Seeding Max CPT for New Keywords: How to Set a First Bid Without Burning Budget

Every new keyword in Apple Search Ads starts with the same problem: you have to pick a max CPT before you know anything about it. Set it too high and you pay for taps from people who were never going to subscribe. Set it too low and the keyword never gets enough impressions to teach you anything, so you write it off as dead when the real issue was your bid.

This post walks through a practical way to seed a first bid. The method is simple arithmetic built on your own install economics, followed by a disciplined review schedule so the bid moves on evidence instead of anxiety.

Understand what max CPT actually does

Your max CPT is a ceiling, not a price. Apple runs an auction for each eligible ad slot, and you pay the clearing price up to your max. If your max is higher than what the auction requires, you usually pay less than your max. If your max is below what the auction requires, you may get few or no impressions.

That means a high max does not automatically mean you overpay. It means you are willing to win more auctions. The real question is whether the taps you buy are worth what you pay for them.

Work backward from what an install is worth

Before you touch a bid, you need three numbers from your own data, not from forums or generic benchmarks:

• Expected net value per install. Take the revenue your app has earned from installs over a window you trust, after refunds and store fees, and divide by installs from the same window. If your revenue data comes through a tool like RevenueCat, use the install cohort view rather than a raw daily total, so the timing lines up.
• Your tap-to-install conversion rate for similar intent. Look at existing keywords that are close in meaning to the new one. Do not use your overall account average, because a brand term and a category term will behave very differently.
• The share of value you are willing to spend on acquisition. This is a business decision. Some developers accept a cost per install close to the expected value because they are optimizing for growth; others want a clear margin.

With those, you can compute a target cost per install (CPI) and then a max CPT:

Max CPT = Target CPI × Expected tap-to-install rate

Here is an illustrative example, not a benchmark. Suppose your expected net value per install is $3.00, and you decide your target CPI is $1.50. If similar keywords convert taps to installs at 60 percent, your max CPT would be $1.50 × 0.60 = $0.90. If the same keyword converted at 30 percent, the math says you should only bid $0.45 to hit the same CPI. The conversion rate matters as much as the value.

Seed below the ceiling, not at it

A good starting point is a fraction of your computed max, not the max itself. The exact fraction is a judgment call. What matters is that you start with a bid you can justify and that you write down the logic, so you can check it later.

Set the bid, record the date, the computed ceiling, and the conversion assumption you used. This tiny log is worth more than people expect, because in two weeks you will forget why you chose that number.

Decide your evidence threshold before you start

The most common seeding mistake is reacting too early. Two days of data on a low-volume keyword tells you almost nothing. Before you launch the keyword, decide how much evidence you need to make a first decision. For many small apps, that means a minimum number of taps, and you should accept that the error bars are wide when the count is small.

A practical rule: pick a tap count and a minimum number of days, whichever is longer. Do not change the bid until you hit both. If the keyword is getting taps but no installs by that point, that is a signal to look at the product page and the query fit before you conclude the bid is wrong.

Read the signals in the right order

When the threshold is reached, review the metrics in this order, because each one depends on the one before it:

1. Impressions. If you have almost none, the problem is usually eligibility or bid level, not creative. Check whether the bid is too low relative to what the auction is requiring.
2. Tap-through rate (TTR). Low TTR on a keyword with real impressions suggests your ad is not matching what the searcher expected. For Search Results, your listing is the ad, so this points to the title, icon, and visible screenshots.
3. Tap-to-install conversion. If TTR is fine but installs are weak, the product page is not closing the gap for that intent.
4. Average CPT compared with your max. If average CPT sits right at your ceiling, you are paying the top of the auction, which can mean you are competing for the most expensive taps.
5. Later revenue. Installs are not the finish line. If your subscription or trial window is longer than your review window, judge the keyword on the cohort that has had time to convert.

Move the bid in small steps

When you change a bid, change one keyword or one tightly related group at a time, and make the step modest. A large jump makes it impossible to know what caused the result. Also avoid changing the bid and the product page in the same week, because you will not be able to tell which one moved the numbers.

A reasonable approach is to raise a bid only when impressions are clearly constrained and the conversion math still holds at the new level. Lower it when the CPI is above your target for long enough to trust the data and the product page has already been reviewed.

Mistakes to avoid

• Copying someone else's bid. Their app, price, offer, and conversion rate are different from yours.
• Starting at the ceiling on day one. You are paying for certainty you do not have yet.
• Killing a keyword after a few quiet days. Low volume is not the same as a failed keyword.
• Ignoring the exact and broad overlap. If the same term runs in two places, you may be bidding against yourself and muddying the read.
• Forgetting to write it down. Without a log, every review restarts from zero.

Takeaway

A new keyword's first max CPT should come from your own install economics: value per install, expected conversion from taps, and the acquisition cost you accept. Seed below the ceiling, decide your evidence threshold up front, read the metrics in order, and change one thing at a time. The goal is not a perfect first guess. It is a bid you can explain and a process that corrects it without panic.

If you want help keeping this kind of decision log and surfacing which keywords have hit their review threshold, AdsBuddy reads your Apple Search Ads and revenue data and suggests a short, prioritized list of changes each day. You still review and approve every change yourself.