App growth marketing is a massive global industry, projected to reach $561 billion by 2032.1 At the heart of this ecosystem is a fundamentally flawed measurement standard: last click-attribution. By awarding 100% of the credit to the final interaction, a few key measurement providers have built a system that incentivizes "Attribution Theft" and perpetuates the industry's most pervasive fraud problem.
The problem with last-click attribution
Last-click attribution assigns full credit for a conversion to the final touchpoint before an app install. While this approach was once considered state-of-the-art, modern data science has exposed its limitations. It’s both inaccurate and, essentially, attribution theft. Channels like TV often create intent, only for the credit to be stolen by a later click, such as a mobile ad. This misattribution distorts return on investment (ROI) calculations and fuels inefficiencies across app growth marketing.
Worse, last-click attribution has paved the way for attribution fraud. By oversimplifying the customer journey and awarding 100% credit to the final interaction, it creates exploitable loopholes. Fraudsters can fake the "last click" to steal credit for organic installs or those driven by other legitimate channels, leading to inflated performance metrics and wasted billions in marketing budgets. This fraud not only drains resources but also erodes trust in the entire ecosystem.
The problem doesn’t stop at clicks. View-through attribution (VTA), which awards credit to impressions even when no click occurs, introduces its own distortions. Without rigorous incrementality testing, VTA can blur the line between genuine influence and incidental exposure, allowing channels to claim credit for conversions that may have happened anyway.
How last-click attribution enables fraud
Last-click's simplicity makes it vulnerable to fraudsters because they only need to manipulate the timing of a single touchpoint to claim full credit.
Click spamming (or click flooding)
Fraudsters use bots or malicious apps to generate massive volumes of fake clicks. These clicks are timed to occur frequently, increasing the likelihood that one will be registered as the "last click" before an organic install. This floods MMPs with bogus data, allowing fraudsters to siphon payouts without delivering real value.
Click injection
This involves apps monitoring for new installs via "install broadcasts." Once an install begins, the fraudulent app injects a fake click just before completion, hijacking the attribution. This steals credit from the true source—often organic or another ad network—and tricks MMPs into paying the wrong party.
SDK spoofing
Fraudsters tamper with Software Development Kits (SDKs) in apps to fabricate install reports or clicks. By mimicking legitimate traffic, they create false attributions, inflating metrics like install counts and engagement rates.
Impression stacking
This is similar to click spamming, but for VTA. Fraudsters trigger fake impressions in the background (without showing ads to users). If a user installs the app later, the fraudster claims credit as the "last view."
These methods exploit the MMPs' reliance on last-click, as the model doesn't verify the authenticity of touchpoints deeply enough.
How black boxes exacerbate the problem
Even if we ignore the mechanics of fraud, we cannot ignore the lack of visibility. Most major ad networks operate as "Black Boxes." Their attribution models and decision-making logic are completely opaque, with zero transparency into how they claim credit for performance or how they "match" a user to an ad. When the decisioning logic is hidden, you are forced to "trust" the person taking your money or keep score of your investment ROI.
Why opaque logic leads to legal firestorms
This lack of transparency leads to legal and regulatory attention. Because these methods cannot be inspected, platforms become a lightning rod for allegations of systemic fraud. For example, in 2024 and 2025, several Securities Fraud Class Action Lawsuits were filed, alleging that revenue growth relied on deceptive practices like click spoofing.
Short-seller reports from Muddy Waters and Fuzzy Panda further alleged that MMP platform terms were violated to manufacture performance results. These lawsuits are the natural result of a black box culture: when you don't show the math, the market eventually assumes the worst.
The marketer’s manifesto: Taking back control
To move from a "marketing cesspool" to a sound scientific engine, marketers must stop being passive observers. Here are three things they should do instead:
Demand radical transparency
Don't accept "proprietary algorithms" as a valid answer. Mandate that your MMP provide an audit of exactly how they assign credit in "non-click" environments like TV.
Execute "go dark" experiments
Periodically turn off specific, high-spend, non-transparent channels for 48-72 hours while keeping your intent creation channels, like TV, running. If total installs don’t drop, you’ve caught an attribution thief.
Tether AI to purchase data
Hard-code your optimization tools to optimize for Post-Install Revenue, not just "installs" or "clicks."
The CFO cheat sheet: 5 questions to audit your MMP
Causality vs. correlation
Can you provide a report showing the "incremental lift" of our spend? How are view-through conversions validated for incrementality?
Audit trail
Is there a clear log showing the decisioning path for every "last-click" credit awarded to a non-transparent partner?
The "creation" credit
How are we weighting the "Intent Creation" of TV/Brand ads against the "Intent Harvesting" of Search/Social?
Anomaly detection
What is our current Click-to-Install Time (CTIT) distribution? (Spikes indicate Click Injection).
Unit economics
Does our reported ROAS from the MMP reconcile with our actual bank-cleared revenue
The era of unverified trust is over
The era of trusting dashboards without verification is over. We cannot hand marketing dollars to opaque systems with misaligned incentives and blindly accept the outcomes they report.
By demanding transparency and using the scientific method to test for true incrementality, you can both save your marketing budget and rewrite the power structure of the industry.
The future belongs to the marketers who choose science over voodoo.
Inside Performance Advertising with Jason Fairchild delivers unfiltered insights, strategic perspective, and hard truths from inside the evolving world of adtech—cutting through the noise to focus on what really drives outcomes. Subscribe here.
Sources:
- SNS Insider, “In-app Advertising Market to Reach USD 561.24 Billion by 2032,” March 27, 2025
