Don’t Believe The Hype - Tight Site Lists Are Not A Scalable Solution
By Rodger Wells
There was an article in The New York Times published last Thursday designed to send fear through small and mid-sized publishers and advertisers. If you are someone who believes what you read in the paper, this is the end of programmatic.
In the article, JPMorgan Chase claimed they cut their digital approved site list by almost 99% from 400,000 to 5,000 sites, and yielded the same results. Why do we need programmatic advertising, networks, or exchanges if 5,000 sites can replace the work of 400,000 sites? Why do we need AI and machine learning if all we need is an intern to manually check sites?
The scenario that JPMorgan Chase presented only works when you are the only company moving to a restricted site list. JPMorgan Chase may indeed have seen success on these 5,000 sites. However, what happens when other brands follow suit and put their ad dollars towards the same 5,000 sites? The result is driving CPMs up, as the number of available impressions is limited. Ultimately the price of inventory gets so high that the ROAS is near zero.
The logic behind the story itself is nonsense and an attempt to scare advertisers into thinking that the only path for digital success is to advertise with tier-one blue-chip publishers, such as the publication running the story - The New York Times. Procter & Gamble’s Marc Pritchard has already threatened to pull money from publishers that don’t meet their standards. AT&T and Verizon both said they are pulling advertising from YouTube and Google due to brand safety concerns. If you are a marketer at Procter & Gamble, GEICO, Citi, Verizon, AT&T, Ford Motor Company, General Motors or any other major brand and you saw this article as a path to success, then you just got sold!
We aren't denying the fact that there are major problems in the ad tech industry involving quality and brand safety. Going backward is not the answer.
Here’s a simplified example of the faulty logic behind this article. Today there are billions, literally billions of websites. Advertisers appear on many of them. In light of this information, the 400,000 URL site list JPMorgan Chase used doesn’t seem so big any more. Now, imagine that someone decided the only site that gets real return for them is www.widget.com. All those ad dollars chase www.widget.com and its limited amount of traffic and avails. What happens to the price of inventory on that site? It skyrockets. What happens to returns for advertisers - they go way down while risk goes way up. This will be the same result if you hand control to a select number of premium publishers.
The problem in the industry right now seems to be a debate between benefits of a human vs. machine solution. We need both. Humans are more prone to error than machines. We need a solution that uses humans to compliment machine learning or small publishers and advertisers alike are doomed.
The answer to scalable efficiency and satisfactory ROAS lies in delivering a simple promise. It's the entire baseline concept behind advertising. The goal of advertising is to deliver an ad to a person and make them do something – whether that is to visit a website, go to a store, or buy a product. In the article, JPMorgan Chase noted the company saw little change in the cost of impressions. However, looking at impressions is only 2/3 of the promise of advertising, all you have done is deliver an ad to a person. Making them do something is engagement and engagement is the solution. It is not limited by a tight set of domains that arbitrarily create a perceived efficacy.
Providing a brand safe, human, targeted and engaged audience regardless of the site is the way that brands will succeed. Following news stories that are designed to scare people into erroneous actions only results in paying too much for too little.
In the words of Public Enemy - "Don't believe the hype..."