Frequency capping, also known as impression or ad capping, is an ad targeting option that limits how many times a user sees a particular ad or advertiser in a given time frame.
As you build your own ad platform, undoubtedly this feature will come up - either internally or as a request from an advertiser. The question is - should you spend months researching and building this functionality?
Frequency capping is an ad targeting option that limits how many times a user sees a particular ad or advertiser in a given time frame.
Generally ad capping involves three components:
For instance:
There are three main benefits of implementing ad capping into your product:
Your ad server will likely be competing with Google, Facebook, Amazon, etc for budget. It’s important, therefore, to have as much feature parity with them as you can.
Is frequency capping one of those must-have features?
The answer isn’t a clear yes. Below looks at the major ad buying platforms and whether they allow advertisers to manually manipulate impression capping rules.
Platform | Do they offer? |
---|---|
Google Ads | Yes |
Yes | |
Snapchat | Yes |
Yes | |
DSPs/Exchanges | Most |
No | |
No | |
Quora | No |
Microsoft Ads | No |
No | |
Amazon Sponsored Products | No |
eBay Promoted Listings | No |
Etsy Promoted Listings | No |
As you can see, it’s a mixed bag, with the two largest players offering it (Facebook/Google), but the others divided. From this standpoint, we would not consider it a tablestakes feature. That said, do note that brand-conscious advertisers with larger budgets may see this as a non-starter.
Maybe more important than feature parity is the question of whether your decision engine should employ ad capping by default. This would involve caps built into the logic, with or without an ability for the advertiser to adjust them.
You wouldn’t be the first engineering team to do this. LinkedIn Ads have fixed frequency caps, with their docs saying, “Because LinkedIn is very focused on member experience, frequency caps are strictly enforced. This means that if an ad was already served to a particular member he or she will not see it again for 48 hours.”
Twitter too incorporates impression capping by default into their logic, but they only state they do it, not the exact rules they have in place.
Yup! For this article we ran a month-long campaign on Google Ads. We limited targeting to one country, one site, one ad unit, and an intent audience around ‘ad serving’. We then built four campaigns with the same Kevel ad, the only difference being what frequency caps they had.
We made sure to give the stricter-cap campaigns higher CPM bids - ensuring, for example, the campaign that was limited to one ad per week always won when eligible.
The results are not too surprising, with performance (as measured by click-through-rate and cost-per-click) degrading with looser frequency caps.
Campaign | Caps | Impressions | Clicks | CTR |
---|---|---|---|---|
#1 | 1 per week per user | 339,400 | 560 | 0.16% |
#2 | 3/week (4 total) | 213,900 | 280 | 0.13% |
#3 | 6/week (10 total) | 122,200 | 65 | 0.05% |
#4 | Unlimited | 215,400 | 50 | 0.02% |
Total | 890,800 | 955 | 0.11% |
Users who saw the ad only once per week clicked on it 0.16% of the time - but the cohort that saw it more than 10 times clicked on it just 0.02% of the time (an 87% drop). Since I was buying on CPM, my cost-per-click was about 6x higher for the fourth campaign versus the first!
I’m now going to use these figures to extrapolate how frequency caps could drive more revenue for you, depending on how you charge (using the two most common pricing methods).
Let’s say you sell ads based on CPM, where advertisers pay a set amount for every 1,000 impressions. I come to you looking to run a month-long Kevel test, and we negotiate a $5 CPM. At the end of the month, I see these results (the below imps, clicks, and click-through-rate come from the actual Google campaigns):
Impressions | Clicks | Fixed CPM | My Cost (Your Revenue) | Effective CPC |
---|---|---|---|---|
890,800 | 955 | $5.00 | $4,454 | $4.66 |
Perhaps I’m fine with $4.66 CPCs - but potentially my boss has demanded we only work with ad sources that drive $4.00 CPCs or lower. If that’s the case, I would not re-up spend with you; if this were a pervasive problem across advertisers, you may have trouble scaling your ad platform.
But what if your ad logic by default incorporated a cap of four ads per week per user? How would that have impacted performance? Here would have been the results for just the first two campaigns:
Impressions | Clicks | Fixed CPM | My Cost (Your Revenue) | Effective CPC |
---|---|---|---|---|
553,300 | 840 | $5.00 | $2,766 | $3.29 |
You would in this scenario make less from me in the short-term, but presumably you’d have additional advertisers to fill those impressions.
To be clear, this logic would be built into the platform by default, not a feature you would let advertisers toggle (such as what LinkedIn and Twitter do).
Here, you are charging advertisers by a set CPC. In this case, you don’t have to worry about the advertiser seeing variable CPCs and becoming unhappy (they will bid the price they want). Instead, you have to worry that you are leaving money on the table by showing ads to users who don’t click (aka, if you charge by CPC, any ad that doesn’t get clicked on is lost revenue).
Like above, let’s say I run a test with you at $4.00 CPCs. With no frequency caps, I would have seen:
Impressions | Clicks | Fixed CPC | My Cost (Your Revenue) | Effective CPM |
---|---|---|---|---|
890,800 | 955 | $4.00 | $3,820 | $4.28 |
Let’s look at what would have happened if you frequency capped at four ads per week:
Impressions | Clicks | Fixed CPC | My Cost (Your Revenue) | Effective CPM |
---|---|---|---|---|
553,300 | 840 | $4.00 | $3,360 | $6.07 |
Here, you would have made about 90% the revenue on just 60% of the impressions. And that remaining 337K impressions could have been allocated to someone with a higher CTR, such as:
Advertiser | Impressions | Clicks | CPC | Your Revenue |
---|---|---|---|---|
Kevel | 553,300 | 840 (0.15% CTR) | $4.00 | $3,3,60 |
Advertiser #2 | 337,500 | 512 (0.15%) | $4.00 | $2,049 |
Total | 890,800 | 1,352 | $5,409 |
Of course, I’m using hypotheticals and data from one test, which makes it easy to reach the conclusion I want. This article isn’t meant to say implementing frequency capping means you’ll drive 40% more revenue; it’s more to plant a seed that this feature is one worth looking into.
There is no magic answer here. In our Google Ads test, I arbitrarily chose the four breakdowns. Maybe the results would have been starker if I had capped the second campaign at two impressions per user instead of three.
Moreover, no ad capping analysis will hold true across the board, even on the same site or app. For instance, a full-page mobile takeover may see a huge CTR drop-off after just one impression, but a small native ad may take five impressions to reach that threshold.
Chris has worked in ad tech for over twelve years in a variety of roles - giving him customer support, PM, and marketing perspectives from both the advertiser and publisher sides. He's the VP of Marketing at Kevel.