Facebook ad frequency

Ask what a good Facebook ad frequency is and you get a number. The number is hard to use, because of what the metric is: a single division whose result depends on the dates you pick and on who is counted.

Frequency is an average, so it can read a comfortable 2.6 while a fifth of your audience has seen the ad seven times. On most campaigns optimized for sales or leads you also cannot cap it, so the number informs you but controls nothing. A cutoff borrowed from someone else's account says little about yours.

This page defines frequency, shows how an average hides heavy viewers, explains where Meta lets you cap it and where it does not, and gives a four-step read-out you can run on your own ad account. Each step has a pass and a fail, so you finish knowing whether to leave an ad alone or replace it.

Frequency is impressions divided by reach for the date range you select, and Meta labels it an estimate. Impressions count every time the ad was shown. Reach counts each account once. Meta's own example is 30,000 impressions across 10,000 accounts, a frequency of 3 (Meta Business Help Center).

The window changes the answer. In one worked example, an ad set had 119,045 impressions across 76,551 accounts, a frequency of 1.56 for the period, while a single day in the same report read 1.25 (Jon Loomer). The ad set is the same and only the window differs.

Frequency is a mean, so it says nothing about how exposure is spread across people. Meta's own example is a 14-day reservation campaign where 44% of people saw the ad four times, with smaller shares seeing it once, twice, three, five or six times (Meta Business Help Center).

One average, three very different viewers

Reach counts each person once per window. A 30-day frequency is therefore not the sum or average of four weekly ones. Pull it at the window you care about.

No single figure holds. Meta's 2016 effective-frequency study covered 11 brand campaigns from January 2015 to February 2016, each running 3 to 4 weeks with reach of 4.5 million or more. Ad recall lift slowed after a cap of 1 per week and purchase intent lift slowed near 1.5 per week. A cap of 2 per week captured 95% of potential purchase intent lift. The study concluded there is "no 'magic frequency level'" (Meta).

That study covered reach-and-frequency brand buys from 2015 to 2016, and it measured brand lift, not sales or cost per result. It does not set a threshold for conversion campaigns. Cutoffs published elsewhere are practitioner rules of thumb, not Meta guidance, and they differ by audience and by window. Treat any cutoff as a rule of thumb.

A workable baseline is the ad set's own last four weekly frequencies. Look closer when this week sits above all four and results move the wrong way.

Only where Meta provides the control. As of October 2026, frequency controls exist on reservation campaigns (Meta's pre-booked reach-and-frequency buying) and on auction ad sets (bought in Meta's standard bidding auction) with Reach (showing the ad to as many different people as possible) or ThruPlay (a video view of 15 seconds or more, or to completion if shorter) performance goals. Ad sets optimized for clicks, leads or purchases have no such controls (reservation controls, auction controls).

On an ad set optimized for clicks, leads or purchases you lower frequency indirectly. Widen the audience relative to budget, exclude people who already bought, or add new creative. Once results have turned, new creative is the lever. Adding a new creative counts as a significant edit that can send the ad set back into the learning phase (Meta ads learning phase).

Run these four checks in order. Each has a pass and a fail.

1. Pull three windows. Take the same ad set at the last 7 days, the last 14 days and the full flight. Record impressions, reach and frequency for each. Pass: every frequency you quote comes with its window. Fail: you hold one number and cannot say which dates it covers.
2. Compare with its own history. Pull the last four weekly frequencies. Pass: this week sits inside that range. Fail: this week is above the highest of the four, so continue.
3. Break it out by ad. Pass: no single ad has more than half the impressions at a frequency 1.5 times the ad set average or more. Fail: one ad does (a rule of thumb), and that ad is the one to check first.
4. Set it beside results. Put frequency next to click-through rate (see what a good CTR looks like) and cost per result for the same weeks. Pass: frequency is up but both are flat, so leave the ad alone. Fail: frequency is up, click-through rate is down and cost per result is up in the same weeks. To check the creative itself, use a Facebook ads creative audit scorecard.

Frequency is confused with reach, impressions and the frequency cap. A frequency cap is a limit you set on delivery, while frequency is the number Ads Manager reports back. "Facebook post frequency" means how often a page publishes and is unrelated.

The related terms that matter next are ad fatigue, which is a drop in results from repeated exposure, and creative iteration, which is how a team decides what replaces a tired ad.

Change the idea, not only the file. When check 4 fails, people have seen the angle enough, and a recut of the same angle often repeats the problem. Detection and recovery are covered under spotting ad fatigue before cost per result rises.

Deepsolv's creative intelligence platform combines ad-performance data, competitor activity, customer signals and past ad learnings into a ranked weekly creative test plan and execution-ready creative briefs. The replacement ad tests an angle you have a reason to believe in.