One of the lesser known concepts in online advertising is yield optimization, which is a way of managing ad inventory to produce the highest average RPM or revenue per thousand impressions.
It starts with the basic idea that a site publisher or sales team may not be able to sell out 100 percent of all ad impressions on a site.
The inventory that is left is an opportunity to make more money.
But the strategy is not simply about filling that leftover inventory with house ads, extra unpaid impressions for a client or even with a single remnant advertiser.
It is about signing up multiple remnant advertisers who all compete for that inventory. The remnant with the highest RPM is the winner.
That sounds easy enough in concept, but it takes some consistent effort to get there.
How to Optimize Yield
Optimizing yield requires an ad serving platform that can automatically adjust remnant campaigns to deliver the ones with the highest RPM.
Google DFP will serve as a model in this scenario because it has a feature called Price Priority and because it is widely used in the online industry.
When setting up a remnant campaign, create a line item and scroll down to Type. Change the type to Price Priority. Scroll down just a tad further and change Rate to the RPM at which that advertiser will start delivering impressions.
Do the same for at least two or three more advertisers, all with the same rate.
One week later, run a report for each advertiser showing the actual RPMs they are delivering.
Go back into the line items and change the Rate again to those RPMs that showed up in the remnant reports.
Let’s say that remnant networks named A, B and C all have an initial Price Priority rate of $1. After running the reports, they reveal that A is delivering a $1.50 RPM, B is delivering $1.00 and C is delivering $0.50.
It’s clear that A is the most valuable of the three providers and should receive the most impressions.
With the new rates in place, A will start delivering impressions at a $1.50 RPM and keep delivering them above a $1 RPM until it runs out of them.
Once that happens and the average rate reaches $1, impressions from advertiser B will start to appear.
B will keep delivering until IT starts to run out of impressions and the RPM drops to 50 cents. At that point, C starts to deliver.
One week later, run another report. Keep running reports once a week and adjust accordingly because no ad network will provide the same RPMs and the same number of impressions every week, although patterns do develop over time.
If advertiser C continues to provide low RPMs or any of them don’t provide enough impressions, it’s time to go shopping for more networks to add to the list. Drop the poor performers to keep the list of providers manageable.
Using Yield Optimization with Direct Sales
Direct sales of online ad inventory often means landing accounts that pay a fixed CPM (cost per thousand) for a fixed number of impressions.
Certain advertisers could end up buying remaining impressions using this concept.
If remnant networks are producing an average RPM of $1.00, maybe a direct sales advertiser would be interested in paying $1.50 instead for a variable number of non-guaranteed impressions.
Besides, the remnant networks usually take 30 to 50 percent of the revenue from the originating clients. Why not get that money instead?
One risk with that approach is the possibility that a client could end up devaluing the ads and will be unwilling to pay more for a fixed number of guaranteed impressions in the future.
Either way, a small amount of ongoing effort will produce better yields and more revenue.
Understanding the Results
A 100 percent improvement in RPM sounds impressive, but it doesn’t have a major impact on total revenue for a small site with 10,000 extra impressions a month.
The reality is that many remnant networks might provide anywhere from a $1 to $3 RPM depending on the site, current economic conditions, time of year, competition and other factors.
Moving the needle for the site above with 10,000 extra impressions from a $1 RPM to $2 means only $10 a month more. If nothing else, it will cover the cost of a small hosting account.
Moving it on a site or network with 10 million extra impressions will produce another $10,000 a month. That’s the equivalent of two to three full-time employees — or a big improvement in profit.
So the volume of impressions will matter as much as how the percentage increase in RPM. If adjusting the campaigns take only five minutes a week, the effort will be worth it for many sites.