Placement Mix Audit: How to Stop Wasting Apple Search Ads Budget on the Wrong Surfaces

Apple Search Ads doesn’t just match keywords to search queries—it also decides where your ads appear. If you’re getting “okay” ROAS but it won’t scale, or your CPA looks unpredictable, your placement mix might be the silent culprit.

This post is a placement audit you can run as an indie dev: compare performance by placement, identify which surfaces are producing installs (not just taps), and then reallocate budget and bids toward the placements that actually convert for your app.

Why placement mix messes with ROAS (even when keywords are fine)

Apple Search Ads uses a CPT auction (cost-per-tap) with a max CPT bid. Within that auction, Apple can place your ad on different surfaces, including:

Search Results (most indies start here)
Search tab
Today tab
Product Pages (browse)

Here’s the core problem: taps are not created equal by surface. A tap on Search Results is often closer to purchase intent than a tap from browse-style discovery. That means the same keyword set can yield very different:

TTR (taps / impressions)
Conversion rate (installs / taps)
CPA/CPI (cost per install)
ROAS (revenue / spend)

And because ROAS comes from the install → purchase chain, placement-induced intent differences can dominate what you see in reports.

Step 1: Pull a placement snapshot (and normalize what you compare)

Before you touch anything, get a clean table of performance by placement for a consistent time window.

In your ASA reporting, export metrics by placement for (for example) the last 14–30 days:

Impressions
Taps
TTR
Installs
Conversion rate (installs / taps)
Spend
CPT
CPA/CPI
Revenue and ROAS

Normalize your interpretation:

If Search Results has lower ROAS but better taps, you can’t conclude “keywords are bad.” It might be that users are tapping, then failing to convert.
If Today tab has higher ROAS but tiny spend, you can’t conclude “it’s a winner.” You need enough volume to make a bid shift worth the risk.

A simple mental model:

Bad CPA + high TTR → you’re paying for clicks that don’t install.
Good conversion rate but high CPT → bid/surface auction positioning is too expensive.
Good install rate but bad ROAS → likely revenue/offer mismatch (trial, subscriptions, refunds), not conversion friction.

Step 2: Classify each placement into one of four buckets

For each placement surface, label it based on two questions:

Does it convert well? (conversion rate)
Does it monetize well? (ROAS)

That gives you four buckets:

A) “Efficient + Profitable”

High conversion rate
Healthy ROAS

Action: Protect delivery. Focus budget/bids here.

B) “Attracting Taps, Not Installing”

High TTR
Low conversion rate

Action: Fix relevance on the surface—usually product page (screens, custom page, offer) and keyword intent. Also consider bid caps.

C) “Installing, Not Monetizing”

Lower ROAS
Decent conversion rate

Action: Check your offer fit (pricing, subscription/trial messaging), onboarding flow, and whether your revenue reporting window makes it look worse.

D) “Neither converts nor profits”

Low conversion rate and weak ROAS

Action: Reduce waste by lowering bids or restricting exposure via keyword strategy.

Important: you can be in Bucket B and still “look okay” in ROAS if the CPT is low. Don’t let one metric hide the other.

Step 3: Identify your likely placement root cause

Now connect the bucket to the most common underlying levers.

Search Results (often intent-rich)

If Search Results is underperforming, it’s usually one of:

Keyword relevance is off (your query set is too broad or mismatched)
Landing experience (default vs custom product page, screenshots, first-screen clarity)
Bid is winning taps that don’t convert (high CPT with low conversion)

Search tab

This often behaves like “search-adjacent discovery.” If Search tab is weaker than Search Results, it’s frequently a keyword intent mismatch or product page mismatch for less-specific queries.

Today tab

This surface usually skews “what’s trending / editorial-like discovery.” Underperformance here is commonly:

You’re paying for clicks from users with lower intent
Your product page doesn’t clearly communicate why your app is worth trying immediately

Product Pages (browse)

Browse-style placements can be a strong growth lever if your app is well-positioned for related audiences. If it’s weak, it’s often because:

Your keywords are driving browse behavior you don’t want
Your app’s positioning doesn’t align with how users discover similar apps

Step 4: Reallocate budget and bids without “thrashing” delivery

Here’s the trap: if you immediately slash bids everywhere, your learning and delivery can wobble and you won’t know what changed.

Instead, use a controlled reallocation strategy:

Rule: Move one dimension at a time

Pick either:

Bid changes (max CPT) for specific ad groups / keyword sets
Keyword exposure (which keywords can trigger ads)
Product page destination (default vs custom product page) — especially for high-traffic ad groups

Concrete approach

Choose your “anchor” placements: typically Search Results + whichever of Search tab / Today tab / Product Pages is in Bucket A.
Reduce exposure to Bucket D first, but only in the ad groups that are responsible.
Keep Search Results stable long enough to preserve comparability.

Because your exact ad group structure matters, the practical way is:

If you run separate ad groups for branded vs non-branded, or for exact vs broad, you can adjust bids at that granularity.
If Search Results is working but Today tab isn’t, don’t assume keywords are broken—use bid/keyword structure to reduce expensive low-intent taps.

Step 5: Use product-page alignment to improve conversion on weaker placements

Placement mix problems often look like keyword issues but show up as conversion friction.

Since Apple Search Ads has no “creative auction edge,” your screenshots/video won’t help you win the auction. But they can absolutely change installs / taps.

A checklist that’s worth doing specifically for placements that underperform:

Does the first screen answer “what is this” in 5 seconds?
Is the value proposition consistent with the user’s intent? (query → app description)
If you use custom product pages, confirm you’re sending the right ad group to the right page for the intent level.

A quick diagnostic: if a placement has low conversion rate, compare the app store page feel for that surface. Even if the metrics are noisy, persistent gaps are a signal.

Step 6: Decide using a decision threshold, not vibes

ROAS can swing, and placement performance can be low-volume. To avoid overreacting:

Only act decisively when you have enough spend on each placement to trust direction.
If a placement is underfunded, it might be Bucket A but you simply don’t have volume.
Treat bid changes like experiments: one change, observe for at least a week (or until attribution settles; purchase tracking through AdServices can resolve within ~24h, but real revenue patterns can still lag).

A note on attribution: don’t blame placements for reporting artifacts

Apple Search Ads attribution flows through AdServices tokens and is resolved around a day later. RevenueCat (or your own mapping) then ties installs to purchase outcomes.

Two common reasons placement ROAS appears “weird”:

Refunds/chargebacks or net revenue reporting differences
Delayed value (subscriptions, trials converting later)

So: if Today tab shows weak ROAS but strong install conversion, verify your revenue window matches your business model. Otherwise you’ll cut a placement that’s actually working on a delay.

Where AdsBuddy can fit (optional)

If you’re already pulling exports and still unsure what to change first, a tool like AdsBuddy can help by reading your ASA performance and revenue mapping and then producing a short prioritized list of placement/keyword/product-page changes for you to approve (no auto-spend).

Closing takeaway

A lot of Apple Search Ads optimization starts at the keyword level, but placement mix can override everything. Audit by placement, bucket the surfaces by conversion + monetization, and then reallocate budget and bids in a controlled way—while aligning the product page to the intent level each surface tends to bring.

If you do that, “bad ROAS” stops feeling random—and starts looking like a solvable targeting and landing problem.