If you doubled your paid install budget last quarter and total installs grew by only 30%, where did the other 70% of expected growth go? For many mobile app teams, the answer is uncomfortable: a significant portion of paid installs would have happened organically. You did not acquire new users. You paid for users you were already getting for free.
This is paid-organic cannibalisation, and it is one of the most underdiagnosed problems in mobile user acquisition. Attribution dashboards make it invisible because they report what happened (a user clicked an ad and installed) not what would have happened without the ad (the same user would have found you through search, a friend's recommendation, or an app store browse).
The financial impact is real. An app spending Rs20 lakh monthly on paid UA with 30% cannibalisation is effectively wasting Rs6 lakh per month on users it would have acquired for free. At scale, this is the difference between healthy unit economics and a growth machine that looks efficient on dashboards but leaks money in practice.

What Is Paid-Organic Cannibalisation (And Why It Happens More Than You Think)?

Paid-organic cannibalisation occurs when paid advertising claims credit for conversions that would have happened without the ad. The user was already on the path to installing your app. Your ad intercepted them at some point on that path, and your attribution system gave the ad full credit.
This is not fraud. The ad was seen. The click was real. The install happened after the click. By every attribution rule, the paid campaign gets credit. But the causal reality is different: the user would have installed anyway.
Why it happens frequently:
Branded search campaigns are the most common culprit. When you bid on your own brand name in Google UAC, you capture users who are already searching for your app by name. These users have high intent and would almost certainly have found your organic app store listing. Your branded campaign delivers excellent reported ROAS because it captures high-intent users, but the incremental value of that spend is near zero.
Meta broad targeting in mature markets overlaps with organic audiences. When your app has strong brand awareness in a specific market, Meta's broad targeting will naturally show ads to people who already know your brand. Some of these users would have found you through word of mouth, app store browse, or social sharing.
Retargeting website visitors who were already converting. If you retarget users who visited your website and were already moving toward an install, you are paying to accelerate a conversion that was already in progress.
Last-click attribution masks the problem entirely. In a last-click model, if a user was going to install organically but happened to see and click a paid ad first, the paid channel gets 100% credit. The organic channel gets nothing. Your dashboard shows paid growth. In reality, organic growth was suppressed. Understanding the limitations of different attribution models is essential for spotting cannibalisation patterns.

Five Signs Your Paid Campaigns Are Cannibalising Organic Growth

Before running formal tests, these directional signals can indicate whether cannibalisation is a problem worth investigating.
Sign 1: Organic installs drop when paid spend increases. Plot your weekly paid installs and organic installs on the same chart. If organic installs decline during weeks when paid spend scales up, especially in the same geographies, you likely have overlap. A healthy growth engine shows both paid and organic growing together, or at least organic staying flat as paid scales.
Sign 2: Total installs do not scale proportionally with paid spend increases. If you increase paid spend by 50% and total installs (paid + organic) grow by only 20%, the gap suggests paid is capturing users who would have been organic. True incremental paid growth would produce closer to proportional total install growth.
Sign 3: Branded search campaigns show unusually high ROAS. If your Google branded search campaigns report 5x+ ROAS while generic campaigns show 1.5-2x, the gap is partly explained by branded search capturing high-intent organic users. These users are valuable, but the spend to capture them may not be.
Sign 4: Paid and organic user quality nearly match, and pausing spend triggers minimal total install decline. If D7 retention, revenue per user, and LTV are nearly identical between your paid and organic cohorts, paid campaigns are reaching the same user profile as organic. If Meta reportedly drives 35% of your installs but pausing it for a week only reduces total installs by 15%, the gap represents overlap between paid and organic installs.

How to Measure Cannibalisation: Three Practical Methods

Directional signals indicate the problem. These methods quantify it.
Method 1: The pause test. Select a specific geographic region (a state or group of cities). Pause paid spend on one channel in that region for 2-4 weeks. Keep everything else unchanged. Compare total installs in the test region (paid paused) against a matched control region (paid active). If the test region's total installs drop by significantly less than the paused channel's attributed share, the gap is your cannibalisation rate.
Example: Meta reportedly drives 300 installs per week in Pune. You pause Meta in Pune for 3 weeks. Total installs in Pune drop from 800 to 650 per week (a decrease of 150, not 300). Estimated cannibalisation: 50% of Meta-attributed installs in Pune were cannibalised organic installs.
Method 2: Correlation analysis. Pull 6-12 months of weekly data with columns for paid installs, organic installs, and total installs. Plot paid versus organic on a scatter chart. If the correlation is negative (organic drops as paid rises) or if total installs plateau as paid scales, you have cannibalisation. This method is less precise than pause tests but requires no changes to active campaigns.
Method 3: Holdout regions. Identify pairs of similar cities (matched by population, app penetration, and user demographics). Run paid campaigns in one city of each pair but not the other. Compare total installs in the paid city versus the organic-only city. The difference in total installs is your true incremental contribution from paid spend.
Setting up each test: Run for a minimum of 2 weeks, ideally 4 weeks, to generate reliable signal and account for day-of-week patterns.
Minimum thresholds for reliable results: You need at least 200 installs per week per region to detect a 20% cannibalisation rate with reasonable confidence. Below that, random variance overwhelms the signal.

Calculating Your True Incremental Cost Per Install

Once you have estimated your cannibalisation rate, translating it into financial terms changes how you evaluate every channel.
The formula: Incremental CPI = Total paid spend / (Total paid installs - Estimated cannibalised installs)
Example: You spent Rs5 lakh on Meta last month. Meta's attribution reports 5,000 installs. Reported CPI: Rs100. Your pause test estimated 40% cannibalisation. True incremental installs: 3,000 (5,000 x 0.60). Incremental CPI: Rs167 (Rs5,00,000 / 3,000).
This is the number that changes the conversation in your Monday budget meeting. Rs100 CPI feels efficient. Rs167 CPI raises questions. Both describe the same campaign.
Your true cost of acquiring a genuinely new user through Meta is Rs167, not Rs100. This changes whether the channel delivers acceptable marketing ROI.
How this changes channel-level ROAS calculations. If your reported Meta ROAS is 3x but 40% of attributed installs are cannibalised, your true incremental ROAS is closer to 1.8x (3x x 0.60). Whether 1.8x ROAS justifies the spend depends on your unit economics, but it is a fundamentally different number from the 3x your dashboard shows.
The insight is not "stop spending." It is "know the true cost." An incremental CPI of Rs167 might still be profitable if your LTV supports it. But making that assessment with an inflated Rs100 CPI leads to over-allocating budget to a channel that is less efficient than it appears.

Reducing cannibalisation Without Killing Paid Growth

The goal is not to eliminate paid spend. It is to maximise the proportion of paid installs that are genuinely incremental.
Tighten targeting to exclude high-intent organic audiences. If users who are already searching for your brand or visiting your website would convert organically, excluding them from paid campaigns reduces overlap. Create exclusion audiences from recent website visitors, existing app users, and users who have engaged with organic content.
Reduce branded search spend. Test lowering your brand keyword bids on Google by 50%. If total installs from branded search drop by only 10-15%, you were overpaying for organic traffic. Many apps find they can cut branded search spend significantly without meaningful impact on total install volume.
Shift budget toward prospecting. Focus paid spend on audiences with zero organic intent signals: broad targeting in new geographies, lookalike audiences based on high-LTV users, interest-based targeting in categories adjacent to your app. These users are less likely to find you organically, making every paid install more incremental.
Use exclusion lists aggressively. Suppress existing app users from install campaigns (your MMP audience data enables this). Suppress users who have visited your website in the last 7 days. Suppress users who have engaged with your organic social content. Each exclusion reduces the chance that paid spend captures an organic-bound user.
Channel-specific tactics: On Meta, test narrow lookalikes (1-2%) against broad targeting. Narrow lookalikes are more incremental when your brand awareness is high. On Google, separate generic campaigns from branded campaigns and evaluate them independently. The budget allocation framework should incorporate incrementality-adjusted ROAS, not just raw attribution-reported ROAS.
Consider feeding incrementality findings into your budget reallocation framework to shift spend from high-cannibalisation channels toward channels with higher incremental contribution.

Building Ongoing cannibalisation Monitoring into Your Reporting

cannibalisation is not a one-time measurement. It changes as your brand awareness grows and as you scale spend.
Track the organic/paid ratio as a weekly health metric. A healthy, incrementally growing app maintains a stable or improving organic share of total installs. If organic share declines as paid spend increases, cannibalisation is likely growing.
Set alert thresholds. If organic installs drop below a defined floor (e.g., 30% of total installs), trigger an investigation before efficiency declines.
Run quarterly pause tests. Over a year, build a data-backed picture of each channel's true incremental contribution. Report incremental metrics alongside attribution to leadership: "Meta reported ROAS: 3x. Meta incremental ROAS (adjusted for 35% cannibalisation): 1.95x." This prevents inflated numbers from driving scaling decisions.
Linkrunner's channel-level attribution with organic install tracking provides baseline data for pause tests and correlation analysis. When you see paid and organic trends by channel and geography in one dashboard, identifying cannibalisation patterns becomes routine.

Frequently Asked Questions

How much of my paid spend is typically cannibalised?
It varies by brand awareness and channel mix. Strong-brand apps typically see 20-40% cannibalisation on branded search and 10-25% on broad social. Early-stage apps with low brand awareness see much less.
Are branded Google search campaigns the biggest source of cannibalisation?
Yes. Branded campaigns target users who already know your brand, the highest-probability organic converters. They show the best ROAS because they capture intent-rich users, not because spend is incremental.
How long should I pause paid spend to measure accurately?
A minimum of 2 weeks, ideally 4 weeks, to generate reliable signal and account for day-of-week patterns.
Can I reduce cannibalisation without cutting total installs?
Yes. Exclusion lists, shifting to prospecting, and reducing branded search typically maintain or grow total installs while improving incremental share.
Does cannibalisation affect SKAN-reported iOS installs?
cannibalisation dynamics apply to SKAN as well. Geo-lift tests work identically on iOS because they measure aggregate volume, not user-level attribution.

Knowing Your True Numbers

You started this post with a hard question: if you doubled your paid install budget and total installs grew by only 30%, where did the other 70% go?
Now you know. Some of it went to users you would have acquired for free. Your dashboard showed paid growth. In reality, organic growth was suppressed. The incremental value of that spend was lower than attribution suggested.
This is not a failure of paid campaigns. It is a failure of measurement. The most efficient growth engines are not the ones that scale paid spend fastest. They are the ones that know the true incremental value of every channel and allocate budget accordingly.
Start with a single pause test on your highest-spend channel in one geography. Measure the delta. Calculate your incremental CPI. Then decide whether your budget allocation still makes sense with the real numbers.
If you need clean organic and paid install data by channel and geography to run these analyses, Linkrunner provides both in a single dashboard with unrestricted exports. Request a demo to see how the paid-organic split looks for your traffic mix.
The uncomfortable truth is the only path to efficient growth.