatic ad market presents a more lucrative stomping ground for criminals than even international drugs rackets: There is no product to move, hardly any risk of getting caught or punishment if you do get caught—and then there's the sheer volume at which ads can be sold. Real publishers can't compete with this.

3) It Doesn't Actually Perform

For every single campaign that I’ve seen, programmatic display and video is more expensive than social, for both reach and clicks, drives less qualified traffic and almost zero viewability. Most show on the screen for a second and then disappear.

The idea that programmatic is cheaper doesn’t take into account the real costs of advertising and realistic KPIs.

Part of this is an extreme overlap of the ecosystems. Agencies will often buy from one vendor, and then buy from another, not realizing that like almost all of the display programmatic companies, they are showing on the same exact sites. So while the reach seems "better," it's just showing the same ads over and over again to the same users.

Most advertisers want to buy from as many ad systems/exchanges as possible, thinking that gives them more reach. Looking at the top 10 domains by volume for a campaign run on 4 different systems according to Dr. Fou—Trade Desk included—reveals almost no difference between any of those supply sources.

The NYTimes even noticed this issue, and wrote an article about how Chase bought ads on 400,000 sites and had the same exact results as buying on 5,000 sites.

This data suggests that the fraud rate in their first media buy was at least 99%, meaning that entire ecosystem is just plain dirty and bad.

What is the result of all this? What is going to happen?

For a while I thought that advertisers were just too stupid to realize that they are being ripped off by the plethora of questionable programmatic vendors. However, it seems that people are starting to pay attention.

53% of marketers expect that they will spend less on programmatic advertising this year than last year, largely because companies are investing more money in walled gardens.

A survey commissioned by identity platform Lotame and conducted by PureSpectrum has found that 15% of marketers expect to reduce programmatic spending more than 50%, while an additional 20% plan on cutting it 26%-50%.

Chris Kane, founder of ad-tech consultancy Jounce Media, said that by bidding for ads across the open internet—on web, mobile and CTV (television) —the risk is too high for buyers to inadvertently advertise on low-quality inventory.

In a survey conducted by Jounce, ad-tech professionals estimated that 12% of the most common inventory marketers bid on were poor investments for the typical marketer. The actual numbers are likely much higher.

In an effort to avoid ad exchanges and keep consumers loyal, many industries have developed direct-to-consumer marketing campaigns. And TikTok -- let's be honest. Programmatic is taking a back seat to that right now, and people realize that social media is killing it in comparison when it comes to actual measurable results. 

Pesach Lattin
Pesach@lattin.us

Please connect to my new linkedin profile just for this publication