How Often Should You Refresh Meta Ad Creative in 2026?

Ask ten media buyers how often to refresh Meta ad creative and you'll get ten different answers. Every two weeks. When frequency hits 3. When CTR drops 20%. All of them are right sometimes and wrong a lot.

Refresh cadence is not a fixed number. It's a function of your spend level, your audience size, and how you read the signals your ads send. Here's how to figure out your actual number.

Why "when frequency hits X" is the wrong trigger

The frequency-based trigger made sense when Meta's delivery was more predictable and audience saturation happened in a fairly linear way. Under Andromeda's current delivery model, it's an unreliable signal for two reasons.

First, frequency is a population average, not a maximum. A reported frequency of 2.5 means the average person in your audience has seen the ad 2.5 times. The actual distribution is skewed: your high-intent audience segments are being shown the ad more often than average, because Meta delivers more to people it thinks are likely to convert. By the time your average frequency is 2.5, your most valuable buyers may be at 5-6 exposures.

Second, Andromeda's reach expansion behavior means frequency climbs differently depending on audience size. Small audiences (under 300K) saturate in terms of effective reach much faster than the frequency number suggests. Large audiences (1M+) may show a very low reported frequency while still having a burned-out core.

A better trigger is engagement decay: when your 7-day rolling CTR has dropped 20%+ from the ad's peak CTR. This is a direct measurement of audience response rather than an inference from delivery patterns.

Spend-based refresh windows as a starting point

Spend level gives you a reasonable baseline for creative production cadence, independent of performance signals. Here's what holds up across accounts:

$1,000-$5,000/month total spend

Plan for new creative every 2-3 weeks.

Andromeda's delivery is slower at this spend level and audience saturation happens more gradually. You have more runway per creative. Plan production in 2-3 week cycles.

Practical production target: 2-4 new creatives per cycle. These don't all need to be produced from scratch. A new hook on an existing winning concept counts. A UGC version of a polished ad counts. Variations, not always entirely new concepts.

$5,000-$20,000/month total spend

Plan for new creative every 1-2 weeks.

Andromeda burns through your relevant audience faster at this spend level. The window between early fatigue and performance collapse is often 48-72 hours rather than the 4-5 days you get at lower spend.

This is the range where most advertisers get caught. The old monthly creative refresh cadence that worked at $2K/month doesn't work at $10K/month. If your creative production hasn't scaled with your spend, this is likely where your efficiency losses are coming from.

Practical production target: 4-8 new creatives per week. At this level, you need a real creative pipeline, not ad-hoc production.

$20,000+/month total spend

Plan for new creative every week, with some ad sets needing twice-weekly additions.

High-spend accounts under Andromeda can fatigue creatives in 5-8 days at peak performance. At this level, creative production is a full-time operational function, not a periodic project.

Practical production target: 8-15 new creatives per week minimum. This requires a structured UGC pipeline, a clear brief system, and a creative team that understands performance metrics, not just aesthetics.

The three-tier creative system

Rather than thinking about "when to refresh," it helps to think about running three tiers of creative simultaneously:

Tier 1: Active performers Ads currently delivering within 15% of their peak performance. Leave these running. Do not touch them.

Tier 2: Early fatigue Ads where CTR has dropped 15-25% from peak over the last 7 days. Stage replacements now. Don't pause yet -- let them run while you have fresh creative ready.

Tier 3: Active fatigue Ads where CTR has dropped 25%+ from peak and CPM is rising. Pause these. Replace immediately with staged creative.

At any given time, you want to have enough creative in production to fill Tier 1 and have a full replacement set ready for Tier 2 and Tier 3. When a Tier 1 ad moves to Tier 2, a Tier 3 replacement moves in.

This system makes refresh a continuous process rather than a periodic event. You're never scrambling to produce new creative because the pipeline is always moving.

What counts as "new" creative

One reason creative production feels overwhelming is that people conflate "new creative" with "entirely original concept." At the practical level, you need far less than that.

Full new concept (1-2 per month at most spend levels) New product angle, new storytelling approach, genuinely different creative direction. These are expensive to produce and time-consuming to brief. Don't plan for many.

Format switch (high value, medium effort) Taking your best-performing static ad and turning it into a UGC video. Taking your best-performing video and converting key frames to a carousel. Format switches often outperform minor creative variations because they reach different inventory and register differently in the feed.

Hook variation (low effort, high leverage) Keep the body of the ad the same. Change the first 3 seconds. Different opening line, different first image, different visual environment. This is the fastest way to extend the life of a winning concept without a full production cycle.

Angle variation (medium effort, medium value) Same product, different reason to buy. Price angle, testimonial angle, feature angle, problem-first angle. This extends creative mileage significantly and requires only copy and basic visual adjustments.

Most of your "new creatives" should be hook variations and angle variations. Full new concepts are for when your entire hypothesis about what the market wants needs to change.

The creative backlog rule

Never let your active creative backlog go below 2 weeks' worth of untested ads.

If you're refreshing weekly, you need at least 2 weeks of staged, ready-to-launch creative at all times. This buffer means that when an ad set goes into active fatigue at 9pm on a Friday, you can swap in fresh creative immediately rather than producing something in a panic over the weekend.

Building this backlog is a short-term production investment. You need to over-produce for a few weeks to get ahead of your refresh cycle. Once you're ahead, maintaining the backlog requires only keeping pace with your refresh cadence.

If you're refreshing reactively after performance drops, your first priority is getting ahead. Over-produce for two weeks until you have the staging buffer.

Identifying when fatigue is creative vs. audience

Before refreshing, confirm which type of fatigue you're dealing with. Creative fatigue means your ads are burned out. Audience fatigue means you've run out of reachable audience.

Signs it's creative fatigue:

• CTR declining but impression volume holding steady
• New creatives added to the same ad set immediately outperform the older ones
• Broadening the audience partially restores performance

Signs it's audience fatigue:

• Impression volume declining, not just CTR
• New creatives perform just as poorly as old ones in the same ad set
• Frequency is high and your audience is under 200K

For audience fatigue, new creative won't fix it on its own. Expand your audience at the same time: increase lookalike percentages, add interest groups, or test Broad targeting and let Andromeda find new pools autonomously.

How Campaiyn tracks this for you

Monitoring hook rate, CTR decay, CPM trends, and frequency clustering across every active ad with enough precision to catch fatigue 3-5 days early is not a reasonable manual task at scale.

Campaiyn's fatigue monitoring does this automatically. It watches your ad performance continuously, flags ads entering early fatigue before CTR has collapsed, and gives you a clear picture of what needs attention today versus what can wait.

Run a free scan on your account to see which ad sets are in early warning territory and get a recommended refresh window based on your spend level and audience data.

Refresh frequency matters less than having a system. The accounts that perform consistently always have the next wave staged. When an ad dies, something is ready within 24 hours. That's the whole difference.