Ori Mendi, Co-Founder & CEO, Kueez
Think higher CPMs always mean better yield? The numbers reveal that publishers focusing solely on high CPMs might be leaving money on the table. Why looking at averages can be misleading, and what do university admission rates have to do with all that?
Averages can be misleading. Here’s a quick example: Michael Jordan and I averaged 536 NBA games played. Sounds impressive, right? But when you look closer, you’ll see MJ graced the court 1,072 times, while I… well, didn’t even make it to warm-ups.
This is the trick of averages - they can hide the real story. In this article, I’ll uncover how the same principle applies to CPMs in programmatic media, where relying on averages can distort the true value of your ad campaigns. Let’s dive in.
Here's something that might sound illogical at first: In one hypothetical university, women might have a higher acceptance rate in every individual faculty, but could still have a lower acceptance rate overall. How can that be? It's all due to how averages work, and they can often hide the whole story.
The Average Dilemma: University Admissions
This idea of "the average hiding the truth" is just like the university admissions paradox. Imagine women have a higher acceptance rate in every faculty - 35% for engineering and 80% for humanities - while men have a lower acceptance rate of 25% for engineering and 60% for humanities. However, because more men apply to the humanities faculty, which has a higher acceptance rate overall, their total average acceptance rate ends up being higher than women’s.
This sounds surprising, right? Even though women are more likely to get accepted into every individual faculty, the overall acceptance rate for men is higher due to where they applied:
Just as the university admissions paradox, the same principle applies in digital advertising, where a lower average CPM doesn’t necessarily mean weaker results - it can still deliver higher CPMs for specific audiences and stronger overall performance. What truly matters isn't the average, but where and how often you win.
Ad-Tech’s 'Average Dilemma' Exposed
Let's say you're expecting $1.5 CPM bids for your ad impressions, but instead, your ad slot gets filled at a $1 CPM. It feels like you're losing out on value, right? Not so fast. If that $1 bid won your auction, it means that this was the highest price any advertiser was willing to pay for that impression. In other words, the CPM wasn't low at all - it was the top bid for that specific user or ad space, at that specific moment.
The concept of a "Low CPM" can be misleading. Each winning bid represents the maximum amount willing to be paid for that specific impression. This means success isn't determined by a high generic average CPM but rather by how frequently you secure valuable impressions and deliver the highest possible value for each one.
The comparison below between two Ad-Tech vendors clearly illustrates this point, showing that a lower overall CPM can actually represent a higher CPM per audience:
So what matters most when it comes to performance?
At first glance, Company A appears to perform better with a higher overall CPM of $2.50 compared to Company B's $1.83. However, a deeper analysis shows that Company B places higher bids within each audience segment, ultimately driving significantly more publisher revenue.
This highlights that, despite Company B's lower average CPM, total revenue (Driven by higher win rate) is a far more meaningful metric than simply focusing on CPM averages. The only way to win more is by offering higher bids - and this always holds true, even if the average CPM doesn’t reflect it.
What truly matters is how often you win and how much total revenue you generate. As we at Kueez observed, focusing on win rate and overall revenue will always give a clearer picture of success than chasing higher CPMs alone.
A Winning "low" CPM isn’t really low
In the digital advertising world, this is especially evident in environments where user data is limited, such as cookieless environments. CPMs in these environments are often lower due to reduced targeting data. However, platforms that excel in these lower CPM environments can unlock valuable revenue opportunities that might otherwise go untapped. A high win rate in these spaces may seem to reflect lower bids, but in reality, you’re capturing significantly more value and revenue by strategically winning and offering higher bids for these audiences.
Just as it’s surprising to think that men can have a higher overall acceptance rate despite lower acceptance rates in each individual faculty, it’s surprising to realize that a "Low" CPM isn’t really low when it wins the bid. Every winning CPM is the highest possible for that moment.