Less is more: why you need to change your approach to programmatic ad fill

Programmatic advertising is all about the data. Publishers are focused on driving up the numbers – CTR, CPM, revenue, fill rate. Higher numbers are better. Higher numbers equate to greater profitability. Or so the traditional way of thinking goes.

We regularly encounter publishers who want to know how high a fill rate we can offer. They are asking the wrong question. Quite frankly, it misses the point. If your goal is to drive ad revenue, then this should be the north star of any campaign.

Ad fill is not the goal of programmatic advertising

Say that you have a choice between earning a $5 CPM with 50% ad fill or a $2 CPM with 100% ad fill. What do you go for?

Your answer should be the first option. A $5 CPM with 50% ad fill equates to an effective $2.50 eCPM. That’s more than the $2. Not only will you earn higher total revenue, but you’ll do so from half the number of impressions. This is better for the user experience.

Yes, serving fewer ads is a good thing. Showing ads is a means to achieving the goal, but it’s not the goal itself. This is an important distinction to make.

Think about it this way. The real goal of publisher monetization teams should be to generate revenue. The reason publishers often aim for 100% fill is because they believe that higher fill automatically translates to higher revenue.

Find a solution to blank spaces

Another argument in favor of 100% ad fill is that you don’t want blank spaces showing up when an ad placement doesn’t have fill. Although this is a legitimate concern, it shouldn’t be a limiting factor.

One option is to stop hard coding your ad placements. At FirstImpression.io we create ads dynamically to avoid the issue of blank spaces. If there’s no fill for one of our ad placements, it simply won’t show.

Another option is to backfill the ad placements or show your own branding in the space. Don’t insist on 100% ad fill from the get go. Make sure that your team is free to make decisions based on revenue. Backfill can always come after, although from our experience, it often isn’t worth it.

Ad fill does not always correlate with revenue

Let’s go back to the main point we hear when publishers ask for a high ad fill: that it indicates revenue. Mathematically speaking:

Total revenue = ad requests x ad fill rate x CPM / 1000

The logic goes that if you have a higher ad fill on the right of our equation, the total revenue will also be higher. So the two metrics correlate.

Not exactly. If the CPMs were consistent, that would be true. However, CPMs and ad fill rate are both affected by the floor prices that publishers set and often move in opposite directions. You can see this from the graph below. The purple line here indicates the fill and the green line shows the CPM for one ad product on Google AdX.

Raising the floor price may generate a higher CPM because you are now attracting higher quality campaigns from advertisers. However, this can result in a lower fill rate. And vice versa. Lowering the floor price can attract advertisers looking to pay as little as possible for your high quality traffic. This may bring you a higher fill rate but a lower CPM.

In other words, the promise of 100% ad fill does not automatically mean that your programmatic advertising revenue will be higher than if you had 80% ad fill. The most important metric to track and optimize is revenue.

Adjust your floors to optimize revenue

When you’re getting started, set a lower floor price that will give you higher ad fill for your ad placements. Then raise it a little and see what happens. You’re looking for revenue that is the same or higher with a lower fill rate. Adjust the floors gradually until you find the optimal point.

However, bear in mind that changing the floor can cause demand partners to allocate campaigns differently to your inventory. Sometimes a slight increase of the floors will result in a lower revenue while doubling the floor price may generate a new set of campaigns with greater total revenue.

The trick is to constantly test and optimize. Experiment with different floors and track the effect of this on your various metrics to find the optimal point. Demand fluctuates so you’ll need to repeat this on a regular basis, ideally daily.

If you have any questions for our revenue optimization team about the above post, or if you want to learn more about how we monetize ad products here at FirstImpression.io, please feel free to contact us! We’d be happy to help.