Publishers get more; Advertisers pay less - gut check your damn CPMs

Publishers get more; Advertisers pay less - gut check your damn CPMs

Time for the latest quarterly CPM benchmarks. The data table below shows CPMs for display ads in Latin America and U.S. The top 3 are from a year ago (Q4, 2023) and the bottom 3 are from this week (Q4, 2024). These are the CPMs observed in the DSP. This is the price paid at the DSP level or the eCPM ("effective CPM") paid for these types of ads. Note there's not much difference between 2023 and 2024. The green gradient shows the prevalence and the yellow highlight shows the most typical CPMs, in US Dollars.


Advertiser can pay less

Why did I publish the above benchmarks? It's so that you would take a closer look at the CPMs you are paying for display, video, and CTV ads. Let me use some round numbers to illustrate a problem hiding in plain sight because advertisers don't know what CPMs are actually being paid for the ads they buy via media agencies.

Some advertisers are paying $15 CPMs to their media agencies to buy display ads for them. The media agencies are turning around and bidding $1.50 CPMs on their behalf. The U.S. CPMs in the table show $0.80 - $2.65 CPMs paid for display ads at the DSP level.

"Advertisers should ask for placement reports with eCPM per domain"

Advertisers should ask their media agencies to pull placement reports, detailing the list of domains where their ads ran and the eCPM ("effective CPM") paid per domain. The table below is from MY OWN campaign, run in DV360, display ads only, in the U.S. If you are paying the agency 10X higher CPMs than what is reported in the DSP placement report, there's a problem. Advertisers can pay significantly less than they are paying now and save a lot of money on programmatic ads.


Publishers can get more

On the flip side, the publishers/sellers are getting pennies on the dollar, compared to what advertisers paid. I have run the following experiment with 2 dozen publisher groups now. And some publishers are leaving 90% of the CPM on the table, because they don't know what CPM the advertiser paid for the ads, like the display ads mentioned above.

The problem comes from the unknown number of middlemen in between the seller of the ads (publishers) and the buyers of the ads (advertisers). And further, the unknown margins and fees that are hidden in the media CPM. If advertisers are paying $15 CPMs for display ads and the publishers are getting $0.15 CPMs, someone or "someones" took 99%.


So what?

Time to gut check your CPMs, whether you're a publisher/seller or advertiser/buyer. Advertisers, ask your media agencies to pull a placement report from the DSP that details the domain and the eCPM paid per domain, like the report above from my campaign in DV360. If the eCPM paid is far less than what you are paying to the agency, it's time to have a conversation with them.


How to simulate a direct-buy in programmatic

In any case, advertisers can pay less and publishers can get more, by doing what I call a "simulated direct buy" using programmatic channels. This way you don't have to bring buying in-house just yet and you can essentially do what you are doing now, but save money and get more at the same time. Advertisers should ask a seller/publisher what is the 1 exchange they prefer to sell through, and what is the 1 sellerID or dealID they should target in that exchange. By doing this, the advertiser is simulating a direct buy, and eliminating all the other supply paths that led to "leakage" (ads not going where they should be) and fraud (fake sites pretending to be legit publishers' sites). Doing this also dramatically increases the accuracy of your buy, even if you are on inclusion lists already. In the slide below, only 73% of my ads went to the 1 domain in my inclusion list, when "all exchanges" were left on. The accuracy increased to 97% when I unchecked all of the other supply paths and left only 1. That's a 33% increase in accuracy, when using a 1 domain inclusion list. You can imagine the gains in accuracy when you have far larger inclusion lists.

In summary, advertisers can pay less by bidding $8 CPMs in a "simulated direct buy" instead of paying $15 CPMs. At the same time, publishers can get more -- $8 CPMs instead of $0.15 -- by selling through a "simulated direct buy." I'd call that a win-win, wouldn't you?

Anyway, I hope this was helpful for advertisers to save more money and publishers to make more money -- and thus help the entire digital ad ecosystem to be more efficient and effective.


Subscribe for more FouAnalytics case examples, screenshots, etc. - https://www.dhirubhai.net/in/augustinefou/recent-activity/newsletter/








Just like it should be. Your insights are needed for change Dr. Augustine Fou

Bruce Clark

Associate Professor of Marketing at D'Amore-McKim School of Business at Northeastern University

1 个月

Love when you share stuff like this.

要查看或添加评论,请登录