Ad Creative Refresh Cadence Framework

AI Vidia sets ad creative refresh cadence by spend and signal, not by the calendar, so new variants enter live ad sets every week and every concept gets a full refresh inside 30 days. Ad creative refresh cadence is the fixed rhythm at which new creative enters a paid social account, measured in variants shipped per week and concepts replaced per month rather than campaigns launched per quarter. Meta for Business reports that campaigns running five or more creative variations see 30 to 50 percent lower CPA, which is the argument for a weekly cadence in a single number. AI Vidia has shipped 1,834 AI videos and 70,342 AI images across 48 brands in 14 countries on this rhythm, and the tested winners return a median 2.4x ROAS.

What the wrong refresh cadence costs

A refresh cadence that is slower than the fatigue curve makes the account pay twice. On a scaling budget, a winning ad holds its performance for roughly two to four weeks before frequency climbs, click rate slides, and CPA drifts upward. A brand refreshing once a quarter therefore runs fatigued creative for eight to ten weeks out of every thirteen, paying a rising CPA the entire time while it waits for the next drop. The waste is not in the creative that failed. It is in the winner that kept spending after it stopped working.

Put a number on it. A brand spending EUR 40,000 a month on Meta and TikTok that lets CPA drift 20 percent for six weeks between refreshes burns roughly EUR 12,000 of media on creative the account had already stopped responding to. Forrester reports a 20 to 35 percent paid media ROAS improvement when creative volume rises, and the gap between a weekly and a quarterly cadence is exactly where that improvement sits. Wyzowl reported in 2025 that 91 percent of businesses use video marketing and 30 percent name production cost as the top barrier to making more of it, which explains why so many teams set cadence by what they can afford to produce instead of by what the account needs.

The second cost is signal. Meta ad sets need 30 to 50 conversion events per week to exit the learning phase, and a starved account never gets there because each refresh resets the clock instead of feeding a live cohort. The Content Marketing Institute reported in 2025 that 73 percent of B2B marketing teams cite producing enough content as their biggest challenge, and cadence is the part of that number that shows up directly in media efficiency. McKinsey reports a 3 to 5x output increase from AI in creative production, but output only helps if it arrives on a schedule the ad account can absorb.

Refresh cadence by spend tier, compared

Cadence is a function of spend, ad set count, and audience size, not of team preference. The more budget an account pushes through a fixed audience, the faster frequency climbs and the sooner creative fatigues. The table below sets the cadence AI Vidia runs at each spend tier, with the variant volume and concept refresh window that go with it.

Read the table as a floor, not a ceiling. Under EUR 10,000 a month, a two week variant cadence is usually enough because the audience is large relative to the spend and frequency climbs slowly. Between EUR 10,000 and EUR 30,000, weekly becomes the break point: this is the tier where most Nordic DTC brands first notice CPA drift they cannot explain, and it is almost always cadence rather than targeting. Above EUR 30,000 the account needs two batches a week and a 30 day concept window, because a single weekly drop no longer covers the number of live ad sets. Above EUR 75,000, refresh is continuous in practice, and the operating question changes from when to refresh to how many variants the production line can hold at brand quality.

The frequency ceiling column is the tripwire that matters most. When seven day frequency crosses the number in that row, the account is showing the same creative to the same people often enough that response falls, and no bid adjustment fixes it. AI Vidia treats that ceiling as a hard trigger, not a guideline, because waiting one more week past it typically costs more in wasted media than the entire batch that would have replaced it.

The AI Vidia Refresh Trigger Threshold Model

This is the strategic model that decides when a refresh is actually due. It replaces the calendar with four measured thresholds, so a refresh happens when the account asks for it rather than when the quarter ends. AI Vidia reads all four every Monday for every brand in flight.

1. Set the frequency ceiling per tier. Pull seven day frequency at the ad set level and compare it to the ceiling for the account's spend tier. Crossing the ceiling means the audience has seen the current creative enough times that the next impression is worth less than the last one, and that is a refresh trigger on its own.
2. Measure hook rate decay against week one. Every concept gets a week one baseline for three second view rate, and the trigger fires when the concept drops more than 15 percent below its own baseline. Comparing a concept to itself rather than to the account average separates genuine fatigue from a concept that was never strong.
3. Track CPA drift on a rolling seven day window. When CPA on a cohort runs 20 percent above its own trailing four week average for two consecutive weeks, the creative is the most likely cause, and refreshing is cheaper than restructuring the campaign. Two consecutive weeks is the threshold that filters out normal weekly noise.
4. Check the share of spend on aging creative. Sum the spend running on assets older than 30 days: above 40 percent of account spend, the refresh is overdue regardless of what the other three signals say. This is the check that catches a slow leak, because an account can look stable while quietly concentrating budget on three tired winners.
5. Assign the refresh depth. Each trigger maps to one of three responses: a variant refresh keeps the concept and changes hooks or cuts, a concept refresh retires the idea and briefs new ones, and a system refresh rebuilds the style lock itself. Naming the depth before briefing stops the team from rebuilding a whole system when three new hooks would have done the job.

The model works because it makes the refresh decision boring. When the thresholds are fixed and written down, nobody argues about whether an ad still has life in it, and the brief goes out the same day the trigger fires.