Est. reading time: 6 minutes

Retargeting is usually the most efficient line item in the account right up until it becomes the most resented. The complaint that eventually surfaces, the customer asking why your brand follows them across every app, is just the audible end of a decay that started weeks earlier in the metrics, where frequency climbed, clicks thinned, and cost per acquisition drifted up while everyone watched the ROAS column instead.

Most brands never install frequency governance because retargeting’s reported numbers flatter it. Warm audiences convert at rates that make the campaign look untouchable, including plenty of people who would have bought anyway. That flattery is exactly what lets frequency run past the point of usefulness unnoticed. The fix is a cadence system with tiers, caps, cool-off rules, and budget math, and none of it is complicated once the decay mechanism is clear.

How retargeting frequency goes wrong

Intent decays with time. Someone who abandoned a cart yesterday is a fundamentally different prospect than someone who skimmed a collection page three weeks ago, and serving both the same impression load mismatches pressure to intent. The recent abandoner can absorb a few reminders. The stale browser gets the same treatment and experiences it as pursuit.

The decay shows up as a specific signature. Average frequency rises, CTR falls, CPA climbs. Our working trigger is a CTR drop of roughly 30 percent or more once average frequency pushes past the 5 to 7 range without a compensating lift in conversion rate. Past that point, incremental impressions are buying irritation, not consideration.

Two quieter signals deserve the same attention. If branded search volume holds flat while retargeting impressions balloon, the campaign is claiming conversions rather than creating them. And if view-through conversions climb while on-site engagement from the campaign deteriorates (shorter sessions, higher bounce), the reporting is overcounting credit for impressions that did nothing. Frequency distribution reports close the loop. Large cohorts sitting at 10 or more impressions with no lift are budget that should move to fresher tiers, or to fixing the deeper problem of retargeting the wrong audience in the first place.

Build the cadence on recency tiers

Segment retargeting pools by time since last visit (0 to 3 days, 4 to 7, 8 to 14, 15 to 30) and step the impression pressure down as the tiers age. Behavior cuts across the same grid. A cart abandoner in the 0 to 3 day tier justifies daily contact. A no-add-to-cart browser in the 15 to 30 day tier justifies a weekly touch at most.

Message should move with the tier, because frequency tolerance is partly a function of whether the audience is seeing anything new. Early touches handle objections (reviews, shipping, returns, a quick demonstration). Middle touches deepen the case (bundles, use cases, comparisons). Late touches make a final offer or exit gracefully. Rotate at least three to five creatives within each tier so accumulated frequency reads as a conversation progressing instead of one message repeating, a structure that maps cleanly onto audience segments in Google Ads as well as Meta.

Then wire in the rules that most accounts skip. Exclude recent purchasers, active trial users, and anyone currently in a support queue, since pitching an upsell to someone waiting on a service ticket is how retargeting earns its reputation. Add cool-off logic on top. Our working rules are to ease back after 3 to 5 exposures with no click in a week, and to pause anyone with no site return in 2 to 3 weeks for a 7 to 14 day rest before re-qualifying them into a lighter tier.

Retargeting frequency caps: starting ranges and where they actually live

Ranges beat absolutes because sales cycles differ, but every account needs a starting position. Ours: 2 to 3 impressions per person per day on paid social, 1 to 2 per day on display, and weekly totals of roughly 7 to 10 impressions for high-intent tiers, 3 to 6 for mid-intent, and 1 to 3 for light-intent, stretched longer for considered purchases. Email runs on its own clock at about one remarketing touch per week, with an unsubscribe rate above roughly 0.2 to 0.4 percent treated as a fatigue alarm, and with the list cut by behavior using the segmentation tactics most brands skip so the cadence lands on people it’s still relevant to.

The catch is that platforms don’t hand you these caps directly. On Meta, hard frequency caps exist only in Awareness campaigns using the Reach performance goal. Conversion campaigns have no cap field at all, so frequency is controlled indirectly, through audience size relative to budget, tier exclusions, creative rotation, and automated rules that flag runaway frequency, all of which we covered in our breakdown of Facebook ad frequency caps. Google Display campaigns do offer manual caps per day, week, or month through the Frequency management setting, with only viewable impressions counting toward the limit, though the setting only appears after the campaign is created and defaults to letting Google optimize.

Cross-channel is the gap nobody’s tooling closes for you. Meta, Google, and TikTok each cap only their own impressions, so a person inside three retargeting pools experiences the sum. If you can’t deduplicate across platforms, set each channel’s cap conservatively and cut spend where audience overlap is heaviest, which is also the discipline behind building TikTok retargeting audiences without wasting impressions.

The budget math that keeps frequency honest

Runaway frequency is usually a budgeting error wearing a creative problem’s clothes. A retargeting audience is a fixed pool, so any budget beyond what the pool can absorb at your target cadence has nowhere to go except repetition. The math takes two lines. Impressions needed equals audience size times desired weekly frequency. Weekly budget equals those impressions divided by 1,000, times your blended CPM.

Worked through, a 50,000-person pool at a target frequency of 5 per week needs 250,000 weekly impressions, which at a $6 blended CPM is about $1,500 per week. A brand spending $4,000 per week against that same pool hasn’t bought more conversions. It has bought a frequency north of 13 and the fatigue that comes with it. Run the math for each tier, start 20 to 30 percent below the output, and only scale where results hold as frequency rises.

None of this caps upside. It caps waste. Intent decays on a schedule, and a retargeting program built around that schedule, with tiers that lighten over time, messages that progress, exclusions that respect context, and budgets sized to the pool, converts the people who were persuadable and leaves the rest alone before they start leaving comments about it.