Match-Type Isolation: Stop Confusing Exact, Broad, and Search Match When You Optimize Apple Search Ads

If you run Apple Search Ads yourself, you’ve probably felt this frustration: you make a small change to improve CPI or ROAS, and then the data looks “muddy.” Sometimes it’s not because the change failed—it’s because different match types are competing and blending impressions, taps, and installs in the same reporting bucket.

A clean way to regain control is match-type isolation: structure your account so exact, broad, and Search Match each live in their own ad groups (and ideally their own campaigns by country). Then your optimization decisions stop being guesswork.

Why match types get you stuck (even when your bids are sensible)

On Search Results keywords, Apple Search Ads can run:

• Exact: tighter query control (usually higher intent, often higher conversion rate).
• Broad: expands to related queries (more volume, more variability).
• Search Match (automatic): Apple matches based on your relevance signals (discovery, but you can’t directly control the query list like exact/broad).

The problem isn’t that match types “don’t work.” The problem is that each match type produces a different mix of queries and user intent. When they’re blended together, the metrics you use for decisions—TTR (taps/impressions), conversion rate (installs/taps), CPT, and eventually CPI/ROAS—become a weighted average of multiple behaviors.

What blended data looks like

Common misleading patterns:

• Conversion rate drops but CPI doesn’t: broad/automated added volume that didn’t convert, but exact stayed stable.
• CPI improves but ROAS worsens: you may have attracted cheaper taps that convert to lower-value purchases.
• TTR changes mysteriously: match type mix shifted (or auction pressure changed), not necessarily your creatives or bids.

When you can’t tell which match type changed the outcome, you can’t confidently keep changing the right lever.

The goal: one decision = one match type

Match-type isolation is simple in principle:

• Put Exact keywords in one logical bucket.
• Put Broad keywords in another.
• Put Search Match in its own ad group (often its own keyword set/campaign structure).

Then when you adjust bids (or decide to pause), you do it with an understanding of the “audience you’re touching.”

A practical account structure that stays readable

Here’s a structure that works well for indie apps and small studios.

1) Use one country per campaign (don’t mix regions)

Apple Search Ads is campaign = country/region. If you’re already correct here, keep it that way. Match-type isolation should happen within a country.

2) For each country, create three ad groups

Within the same campaign (country):

• Ad group A: Exact Only

  • Keywords: your core phrases as exact match type.
  • Bid: set an initial CPT cap based on what you’re comfortable paying for high-intent traffic.

• Ad group B: Broad Only

  • Keywords: the same (or expanded) phrases as broad match type.
  • Bid: typically lower than exact if your goal is to avoid broad “tax,” but you’ll learn from actual conversion rate.

• Ad group C: Search Match

  • Let discovery happen here instead of letting it leak into your exact/broad reporting.
  • If you have a separate ad group for Search Match, it becomes much easier to evaluate whether discovery is buying users who actually convert.

3) Don’t reuse the same keyword across match types

This sounds obvious, but it matters:

• If you include the same term in both exact and broad, you can still get signal—but you lose clarity.
• Prefer to treat match types as separate “experiments” that happen simultaneously.

If you want to test multiple keyword themes, keep those themes inside the match-type bucket, but keep match type itself isolated.

How to isolate without creating chaos: naming and tagging

Apple Ads reporting won’t automatically tell you “this is Exact vs Broad vs Search Match” in every table. So make it easy.

Use a consistent naming convention:

• AG-Exact-Core
• AG-Broad-Core
• AG-SM-Discovery

And when you check performance, always filter by ad group first. Don’t jump straight to campaign-level ROAS if you care about what to change.

What decisions become easier once you isolate

With isolation, each optimization question maps to one bucket.

1) “Is the traffic interested?” → look at TTR per match type

• If Exact has low TTR: your keywords aren’t aligning with what people search, or your product page doesn’t match expectations.
• If Broad has high TTR but low installs: taps are happening, but they’re not converting.
• If Search Match has erratic TTR: discovery is working differently than your core intent.

You can then target the right fix:

• Keyword refinement (exact keyword list)
• Broad throttling (reduce CPT cap or pause specific broad terms)
• Product page alignment (custom product page for discovery intent if you use that approach)

2) “Is the store converting?” → look at conversion rate per match type

Compute conversion rate using installs/taps (Apple reports these at the ad group level).

• If conversion rate is strong in Exact but weak in Broad: broad is attracting less-qualified queries. Lower broad bids or tighten broad keyword themes.
• If Search Match conversion is weak: discovery may be too broad for now; cap CPT lower, or temporarily pause Search Match while you harden your exact/broad set.

3) “Which lever should I touch?” → change only what matches the root cause

Isolation reduces the blast radius of each change.

Examples:

• Exact CPI is high and conversion rate is low → adjust exact keyword targeting and/or product page.
• Broad CPI is acceptable but installs are low → broad may need a different bid strategy (or a different keyword set).
• Search Match ROAS is volatile → treat it as discovery, not core budget. Control it with a lower CPT cap and evaluate after attribution resolves.

A workflow for your weekly checks

Here’s a repeatable routine that takes advantage of match-type isolation.

1. Wait for attribution to settle Apple resolves installs via AdServices; if your revenue mapping uses RevenueCat or similar, still expect some lag. Don’t decide based on half-resolved ROAS.

2. Check ad group level metrics first For each match-type ad group, review:

   • TTR
   • Conversion rate (installs/taps)
   • CPT
   • CPI
   • ROAS (once revenue attribution has caught up)

3. Classify the bucket behavior

   • Exact: your “quality bar.”
   • Broad: your “volume engine.”
   • Search Match: your “learning/discovery lane.”

4. Make one primary adjustment aligned to the worst bucket Don’t simultaneously change exact and broad bids unless you’re deliberately shifting mix.

5. Keep a short decision horizon If your goal is stabilizing performance, avoid overreacting to daily noise. Use consistent intervals for changes and evaluation.

Common pitfalls (and how to avoid them)

Pitfall: believing campaign-level ROAS

If your campaign blends all three match types, campaign ROAS can hide a truth like:

• Exact is great, broad is dragging.
• Search Match is cheap CPT but poor purchase conversion.

Fix: always pull up ad group metrics.

Pitfall: isolating but not resetting expectations

Match-type isolation doesn’t guarantee better ROAS automatically. It guarantees you’ll know what you changed.

To get value, you still need to:

• Make bid changes only in the bucket you understand.
• Keep your evaluation window consistent.

Pitfall: Search Match “stealing” intent

Search Match can bring valuable installs, but it can also buy uncertainty. Isolation helps you see when it’s behaving like discovery (good) vs like random exploration (bad for your current goal).

If your core ROAS depends on predictable buyers, treat Search Match as controlled experimentation rather than a forever-on core.

Where AdsBuddy can help (lightly)

If you want a second set of eyes, tools like AdsBuddy can read your Apple Search Ads + revenue and output a daily prioritized list of changes. But regardless of automation, match-type isolation is the prerequisite that makes any advice actually trustworthy.

Closing takeaway

Match-type isolation is one of the highest-leverage structural tweaks you can make in Apple Search Ads. By separating Exact, Broad, and Search Match into dedicated ad groups, you prevent blended reporting from misleading your optimization decisions.

Once your data is clean, bid changes and product page fixes become much more direct—and you stop “debugging” the wrong problem.