Negative keywords: the cheapest ROAS win in Apple Search Ads

Everyone wants to improve ROAS by finding better keywords or bidding smarter. Both are real, but both cost money and time. There’s a lever that costs neither: stop paying for the queries that were never going to convert in the first place.

Negative keywords are the most under-used tool in Apple Search Ads, and for an indie subscription app they’re often the single fastest ROAS win available.

Where the waste comes from

Your discovery layer — Search Match and broad match — is doing its job: matching your app to queries you never explicitly targeted. Most of those matches are fine. Some are quietly bleeding you:

• Queries for a feature you don’t have.
• A competitor’s brand name where users want their app, not yours.
• “Free” intent when you’re a paid subscription.
• Adjacent-but-wrong categories (a meditation app matching “music player”).
• Plain irrelevant terms that match on a loose word overlap.

Each of these costs you taps, and every tap costs money whether or not it ever converts. Left alone, they don’t fix themselves — Apple keeps matching them as long as they’re profitable for Apple.

How to find it

Open your search-terms report and sort by spend. You’re hunting for one specific pattern:

Meaningful spend, plenty of taps, and zero (or near-zero) installs.

That’s not bad luck — that’s a query that doesn’t want your app. The taps prove people are seeing you; the absence of installs proves they bounce. Add it as a negative and you reclaim that spend immediately.

A second, subtler pattern: terms that do convert to installs but never to trials or paying subscribers. For a subscription app, an install that never starts a trial is worth roughly nothing. If you can see down to the subscriber level, these are negative-keyword candidates too — they look fine on an install-based dashboard and terrible on the economics.

Build the habit, not the one-time cleanup

The mistake is treating negatives as a spring-clean you do once. Discovery keeps surfacing new junk as the market shifts, your category changes, and seasonal queries come and go. A stale negative list goes stale fast.

Make it a weekly ten-minute loop:

1. Sort the search-terms report by spend.
2. Flag anything with spend and taps but no conversions.
3. Add those as negatives at the right level (ad group for narrow, campaign for broad waste).
4. Note recurring themes — if “free” keeps showing up, a single negative covers a whole class of waste.

Where to place negatives

Placement matters more than people think:

• Ad-group negatives keep a term out of one ad group while letting it run elsewhere — use these when a query belongs to a different theme you target.
• Campaign negatives kill a term across the whole campaign — use these for genuine junk that should never match anywhere.

There’s also a structural use: campaign-level negatives prevent your discovery campaign from competing with your exact campaigns for terms you’ve already harvested. That stops you bidding against yourself.

Don’t over-prune

One caution: a query with a handful of taps and no installs yet isn’t proof of failure — it might just lack data. Give low-volume terms enough taps to judge fairly before you cut them, or you’ll negative away winners that simply hadn’t converted yet. Spend is the signal that matters; a term that’s barely spent anything isn’t urgent.

The whole point is asymmetry. A negative keyword costs you nothing and stops a leak forever. Compared to chasing new keywords or pushing bids, it’s free money sitting in your search-terms report waiting to be claimed.

TapPilot surfaces wasted-spend terms straight from your search-terms data and lets you add negatives in a click, so the weekly hygiene loop takes minutes instead of a spreadsheet session. See the search-terms report.