Apple Ads ROAS Guide: Ongoing vs Cohort ROAS Explained

ROAS (Return on Ad Spend) is the single most important metric in Apple Search Ads. Yet it is also one of the most misunderstood. Apple Ads has both an overwhelming number of surface-level metrics and surprisingly few truly useful ones. Understanding the difference between Ongoing ROAS and Cohort ROAS is what separates advertisers who panic-pause profitable campaigns from those who scale with confidence.

In this guide, I will break down how ROAS actually works in Apple Search Ads, why most advertisers read it wrong, and how to build an optimization philosophy around the metrics that actually matter. This is based on my experience scaling subscription apps to over $1M in revenue with Apple Ads.

The Data Problem in Apple Ads

Here is the biggest frustration with Apple Search Ads: there are no post-install statistics in the dashboard. Apple knows 100% of user data - who downloaded your app, who started a free trial, who converted to a paid subscription, and who renewed. They own the entire funnel. Yet they show you almost nothing beyond the install.

This means that out of the box, Apple Search Ads gives you impressions, taps, installs, and cost. That is it. You cannot see trial starts, paid conversions, revenue, or ROAS directly in the Apple Ads dashboard.

To get the metrics that actually matter, you need one of three things:

• A Mobile Measurement Partner (MMP) - AppsFlyer, Adjust, or similar tools that track post-install events and attribute revenue back to campaigns
• Revenue analytics platforms - RevenueCat, Qonversion, or Adapty that integrate with Apple Ads to provide subscription-level attribution
• Build your own attribution - Use Apple's AdServices Framework to build custom attribution, connecting ad taps to in-app purchases on your own backend

The 4 Metrics That Actually Matter

Once you have post-install data flowing, the noise clears. Out of dozens of available metrics, only four truly determine whether your Apple Ads campaigns are working:

1. Cost per Client (CPA) - How much you pay to acquire a paying customer, not just an install. This is your true acquisition cost.
2. Cost per Conversion - The cost to get a user who starts a free trial or converts to a paid subscription. This includes both trial starts and direct purchases.
3. ROAS (Return on Ad Spend) - The ratio of revenue generated to money spent on ads. A ROAS of 100% means you broke even; above 100% means profit.
4. Impression Share - The percentage of total available impressions your ads captured for a given keyword or market. This tells you how much room you have to scale.

You can simplify this even further. If you are an experienced advertiser, you can narrow your focus to just two metrics: ROAS and Impression Share. ROAS tells you how well your campaign pays back. Impression Share tells you how much you can scale. Everything else is a supporting detail.

But ROAS is not a single number. There are two fundamentally different ways to measure it, and confusing them is one of the most common mistakes in Apple Ads.

Ongoing ROAS: The Retrospective View

Ongoing ROAS is the cumulative return on ad spend that changes over time. It is what most people think of when they hear "ROAS" - the total revenue generated by a campaign divided by the total cost spent on it, measured up to the current date.

How Ongoing ROAS Works

Imagine you spent $10,000 on Apple Ads in September. In that same month, those users generated $4,000 in revenue. Your September Ongoing ROAS at the end of September is 40%.

But those users do not stop paying after September. Some are on weekly subscriptions that renew every week. Others are on monthly plans. By October, the September cohort has generated another $3,000 in revenue through renewals, bringing total revenue to $7,000. Your September Ongoing ROAS is now 70%.

By December, the same cohort has generated $12,000 total. Your September Ongoing ROAS is now 120% - profitable. And no new costs were added. The $10,000 was spent once in September, but the revenue from those users keeps accumulating.

When Ongoing ROAS Is Useful

Ongoing ROAS is excellent for retrospective analysis. It answers the question: "Did this month's spending eventually pay for itself?" It is the metric you use to evaluate historical performance and validate your payback window assumptions.

When Ongoing ROAS Misleads

Ongoing ROAS is not ideal for evaluating current campaign performance. Because it grows over time, a campaign that just launched will always look worse than one from three months ago. If you use Ongoing ROAS to make real-time decisions, you will constantly pause campaigns that are actually on track to be profitable.

Cohort ROAS: The Performance Benchmark

Cohort ROAS measures the return on ad spend for a specific group of users acquired on a specific date, tracked over a defined time window. This is the metric you should use to evaluate and optimize active campaigns.

How Cohort ROAS Works

Cohort ROAS uses a standardized time window to compare performance apples-to-apples:

• 1-day ROAS (D1) - Cost spent to acquire users on Day X, divided into the revenue those users generated on that same Day X. Useful for apps with immediate conversion (e.g., paid apps, one-time purchases).
• 7-day ROAS (D7) - Cost spent on Day X, divided into the revenue those users generated over the next 7 days. The standard benchmark for apps with free trials.
• 30-day ROAS (D30) - Cost on Day X, revenue over the next 30 days. Better for monthly subscription apps and longer trial periods.
• 90-day ROAS (D90) - Cost on Day X, revenue over 90 days. Used for apps with longer payback windows.

Why Cohort ROAS Is Better for Optimization

Cohort ROAS gives you a consistent, comparable benchmark. When you compare the D7 ROAS of Campaign A versus Campaign B, you are comparing them on equal footing - same time window, same maturity of data. This makes it possible to identify which campaigns, keywords, and geos actually perform better.

If your target D7 ROAS is 30% (because your model shows that users who hit 30% by day 7 will reach 100% by month 4), you can confidently evaluate any campaign against that target within a week of data.

Ongoing vs Cohort ROAS Comparison

Why These Distinctions Matter

This is where most Apple Ads advertisers get into trouble. If your business model means that 100% ROAS takes 3-4 months to achieve (which is completely normal for subscription apps), then your Ongoing ROAS for active campaigns will never show 100% right away. That is expected behavior, not a problem.

I regularly see solo developers and indie makers on X (Twitter) complaining about low ROAS on their active campaigns. They see 30-40% ROAS for the current month and panic. But if they checked their campaigns from 2-3 months ago, they would find those months already paid back through weekly and monthly subscription renewals.

The pattern is always the same: the current month looks bad because users have not had time to renew. The previous months look great because renewals have been accumulating. If you only ever look at the current month's Ongoing ROAS, you will always feel like you are losing money - even when you are profitable.

The Practical Implication

Use Cohort ROAS (D7 or D30) as your primary optimization metric. Set a target based on your unit economics model - for example, "D7 ROAS must be at least 25% for the campaign to stay active." Then use Ongoing ROAS on a monthly or quarterly basis to validate that your model is correct and campaigns are actually reaching 100% within your expected payback window.

The Optimization Philosophy

This is the most important section of this guide. Everything above is mechanics. What follows is the framework for making decisions with these metrics.

Three core principles should guide every optimization decision you make in Apple Ads:

Principle 1: Only Financial Metrics Determine Success

The only metric that determines whether a campaign is successful is ROAS. Not tap-through rate. Not cost per install. Not conversion rate. ROAS. Everything else is a diagnostic tool that helps you understand why ROAS is good or bad, but it does not define success or failure on its own.

Principle 2: Every Business Is Different

There is no universal "good" ROAS target, payback window, or cost per acquisition. Different businesses have different conversion rates, margins, subscription models, and goals. An app with 18-month payback can be just as healthy as one with 6-month payback - if the economics support it. Do not copy someone else's benchmarks blindly. Build your own model based on your own numbers.

Principle 3: Only Statistically Significant Data

Never make optimization decisions based on small sample sizes. A keyword with 5 installs and 0 conversions is not "bad" - it is inconclusive. Wait until you have enough data to be confident in the pattern before pausing, scaling, or adjusting bids. The threshold depends on your app's conversion rates, but as a general rule, you need at least 30-50 conversions per segment to draw meaningful conclusions.

Why Only Financial Metrics Determine Success

This point deserves its own section because it is counterintuitive and goes against what most Apple Ads tutorials teach.

Many advertisers optimize for proxy metrics: low CPI (cost per install), high TTR (tap-through rate), or low cost per trial. The assumption is that cheaper installs and more trials automatically lead to better ROAS. This is often wrong.

The reason is simple: different keywords, geos, and creative variants attract different types of users. A keyword that delivers cheap installs might attract users who never convert to paid. A geo with low CPIs might have users who cancel after the first renewal. A screenshot variant with amazing tap-through rates might set the wrong expectations, leading to quick uninstalls.

The chain of assumptions breaks at every step:

• Low CPI does not equal low trial cost
• Low trial cost does not equal low cost per paying user
• Low cost per paying user does not equal high ROAS

Each step in the funnel has its own conversion rate, and those rates vary dramatically by acquisition source. The only way to know if a campaign works is to measure the financial outcome directly.

Real-World ROAS Examples

Let me walk through three real scenarios that illustrate why surface metrics deceive and only ROAS tells the truth.

Example 1: Great Metrics, Terrible ROAS

A discovery campaign targeting broad keywords in Tier 1 geos. The numbers look amazing on the surface: high tap-through rate, low cost per install ($1.20), and cheap trial starts ($3.50 per trial). By every proxy metric, this campaign is a winner.

But when you check the financial metrics, the ROAS is 15% at D30. The users are starting trials and then canceling before the first payment. The cheap installs are coming from users who are casually browsing, not actively looking for a solution. They try the app, decide it is not for them, and leave. Great top-of-funnel metrics, terrible financial outcome.

Example 2: Awful Trial Cost, Great ROAS

An Exact Match campaign targeting high-intent keywords in specific European geos. The cost per trial is $12 - nearly 4x the first example. By the proxy metric standard, this campaign should be paused immediately.

But the D30 ROAS is 85%, and the D90 ROAS projects to 140%. Why? Because users in these geos overwhelmingly choose annual subscriptions. One annual purchase at $49.99 covers the entire acquisition cost and then some. The users who do convert are high-quality, high-LTV customers. Expensive trials, but excellent financial return.

Example 3: Low Short-Term ROAS, Strong Cumulative

A scaled campaign across multiple geos with a D7 ROAS of only 18% - well below most advertisers' comfort zone. At first glance, this looks like a money pit.

But this app has strong trial-to-weekly conversion rates and excellent weekly retention. Users who subscribe tend to stick around. The D30 ROAS hits 55%, and by D90 the cumulative ROAS crosses 110%. The low D7 number reflects the subscription model (weekly, with a 3-day trial), not poor performance. Patience and understanding your payback curve is everything.

Budget, Data, and Decision-Making

Here is an uncomfortable truth: you probably will not have enough data. Most Apple Ads accounts do not generate the volume of conversions needed for perfect statistical confidence across every keyword, geo, and campaign. Luck plays a role, especially at smaller scales.

Your budget size fundamentally changes how you should optimize:

Small Budget ($5K - $10K/month)

With a small budget, you cannot afford to wait for statistical significance on every decision. You need to be more aggressive with your rules:

• Set strict ROAS thresholds and enforce them within shorter windows (D7)
• Pause underperformers quickly - you cannot afford extended testing periods
• Focus on fewer keywords and geos - concentrate spend to reach meaningful data volumes faster
• Rely more on proxy metrics early - when you have 10 installs, CPI and trial rates are all you have. Use them, but verify with financial data as soon as possible
• Accept that some decisions will be wrong - at low volume, randomness plays a bigger role

Large Budget ($50K+/month)

With a larger budget, you can afford patience and precision:

• Wait for statistical significance before making changes - at least 50+ conversions per segment
• Focus primarily on financial metrics - you have enough data to rely on ROAS directly
• Test more aggressively - allocate 10-20% of budget to pure experimentation
• Use longer evaluation windows - D30 ROAS instead of D7 when your model supports it
• Optimize incrementally - small bid adjustments rather than dramatic pauses

The key insight is that budget constraints determine your optimization style, not just your scale. A $5K/month advertiser and a $50K/month advertiser should not use the same playbook, even if they are promoting the same app.

Final Thoughts

ROAS is not a single number you check once and move on. It is a dynamic metric that behaves differently depending on how you measure it - and understanding that difference is fundamental to running profitable Apple Ads campaigns.

Use Cohort ROAS (D7, D30) for daily and weekly optimization decisions. Use Ongoing ROAS for monthly and quarterly validation. Never confuse the two, and never panic about a current month's Ongoing ROAS being low - that is how the math works for subscription apps.

Build your optimization philosophy around three principles: only financial metrics define success, every business has unique benchmarks, and only statistically significant data should drive decisions. Proxy metrics like CPI and TTR are diagnostic tools, not decision-making criteria.

And above all, remember that ROAS optimization is a patience game. The advertisers who win are not the ones with the best Day 1 numbers - they are the ones who understand their payback curves, trust their models, and make disciplined decisions based on the right data at the right time.

For a comprehensive overview of Apple Search Ads strategy, check out the complete Apple Search Ads guide. If you need help setting up proper attribution and building a ROAS optimization framework for your app, get in touch.