Setting a Creative Iteration Cadence for Live Mobile Games

Ask five UA teams how often they refresh creative for a live title and the answers will span from weekly to quarterly, often with equal confidence behind each. Both can be right, because a creative iteration cadence is not a universal number. It is a function of spend level, genre volatility and how fast a team can actually produce and test new concepts without shipping noise instead of signal. The mistake most teams make is picking a cadence based on what a competitor or a case study reports, rather than what their own production capacity and spend level can actually support.

Why a creative iteration cadence is a capacity question first

A weekly refresh cadence only works if a team can produce enough distinct, testable creative variants each week to generate a meaningful read before the next refresh arrives. A team spending 20,000 dollars a week on a given creative concept, with a genre that supports fast conversion signals, can plausibly read performance within days. A team spending 2,000 dollars a week on a niche genre with a longer conversion window cannot generate a statistically usable read on that timescale, no matter how fast its creative team can produce new assets. Setting a weekly cadence in the second case does not produce faster learning. It produces decisions made on noise, which is worse than making them on a slower, more reliable rhythm.

Reading fatigue signals instead of guessing a schedule

Rather than fixing a calendar cadence in advance, a more reliable approach ties refresh timing to observed fatigue signals: rising CPI on a stable audience and bid strategy, falling click-through rate on creative that has not changed, and a widening gap between impression frequency and conversion rate for repeat viewers. When two or more of these move in the same direction over a sustained period, that is a stronger trigger for refresh than a date on a calendar, because it reflects what the audience is actually telling the team rather than an assumption made weeks earlier about how fast fatigue would set in.

This does not mean cadence planning is pointless. A team still needs a production calendar to make sure new concepts exist before fatigue signals appear, not after. The distinction is between planning production capacity on a schedule and triggering the actual swap based on data.

The two approaches fail differently when misapplied. A pure calendar cadence, refreshing creative every two weeks regardless of performance, wastes production effort replacing concepts that were still working, and can introduce noise into performance tracking when a fresh creative underperforms a proven one simply because it has not had time to find its audience yet. A pure signal-based approach, with no calendar planning at all, risks a gap between detecting fatigue and having a replacement ready, during which CPI drifts upward with nothing in the pipeline to fix it. The two need to run together: the calendar sets the pace of production, the signals decide when a produced concept actually enters rotation.

A framework by spend tier

A practical way to set expectations without pretending one number fits every team:

• High-spend live titles, generally above 100,000 dollars a month in a single creative concept, can usually support a genuine weekly refresh cadence on top-performing concepts, because volume produces a usable read fast enough to justify the production cost.
• Mid-spend titles, in the 20,000 to 100,000 dollar range, typically get more reliable signal on a two to three week cadence, giving each concept enough impression volume to separate real fatigue from normal weekly variance.
• Lower-spend or niche-genre titles often do better on a monthly cadence, focusing production effort on fewer, higher-conviction concepts rather than spreading a limited testing budget across variants too thin to read cleanly.

These bands are illustrative starting points, not fixed rules; a team's own historical time-to-signal for its specific genre and audience should adjust them up or down. A title with a long conversion window, where the events that matter for value most often happen well past the first week, needs more time between refreshes than these bands suggest regardless of spend level, simply because the read on whether a concept is genuinely fatiguing takes longer to surface. A title with a short, high-frequency conversion loop can sometimes justify a faster cadence than its spend tier alone would indicate, because each impression carries more diagnostic value per dollar spent.

Building the production calendar around the cadence

Once a tier is chosen, the practical work is building a production calendar that keeps a bench of tested-but-unused concepts ready before fatigue hits, rather than starting production only once performance has already dropped. A rough target worth setting: always have at least one fully produced, untested concept ready to enter rotation, so a fatigue signal triggers a swap rather than a production scramble. Teams that build to this buffer consistently report shorter gaps between detecting fatigue and having a replacement live, which is usually the single biggest lever in avoiding a CPI spike that lingers for weeks.

Closing the quarter on the right rhythm

As Q2 planning starts, this is a good moment to check whether a team's actual cadence matches what its spend and genre can support, rather than a cadence set a year ago under different conditions, before spend levels, competitive pressure or the platform's own bidding models have had a chance to shift underneath it. As we discussed in Detecting Creative Fatigue Before It Shows Up in CPI, the cost of a mismatched cadence rarely shows up as an obvious failure. It shows up as a slow CPI drift that a team spends weeks attributing to the wrong cause before checking whether the creative simply ran too long.

Related archive reading

These articles provide related context and remain subject to their stated review status.

• Creative concept portfolio sizing for a flat budget
• Q4 Creative Capacity Planning: How Much Is Enough