AppSkale.AI
Back to blog

How Long Should You Run an Apple Search Ads Keyword Before Deciding?

By Sam H

A keyword spends for three days. No trials. You pause it. A week later you realize the sample was tiny and the trial window had not even started. Or the opposite: you leave a dead term running for a month because you wanted “more data.”

Timing is part of keyword judgment. This post gives a practical wait window for indie Apple Search Ads accounts before you pause, keep, or scale.

Decide on evidence, not calendar anxiety

You need two clocks:

  • Auction sample clock. Enough taps and spend that the keyword was actually tested.
  • Conversion clock. Enough calendar time for installs to become trials and trials to become paid.

Either clock alone lies. Spend without time misses delayed revenue. Time without spend is just waiting on an empty keyword. See trial started, ROAS didn’t move and daily budget.

A practical timeline for Exact keywords

  1. Days 1–3: Confirm delivery. If a term gets almost no impressions on a reasonable CPT, the problem may be bid or volume, not conversion. See max CPT bidding.
  2. Days 4–14: Look for trial starts and obvious intent mismatches in search terms. Pause clear junk early. See search terms report.
  3. After one full trial cycle: Make keep/pause/scale calls using trials and early revenue, not CPI vanity. See CPT vs CPI.
  4. After 30–60 days: Re-check ROAS including conversions that lagged. That is when scale decisions get serious.

For a seven-day trial, do not declare ROAS failure on day five. For a thirty-day trial, push every phase later.

Pause earlier when intent is obviously wrong

You do not need a full trial cycle to negative android queries on an iOS-only app, or job-seeker terms on a consumer tool. Obvious mismatches can die as soon as they show repeated waste. See negative keywords and pause vs lower bids.

The rule: pause early for bad intent. Wait longer for uncertain economics on plausible intent.

What “enough sample” means on a small budget

Indie accounts rarely get enterprise sample sizes. Use relative judgment:

  • A keyword with meaningful spend versus your daily cap and zero trials after two weeks is usually enough to pause
  • A keyword with three taps is not enough to crown or kill
  • Brand can often be judged faster than category because intent is cleaner
  • Broad and discovery need search-term harvest, not a single keyword average. See discovery campaigns.

A weekly decision checklist

  1. Sort by spend.
  2. Mark obvious waste for pause or negative.
  3. Leave low-sample rows alone.
  4. For high-spend plausible terms, check whether you are still inside the trial window.
  5. Only then pause, bid-cut, or raise budget.

Native ASA will not show subscription lag cleanly beside those rows. AppSkale puts keyword spend next to RevenueCat revenue so wait-vs-kill decisions follow paying users.

Where to go next

Give keywords enough auction sample and enough conversion time. Kill bad intent early. Judge economics after your trial cycle, not after a panic Monday.

Related reading: ASA for beginners, ROAS with RevenueCat, and the attribution setup guide. When you want those clocks on one screen, AppSkale connects ASA to RevenueCat.