TL;DR

• Ad fraud doesn’t just waste budget—it corrupts your ML models, skews KPIs, and rewards fraudulent partners
• Evaluating fraud detection data reveals patterns you can use to sharpen strategy and reclaim wasted spend
• The real power: confidently optimize future spend by turning fraud intelligence into a growth enabler

What most marketers miss about fraud

We all know ad fraud burns money. That’s not news.

But what catches even sophisticated marketers off guard is this: you’re making million-dollar decisions based on distorted data, and that’s what really destroys performance.

Think of fraud prevention like having a water filtration system. Sure, it’s great that you’re blocking contaminants. But what if you could analyze what’s in the filter (the weird chemicals, the unexpected particles) and use that intel to optimize your entire water supply chain? That’s the difference between blocking fraud and learning from it.

In this blog, we’ll show you how evaluating fraud data transforms it from a cost center into a growth engine. You’ll learn how to decode fraud patterns to sharpen your targeting, recalibrate your KPIs to reality, and turn prevention into a competitive advantage. Because once you stop just blocking fraud and start learning from it, you unlock insights that make every dollar you spend work harder.

The hidden damage: when your “winners” are actually losers

Ad fraud continues to drain roughly 12% of digital ad spend globally, with losses projected to hit $172 billion by 2028.

But the most threatening part isn’t actually the money that’s lost. It’s what happens to the money you keep spending.

When fraudulent installs or clicks slip into your dataset, they don’t just inflate your numbers. They corrupt your entire feedback loop. Your machine learning algorithms? They’re learning from false data. Your budget pacing? Optimizing toward fiction. Your “best-performing” channel? Might be your biggest fraud risk.

Here’s a real case that makes this painfully clear:

A gaming advertiser discovered a quarter of their traffic was invalid. No big surprise there. But here’s the real problem: 80% of installs were misattributed. Their optimization engine was literally rewarding the partners inflating fake conversions. It took months to rebuild trust in their data and recalibrate their bidding logic.

This isn’t just a marketing leak. It’s a data integrity breach that infects every decision downstream.

But wait… having fraud protection isn’t enough?

Many advertisers treat fraud detection like setting up antivirus software. Install it, enable the filters, maybe check the dashboard once a quarter, and call it done.

The problem: detection without evaluation is like having a security camera that never gets reviewed. You’re blocking threats, but you’re missing the intelligence goldmine sitting right in front of you.

Every fraudulent install, click, or impression leaves a fingerprint. Timestamps, device clusters, velocity patterns, behavioral mismatches. These aren’t just anomalies to filter out. They’re signals telling you exactly where your defenses are weak.

When you evaluate fraud data, you start asking better questions:

• Detection speed: How fast did we catch this? Pre-attribution or after we’d already paid for it?
• Pattern recognition: Are the same fraud signatures showing up from specific sub-publishers or geos?
• Attribution hijacking: Are legitimate sources getting credit stolen by injected traffic?

Most marketers never look this deep. But this is where the real competitive advantage lives.

Turning fraud evaluation into your growth engine

Let’s flip the script. Instead of thinking about fraud data as a defensive necessity, treat it like performance intelligence. What does that look like in practice?

Optimize spend efficiency, not just spend hygiene

Fraud data doesn’t just show you where budgets get burned. It reveals where your real incremental lift comes from.

Before, you might have thought “cut the bad sources and move on.” Now you’re thinking in terms of budget recapture rate: how much of that reclaimed spend can you redirect into proven, fraud-light channels? Advertisers who continuously reallocate reclaimed spend into fraud-light channels can recover significant portions of their budget and redirect it toward real incremental growth

Recalibrate your KPIs to reality

Your ROI, ROAS, and CPA metrics only mean something if they reflect authentic user behavior. When you strip out fraud, you’re not just improving accuracy. You’re recalibrating every performance model that depends on those numbers.

Here’s the math: if 20% of your conversions are fraudulent, you’re paying 25% more per real customer than you think. Say you spend $10,000 for 100 conversions at $100 CPA. But 20 are fake, so you actually paid for 80 customers at $125 each. Evaluating fraud data early reveals the true cost and keeps your bidding aligned with reality.

Shorten your feedback loop

Fraud insights have a shelf life. The longer they sit unreviewed, the less value they provide.

By integrating real-time fraud evaluation into your analytics pipeline (through APIs or direct data access), you can act before optimization drifts off course. This continuous feedback loop means your marketing teams adapt in days instead of quarters.

Build accountability into partnerships

Fraud data shouldn’t live in a silo. When you share fraud evaluation metrics with networks, DSPs, and affiliates, you create alignment around quality standards. It also signals that you’re measuring performance at a granular level. That’s a subtle but powerful deterrent.

Over time, this shifts partner behavior toward transparency and sustained quality. Think of it as raising the bar for everyone in your ecosystem.

The detection paradox (this one trips people up)

Here’s something that may sound counterintuitive: when your fraud detection improves, your fraud metrics often spike first.

It’s easy to panic and think “oh no, fraud is getting worse!” But actually, you’re just seeing what was always there. You’ve upgraded from a flashlight to floodlights.

The goal isn’t zero detected fraud. That’s impossible. The real goal is increasing detection coverage and reducing detection latency. Catching more fraud, faster, before it contaminates your optimization.

By evaluating your fraud data over time (comparing detection timelines, false-positive rates, and attack vector evolution), you can actually quantify how well your protection stack is learning. This is what transforms fraud prevention from a checkbox into a competitive differentiator.

Making fraud evaluation part of your routine

Building a culture around fraud evaluation means weaving it into your operational rhythm:

Weekly: Review emerging fraud patterns. New publisher IDs, device farm signatures, click velocity spikes.

Monthly: Cross-reference fraud trends against campaign performance to validate your KPIs.

Quarterly: Audit detection efficacy. Latency, false-positive ratios, rule overlap. Refine your thresholds. This is also when you align with partners. If fraud rates are climbing from a specific source, it’s time to renegotiate agreed-upon performance standards. Fraud metrics give you leverage to enforce quality and hold partners accountable.

The critical move? Tie these reviews directly to optimization decisions. If a source shows rising fraud rates despite volume growth, it’s not “performing.” It’s stealing.

And if your team is still treating fraud as an external vendor problem instead of a data problem, that’s the mindset shift to tackle first.

Key takeaways

When you make fraud evaluation a core part of your analytics workflow, things change:

• Fraud evaluation is a data integrity function: It protects the accuracy of your KPIs, ML models, and strategic decisions, not just your budget.
• Continuous pattern review shortens your feedback loop and prevents the slow drift that kills optimization performance.
• Fraud insights create leverage with partners, enforcing transparency and aligning incentives around real, quality performance.
• The goal isn’t eliminating detected fraud. It’s increasing detection coverage and slashing detection latency so you get earlier, more actionable intelligence.
• Fraud optimization enables growth: Confident fraud evaluation lets you take calculated risks, test aggressively, and scale faster because you know how to manage risk intelligently.

Teams who regularly evaluate fraud data outperform those who only block it, because they treat fraud as strategic intelligence, not just a defensive safety net.

The post Why fraud data is your secret growth engine (Not just a filter) appeared first on AppsFlyer.

TL;DR

• In finance, customers enter the app at moments of commitment, not exploration
• Engagement teams own many of the moments that lead customers into the app
• When this transition into the app breaks, customers often don’t retry – and silent failures damage trust long before teams notice
• These failures are often invisible in analytics and misattributed to “low engagement”
• Treating these transitions into the app as intent continuation and a critical part of the user journey – not redirection – changes outcomes
• Teams that ensure their journey into mobile apps is continuous  boost engagement , measure granularly for better optimization, and are able to better predict outcomes

App entry is an engagement problem, not a technical one 

If you work on CRM, lifecycle, or engagement in a finance or fintech company, you’re responsible for some of the most important moments in the customer journey: account creation and registration, first deposit or first transaction, identity verification and document submission, reviewing statements or required account actions, resuming an application or setup after interruption.

What’s the common denominator? In all of these moments, the customer is trying to continue and complete an action inside the app. They’re not browsing, or experimenting but rather are acting with clear intent.

This continuation is handled by deep linking technology – the infrastructure that ensures that the customer lands in the right place in the app with the right state preserved.

When app entry works, engagement feels effortless. When it doesn’t, customers don’t complain – they disappear.

How can we ensure we deliver effortless engagement? Here are five app entry moments engagement teams in finance can’t afford to break.

1. Account activation and first secure entry

The moment
A customer taps a message to activate their account, complete onboarding, or start using the app for the first time.

Why it matters
This is where intent turns into habit. If activation breaks, many customers never form a relationship with the app at all.

What often goes wrong

• The app opens, but the customer lands on a generic home screen
  
• Context from the activation message is lost because deep links don’t reliably carry context into the app
  
• Login interrupts the flow and forces a restart

From the outside, it looks like “customers didn’t activate.” In reality, the journey never continued even if the customer clearly intended to complete the setup.

How to fix it

Treat activation as a continuation of intent. Use deep linking to ensure the app opens directly in the activation or onboarding step the customer was prompted to complete, even if login or verification is required along the way.

2. Secure messages and required account actions

The moment
Customers tap links in emails or SMS messages about statements, alerts, verification steps, or policy updates.

Why it matters
These messages aren’t optional. Customers expect to land on the exact relevant screen, with the right context, so they can immediately take action. .

What often goes wrong

• Messages open the app, but not the specific content referenced
  
• Authentication breaks the flow
  
• Measurement shows clicks, but no meaningful in-app action

Engagement teams see high open and click rates, but downstream completion quietly drops.

How to fix it
Deeplink customers from secure messages into the specific in-app screen referenced in the message. Handle authentication as part of the flow so customers can act immediately without re-navigating or searching.

3. Re-engagement after inactivity

The moment
A dormant customer taps a message to check their balance, resume an application, or respond to an offer.

Why it matters
This is often a last chance. Customers who drop here rarely retry.

What often goes wrong

• Customers with the app installed are sent to the app store
• Customers who reinstall or open the app after inactivity aren’t routed correctly because deferred deep linking isn’t handled consistently
• Context from the message isn’t preserved
• The app treats the customer as new instead of returning
  

Deferred deep linking is what allows customers who don’t yet have the app (or  want to reinstall it) to continue where they left off after installation.

While campaign performance may appear to be good, engagement doesn’t recover. In other words, the point of  failure sits between the click and the app.

How to fix it
Accurately detect install and user state and prioritize opening the app with preserved context. Returning customers should resume what they were invited to do, not restart from the top.

4. Offline or physical moments to complete actions in the app

The moment
Customers enter the app from the physical world: branch signage, statements, direct mail, ATM screens, or packaging.

Why it matters
These moments represent explicit intent. The customer chose to act, often in a high-trust context.

What often goes wrong

• The app doesn’t open even when installed
  
• The customer lands in the wrong place
  
• Links can’t be updated once deployed

These flows rely heavily on deep links that live outside the app and must continue to work long after they’re deployed

Because these campaigns are hard to iterate, teams often stop scaling them altogether after a few bad experiences.

How to fix it
Design offline entry points with the same rigor as digital ones. Ensure links are updateable, governed, and consistently route customers into the correct in-app experience the first time.

5. Login and verification interruptions during sensitive actions

The moment
Customers are interrupted by login or verification and need to resume an action, whether completing an application, initiating a transfer, or going through a setup flow.

Why it matters
Forcing customers to start over after authentication breaks trust and completion.

What often goes wrong

• Context is lost after login
  
• Customers are dropped at a default screen
  
• Teams assume “customers abandoned”

In reality, the journey was abandoned for them.

How to fix it
Decouple authentication from intent resolution. After login or verification, return customers to the exact step they were attempting to complete instead of defaulting to a home screen.

The hidden pattern: broken app entry looks like low engagement

Across all five moments, the pattern is the same:

• Engagement teams do everything right
  
• Messages are timely and relevant
  
• Customers click with intent

But the transition into the app fails and it often does so silently.

Because these failures happen after the click and before meaningful in-app events, analytics rarely capture them clearly. Teams optimize what they can see, not what’s actually broken.

Questions CRM and lifecycle teams should be asking

If you own lifecycle or engagement performance, these are worth checking:

• When customers enter the app, do they land with the same context they had before?
  
• How often do customers with the app installed still end up in the app store?
  
• Can we see what happens between the click and the first in-app action?
  
• Are resume flows after login preserving intent or forcing restarts?
  
• Who owns fixing app entry when performance looks fine but engagement drops?
  
• Do customers who click actually complete the action they were prompted to take?

The real lesson: These are moments of commitment, not “traffic”

Most teams think of app entry as a link, a banner, or a button, but customers don’t.

They’re asking to continue a specific action inside the app, without friction or reset. That requires consistent routing, accurate state detection, and context preservation.

When app entry is handled consistently:

• Engagement increases
  
• Measurement becomes more trustworthy
  
• Completion becomes predictable

While AppsFlyer is widely known for user acquisition measurement and attribution, it also offers a dedicated Deep Linking Suite for owned media which does not require a paid media attribution package of any kind. 

AppsFlyer works with thousands of CRM, lifecycle, growth, and product teams to power reliable X-to-app journeys across web, email, QR, social, SMS, referrals, and deferred install flows, increasing engagement, conversion and overall performance across channels using its OneLink technology.

Key takeaways

• App entry failures often look like low engagement
• Finance customers don’t retry broken entry moments
• Context and state matter more than the channel
• These issues are hard to spot but expensive to ignore
• Treating app entry as intent continuation changes outcomes
• AppsFlyer’s Deep Linking Suite is available for owned media teams. 

The teams that get this right don’t just see better engagement numbers – they stop losing customers they already earned, at the moments those customers chose to act. If owned media journeys are a meaningful part of your lifecycle strategy, see how AppsFlyer’s Deep Linking Suite helps teams preserve intent and context across email, SMS, QR, and deferred install flows, without requiring a paid attribution package. Explore the Deep Linking Suite or talk to an expert about your app entry gaps.

The post 5 app entry moments engagement teams in finance can’t afford to break appeared first on AppsFlyer.

Brands that monetize with subscriptions are navigating two shifts that are reshaping the market faster than most playbooks can keep up. 

AI has moved from feature to infrastructure, shifting competitive advantage from budget size to data quality and the speed to act on it. Meanwhile, the geographic map is being redrawn: North America is maturing, Android is on the rise, and the next subscriber is increasingly coming from India, LATAM, and the Middle East.

The category picture is fragmenting, and the gaps are widening. Short Drama and OTT are surging, Gaming is pulling back spend, while GenAI is rethinking its monetization model. To help make sense of this changing landscape, this report analyzed 1.7 billion paid installs of 2,900 subscription apps in 13 categories, and $2.1 billion in UA spend.

What’s inside:

• Where Android and iOS UA spend is going, and which categories are winning
• How monetization models are diverging across Streaming, Short Drama, Gaming, GenAI, and more
• What the subscription funnel looks like by category: trial adoption, conversion rates, and install-to-paid benchmarks
• How subscription marketers are using AI today, and where the capability-usage gap is widest

The post The State of Subscriptions for Marketers – 2026 Edition appeared first on AppsFlyer.

TL;DR

• Marketers have always paid a “fragmentation tax” — bad signals scattered across platforms, channels, funnels, and tech stacks erode confidence and waste budget.
• AI doesn’t solve fragmentation; it amplifies it. Corrupted or incomplete signals fed into AI systems produce faster, more confident wrong decisions.
• CMOs are caught in a double bind: AI has made the marketing environment noisier and more complex, while leadership expects AI to have already solved the measurement problem.
• The fix isn’t more AI tools — it’s the foundation: governed signals, AI-ready data architecture, and mobile-grade measurement applied across all channels.
• Marketers who get the foundation right will turn AI into a compounding advantage; those who don’t will keep paying an ever-growing tax.

Introduction

In 1966, The Beatles had an earworm for the ages: “There’s one for you, nineteen for me.” George Harrison wrote Taxman in a fury: the idea that simply existing cost you something felt absurd, even offensive. Your street, your seat, your heat, your feet. All taxed.

When I talk to fellow CMOs today, that song keeps coming back. Because in 2026, marketers are paying a tax on everything too. Every platform, channel, tool silo, and, most expensively, every wrong decision made on fragmented data.

We’ve been living with this fragmentation tax for years. But something changed recently: the bill just got a lot bigger.

The reason is AI.

Fragmentation creates and amplifies bad data signals

At the core of every marketing decision is a signal: an impression, a click, a purchase, an identity match, a fraud flag. As AI removes abstraction layers, signals themselves become the product yet most are still bad at the source: self-reported, fraud-contaminated, incomplete, and restricted by walled gardens.

Fragmentation makes it worse. Bad signals flow through siloed systems, inconsistent taxonomies, and unreconciled identities and never reach the decision layer intact across four dimensions:

1. Across platforms: mobile app, web, CTV, PC, console, DOOH. Each with its own measurement logic, schema, and latency. The same customer behaves differently on each, and your tools rarely agree on who they are.
2. Across channels: owned and paid media operate on fundamentally different rules. Paid platforms self-report inflated numbers. Walled gardens restrict data. Traditional marketing clouds were never designed to reconcile journeys across both.
3. Across the funnel: brand, performance, and product teams measure separately. A single customer journey is often split across three systems, three teams, and three definitions of success.
4. Across the tech stack: measurement is disconnected from activation. First-party data can’t reach platforms. Platform data can’t enrich internal systems. There’s no unified foundation.

Each gap costs you time, confidence, and budget.

Why AI made the fragmentation tax soar

Here’s the uncomfortable truth: AI doesn’t fix bad signals or fragmentation. It amplifies both.

AI needs data, and when it’s fragmented, it doesn’t unify it, optimizing  on whatever signals are easiest to access. It doesn’t distinguish clean from corrupted signals, and amplifies whichever is loudest.

Garbage in, garbage out. On steroids.

The consequences are concrete:

Across platforms: AI can’t reconcile identity systems or attribution models. It treats the same customer as multiple users, double-counts conversions, and optimizes toward the fastest, not most accurate, signals.

Across channels: Paid media AI attributes everything to ads, while CRM AI attributes everything to email. Both optimize on incomplete data, with no arbiter.

Across the funnel: Separate AI systems for brand, performance, and product can end up working against each other.

At the same time, a growing share of decisions now happens before anything measurable: in brand perception, communities, and LLMs where influence is shaped by narrative, not clicks. By the time a user enters the measurable funnel, much of the decision is already made.

This isn’t theoretical. According to the IAB’s State of Data 2025, 62% of marketers cite data quality and fragmentation as a top barrier to AI success.

The double bind CMOs are in

Two forces are pulling in opposite directions.

On one hand, AI has made marketing noisier. Content is now infinite, but attention is not. The bottleneck has shifted from production to attention. With more noise, more bias, and autonomous recommendations flying in from every direction, confidence erodes fast.

The Harvard Business Review studied 200 employees over eight months and found AI didn’t reduce work but intensified it. HubSpot reports 73% of marketers have seen increased workload since adopting AI.

On the other hand, leadership expects more: more speed, more efficiency, more precision and tighter accountability on every dollar. As far as the board is concerned, AI has already solved your measurement problem. Now prove it.

The issue isn’t AI. It’s the foundation.

AI is genuinely transformative. The problem isn’t the model, it’s what the model is trained on.

Most marketing organizations are feeding AI systems data built on web-era assumptions, designed for a simpler world. That foundation was built for reporting, not autonomous decision-making.

To reduce the fragmentation tax in the AI era, we need to rebuild from three directions.

The right signals. At the core of every decision is a signal: an impression, a click, a purchase, an identity match. But not all signals are equal. A fraud-filtered, deduplicated conversion tied to a real identity is fundamentally more valuable than a platform-reported event with no cross-device validation. Measurement doesn’t just report — it validates and structures signals into something AI can act on. When signals are governed, AI compounds advantage. When they’re not, it compounds error.

The signals that matter most span mobile and web, connect creative performance to attribution, and tie purchase behavior to acquisition data. And because the best signals span the full funnel — from acquisition to remarketing — they enable connected customer experiences that convert over time, not just acquire.

AI-ready data architecture. “AI-ready” isn’t a label, it’s a set of properties: governed (traceable, validated, privacy-compliant), structured (consistent definitions across sources), contextual (complete journeys, not fragments), comprehensive (full coverage across platforms and channels), and consent-aware. Most marketing data today fails several of these. That’s not just a measurement problem, it’s an AI readiness problem.

Mobile-grade measurement, applied everywhere. Mobile set the highest bar for signal governance, out of necessity. It had to solve privacy constraints before the web did, fragmentation across iOS and Android, sophisticated fraud schemes , and identity resolution without cookies using several measurement frameworks. The takeaway isn’t to treat mobile as one channel among many. It’s to apply mobile-grade measurement as the standard across all channels: web, CTV, PC, console, and whatever comes next.

Why this is actually the golden age of marketing

It may not feel like it, but moments like this create resets and resets create new leaders.

Marketers are uniquely positioned for what comes next, not because they’re the fastest adopters or the loudest voices, but because they already understand the full equation: people and behavior, brand and narrative, data and experimentation, and how execution connects to revenue. No other function in the business owns all of that together.

AI hasn’t made that combination obsolete. It’s made it more valuable because someone still needs to know when the system is right, and when it’s compounding an error.

But knowing isn’t enough on its own. To actually get to the golden age, marketers need a foundation that brings measurement, data collaboration, and AI together in one place: trusted, privacy-first, and built for the complexity of today’s customer journeys. That’s what Modern Marketing is all about. 

Here comes the sun

Three years after Taxman, George Harrison wrote Here Comes the Sun — sitting in a friend’s garden after a long, brutal winter, guitar in hand, feeling the relief of something difficult finally lifting. “It’s alright,” he kept repeating. Not because the hard part hadn’t happened. Because it was over.

That’s where marketing is heading, if we do this right.

The fragmentation tax of the AI era is real. AI amplifies the problem when the foundation is weak. But the path forward isn’t to fear AI or stack more tools on top of broken infrastructure. It’s to fix what’s underneath: trusted measurement, AI-ready data, and real visibility across a fragmented reality where others are still flying blind.

Fix the foundation, and AI becomes an advantage – not a liability.

Then, finally, it’s alright.

The post Fix the Foundation: Marketing’s Signal Problem in the AI Era appeared first on AppsFlyer.

APAC fintech and finance marketing has entered a new phase. Growth has shifted from expansion to budget discipline, acquisition is more selective, lifecycle investment is accelerating, and monetization remains resilient in scale markets even as ad spend tightens.

The State of Finance for Marketers in APAC 2026 analyzes the data behind this transition, how finance marketers are reallocating spend, deepening monetization, and navigating uneven market maturity across the region.

From Indonesia’s paid expansion to Southeast Asia’s remarketing surge and fraud reduction in major markets, this report shows where performance is stabilizing, where it is concentrating, and where new efficiency signals are emerging.

Key findings:

• Finance installs fall for the first time, even as iOS hits a record 16% share
• A 193% surge in SEA remarketing spend as lifecycle investment gains ground on UA
• Indonesia and India as APAC’s dual-platform IAP anchors, commanding 66% of Android and 58% of iOS revenue across the region

The report covers:

• Installs, paid acquisition, remarketing, retention, revenue, and fraud benchmarks across SEA, India, Japan & Korea, and ANZ
• Data sample: 5.31B finance app installs, $5.7B in UA ad spend, and $29M in remarketing spend analyzed across APAC in 2025
• Best practices for finance marketers navigating capital discipline, platform mix, and lifecycle investment
• Expert insights from

The post The State of Finance for Marketers in APAC – 2026 Edition appeared first on AppsFlyer.

The State of Subscriptions for Marketers – 2026 Edition

01 KEY findings

+57%

YoY growth in Android paid installs, driven by Short Drama (+155%) and OTT & Live Streaming (+170%)

The Middle East grew +197% and Eastern Europe +95% as North America remained essentially flat. iOS organic declined 8% over the same period, making paid the primary growth lever on both platforms.

+24%

YoY growth in subscription app UA spend, with Android growing at 4x the rate of iOS (42% vs. 10%).

OTT & Live Streaming led the surge (Android +240%, iOS +120%), Short Drama slashed budget in North America (-40%) and significantly increased spend in the Indian Subcontinent (+423%) and LATAM (+77%). Gaming cut spend on both platforms.

90%+

Share of UA spend captured by the top 5 apps in Short Drama and OTT & Live Streaming

Both categories are structurally winner-take-most and holding. Health & Fitness is consolidating further, with the top 5 share jumping from 54% to 73%. Photo & Video is the exception: its top 5 share fell from 64% to 45% as AI-powered tools disrupt incumbents.

+17% Increase in ‘IAS Only’ share from 53% to 62% among OTT & Live Streaming apps to lead the shift toward pure subscription.

IAS & IAP fell from 47% to 38% over the same period. Short Drama moved the other way: IAS & IAA surged from near zero to 7.4% as the category tests ad-supported models to monetize install volume that isn’t converting to subscriptions. Gaming traded ad revenue for purchases, with IAS & IAP up 25% and IAS & IAA down 44%.

10:1 The ratio of retrieval to diagnostic queries among subscription marketers using AI chat tools

50% of queries cover channel performance and cost, which are the easiest questions to ask, not necessarily the most important. For every marketer asking why ROAS dropped, ten ask which campaign performed best. Business is the exception at 18% retention and cohort queries (4x the average).

19% Free trial-to-paid conversion rate in Gaming: the lowest in the dataset despite the highest trial adoption

Gaming pulls 12.2% of installs into trials but converts only 19% to paid. Education (42%) and Lifestyle (41%) convert at more than double the rate on lower volumes. Health & Fitness and Dating lead direct paid conversion without free trial at 7.1% and 6.5%.

02 introduction

AI and Android growth are reshaping the subscriptions economy

The subscription app market is growing, but not necessarily in the ways or places most marketers expected.

Two shifts stand out, converging to reshape what subscription app marketing looks like in 2026. The first is AI, but not as a feature, as infrastructure. Bidding algorithms now optimize across thousands of variables in real time. Creative generation is compressing campaign cycles from weeks to hours. Predictive models are changing which users get targeted and at what price. The competitive advantage is shifting from budget size to data quality and the speed to act on it.

The second is a geographic and platform reorientation that is quietly redrawing the map. User acquisition spend and paid installs are rising. But iOS organic installs are declining, and North America appears to be maturing across multiple categories. A growing share of subscribers is now coming from India, LATAM, the Middle East, and Southeast Asia, leading Android to cross the paid majority threshold of its own installs for the first time.

For marketers whose playbooks were built around iOS-first, North America-first acquisition, this may represent a meaningful reorientation. The economics differ: CPI, monetization dynamics, and platform mechanics all point to a different kind of growth.

The category picture adds another layer of complexity. Short Drama and OTT & Live Streaming are in aggressive global expansion mode. Ad spend from games that offer subscriptions is declining on both platforms. Health & Fitness is consolidating around a smaller group of well-capitalized iOS apps, while Generative AI is renegotiating its monetization model mid-flight.

This report maps where subscription app marketing stands today: where growth is coming from, which categories are pulling ahead, how monetization models are shifting, and how marketers are, and aren’t yet, using AI to make sense of it all.

Data sample *

1.7B

paid installs of subscription apps from Oct 2024 to Feb 2026 inclusive

$2.1B

in UA ad spend analyzed

2.9K

Subscription apps across a wide range of categories **

* All results are based on fully anonymous and aggregated data. To ensure statistical validity, we follow strict volume thresholds and methodologies and only present data when these conditions are met. ** OTT & Live Streaming, Short Drama, Health & Fitness, Education, Utility & Productivity, Gaming, Dating, Business, Photo & Video, Lifestyle, Generative AI, Books & Reference, Graphic & Design.

03 Top trends

UA ad spend rises 24% YoY with Streaming, Short Drama surging

Total user acquisition (UA) ad spend among subscription apps grew +24% year-over-year (Oct 25-Feb 26 and Oct 24-Feb 25) with Android driving the rise with +42% vs. iOS at +10%.

OTT & Live Streaming is the standout on both platforms: Android +240%, iOS +120% with North America Android surging 157%. Short Drama Android grew 42% overall, but the geo reallocation is the real story: North America (-40%) and SEA (-39%) fell sharply while the Indian Subcontinent surged +423% and LATAM grew +77%. Marketers are rotating budget toward markets where install efficiency is still high.

Education Android +114% is dominated by the Indian Subcontinent, jumping 184%, while Health & Fitness shows the sharpest platform divergence: iOS up 87%, Android down -41%. The Android decline is broad-based across every geo, while iOS spend is concentrating where subscription LTV justifies the higher CPI.

Gaming is cutting everywhere: Android -14%, iOS -32%. North America Android saw a -21% drop, whereas Eastern Europe was the one exception at 20% growth. Utility & Productivity splits cleanly by platform: Android rising 93% while iOS falling -23%. On Android, Indian Subcontinent and Middle East are driving the increase; iOS pullback is concentrated in North America and Western Europe.

The overarching signal: spend is rotating toward Android in volume-driven categories (Short Drama, Education, OTT, Utility & Productivity) and toward iOS where conversion quality matters more (Health & Fitness, Dating). Gaming is being defunded on both platforms.

When it comes to the ad spend split by category, we can also see that iOS concentrates in monetization-quality categories: Health & Fitness (23%), Dating (14%), Utility & Productivity (17%), and Business (9%) together account for 63% of all iOS UA spend. The equivalent categories on Android total just 14%. 

Android in turn dominates volume categories: Gaming (25%), Education (21%), and Short Drama (19%) account for 65% of Android spend, versus only 13% on iOS. OTT is the one category where both platforms are betting equally (11% share on iOS and Android) — the only major category with no platform skew.

Year-over-year % change in user acquisition ad spend for subscription apps *

* Comparing Oct 25-Feb 26 and Oct 24-Feb 25

Ad spend split by category among subscription apps (Oct 25-Feb 26)

---

Budget concentration: Winners keep winning in fast-growth categories

An analysis of market concentration among the top 5 apps in each category shows that Short Drama and OTT & Live Streaming are structurally concentrated at 90%+ and staying that way in each time frame. The likely cause: both categories are content-driven, where catalog depth and brand recognition create compounding advantages. New entrants can acquire users but can’t easily displace the content libraries of incumbents. 

Concentrating further are Health & Fitness, where the top 5 share jumped from 54% to 73%, the largest shift in the dataset. Spend grew 87% on iOS but flowed to a small group. It appears that subscription LTV in this category is high enough that only well-capitalized incumbents can afford the CPIs while smaller apps get priced out of the auction as competition intensifies.

Democratizing trends lead Photo & Video top 5 share to drop from 64% to 45%. The most plausible cause is a wave of AI-powered photo and video tools entering the market with enough differentiation to attract UA budgets, disrupting incumbents rather than fighting for the same positioning.

Gaming, Dating, and Utility & Productivity barely moved despite very different spend trajectories. Gaming is shrinking overall (-14% Android) but no single app is consolidating the cuts. Dating is growing (+86% Android) but gains are spread as the category is inherently geo- and demographic-specific, limiting any single app’s addressable market.

The fastest-growing categories (Short Drama +155%, OTT +170%) also have the highest concentration. One explanation is a flywheel effect where leaders use install volume to train better bidding algorithms, lowering their effective CPI relative to challengers over time.

Concentration of ad spend among top subscription apps by category

---

Android paid installs up 57% YoY vs. only 9% on iOS

Paid install growth in subscription apps remains strong, but the real story is where it’s coming from. Overall paid installs grew 47% YoY in Oct 2025–Feb 2026 vs. Oct 2024–Feb 2025, and Android is driving the bulk of it.

Android paid installs grew 57% YoY vs. 9% for iOS. The former’s paid share of total installs rose from 43% in the prior period to 51% — paid has crossed the majority threshold. iOS moved from 30% to 34% over the same period. The organic signal makes the gap more meaningful: Android organic was up only 13% while iOS organic declined 8%.

Lower CPIs, broader addressable reach, and improving monetization on Android have made it the growth engine for paid acquisition. iOS remains in the mix but with single-digit YoY growth.

Two categories accounted for the bulk of Android’s net paid install gain. Short Drama drove 60% of the total delta (+155% YoY). OTT & Live Streaming contributed 13% (+170%). Education added 22% of the delta (+66%), concentrated almost entirely in the Indian Subcontinent. 

Geographically, Indian Subcontinent alone accounted for 49% of the net Android delta (+95%), followed by LATAM (+52%, 18%) and SEA (+36%, 11%). Eastern Europe (+95%) and Middle East (+197%) contributed a combined 18%. North America was essentially flat.

A look at the share of paid installs, we can see that Short Drama and OTT & Live Streaming are both moving toward more paid, but from different starting points. Short Drama Android was already predominantly paid (77%) and edged up to 81% in Oct 2025–Feb 2026 vs. Oct 2024–Feb 2025. OTT made a bigger jump: Android +56% (25% to 39%), iOS +34% (35% to 47%), with Eastern Europe Android (+131%) and Middle East iOS (+200%) leading. Japan & Korea and SEA moved the other way as organic carries growth without UA support.

Utility & Productivity and Generative AI point in opposite directions. Utility & Productivity iOS is getting more paid-dependent: North America crossed 50% (+55%, 33% to 51%), while Android’s flat global average masks a sharp organic takeover in Japan & Korea (-51%, 81% to 40%). Generative AI fell on both platforms (-32% Android, -45% iOS), with Indian Subcontinent Android (-81%) seeing the steepest drops. LATAM iOS is the exception at +156% (18% to 46%), still in active acquisition mode.

Year-over-year % change in paid installs for subscription apps *

* Comparing Oct 25-Feb 26 and Oct 24-Feb 25

Share of paid installs trend by region and category

---

OTT doubles down on pure subscription, Short Drama and Gaming diversify

Across subscription app categories, monetization models are moving in different directions and the divergence is more complex than a simple shift toward or away from pure subscription.

OTT & Live Streaming is the clearest consolidation story. IAS Only share rose from 53% to 62% (+17%) while IAS & IAP fell from 47% to 38%. The direction is consistent month over month with no reversal. This aligns with the category’s aggressive UA investment (+240% Android, +120% iOS): when the monetization model is clean and predictable, marketers are more willing to spend heavily on acquisition.

Short Drama is moving in the opposite direction, and the shift is notable. IAS & IAA went from near zero (0.9%) to 7.4% in Period B, while IAS Only fell from 46% to 42%. Apps in this category are layering in ad-supported models, likely a response to the economics of emerging markets where subscription conversion rates are low relative to install volume. The category is scaling fast on installs but appears to be testing whether subscriptions alone can convert that volume into revenue.

Gaming shows the most complex model evolution. IAS Only rose modestly (45% → 49%), but the more significant move is IAS & IAP growing from 29.5% to 37% (+25%) while IAS & IAA dropped sharply from 12.8% to 7.1% (-44%). Gaming is actively trading ad revenue for in-app purchases, which seems to be a deliberate shift away from ad-dependent models toward higher-value transactions, even as overall UA spend declines.

Generative AI continues adding IAP. IAS Only fell from 65% to 60% while IAS & IAP grew +20%. Apps that launched as pure subscription are hedging with purchases, likely to capture users unwilling to commit to recurring billing, which is consistent with the pattern across multiple prior quarters.

Lifestyle is adding ads, not simplifying. IAS & IAA grew from 5.4% to 9.2% (+70%) while IAS & IAP declined. Dating is slowly consolidating around pure subscription: IAS Only moving from 51% to 56% as IAS & IAP gradually retreats, consistent with the category’s growing paid UA investment.

Share of apps split by business model *

IAS Only: Apps generating revenue from subscriptions only 
IAS & IAP: Apps generating revenue from subscriptions and in-app purchases
IAS & IAA: Apps generating revenue from subscriptions and in-app ads
IAS & IAP & IAA: Apps generating revenue from subscriptions, in-app purchases, and in-app ads

---

The subscription funnel: Not all trials are created equal

The subscription conversion funnel shows Gaming has the highest free trial adoption but the worst conversion. At 12.2% install-to-trial, Gaming pulls more users into trials than any other category, but only 19% of those convert to paid, the lowest CR in the dataset by far. Every other category converts at 32%+. Gaming’s trial model appears to be a discovery mechanism rather than a purchase intent signal, meaning users try but then churn. High trial volume is masking a monetization problem, not solving it.

Education and Lifestyle punch above their weight on conversion. Education converts 42% of trialists to paid (highest in the dataset) suggesting users who start a trial are highly motivated and the product delivers enough value to justify the subscription. Lifestyle also converts 41% despite the lowest install-to-trial rate (3.9%), meaning it attracts fewer trialists but keeps nearly half of them. Both categories are the opposite of Gaming: lower funnel volume, higher funnel quality.

The no-trial chart reveals a different competitive landscape. Health & Fitness (7.1%) and Dating (6.5%) lead paid subscription conversion without a free trial. These are categories where users arrive with high purchase intent and don’t need a trial to decide. 

Short Drama and Utility & Productivity sit at the bottom of the no-trial funnel at 1.8% and 1.7% respectively, well below the 3.5% overall average. For Short Drama this is expected as the category is almost entirely UA-driven with strong acquisition scale but weak monetization conversion relative to install volume.

The subscription funnel (Oct 25-Feb 26)

---

How subscription marketers use AI chat: more to retrieve, less to reason

The pattern across over 4,000 queries tells a story about how subscription marketers relate to their data. Nearly 40% of all queries fall into channel performance and cost topics. This is not because these queries are the most strategically important questions, but because those are the questions the tool makes easy to answer. The AI assistant is being used for the path of least resistance.

The 6% diagnosis rate within the Media Source and Optimization bucket is equally telling. Most users ask “which campaign had the highest ROAS D7” rather than “why did ROAS drop on this campaign last week.” The retrieval-to-diagnosis ratio is roughly 10:1. This likely reflects a prompting ceiling: most users don’t structure a diagnostic question for an AI, so they default to ranking and lookup tasks they know will return a clean answer.

The Business category’s 18% Retention & Cohort share, which is 4x the overall rate, suggests that when the business model demands lifecycle thinking, it surfaces in how marketers prompt. Media & Entertainment’s 27% Chart Summarization share points to a different dynamic: these marketers are using the tool as a passive reader rather than an active analyst, asking the AI to interpret dashboards rather than investigate performance. Both are legitimate use cases, but neither reflects the full potential of AI-assisted decision-making.

Dating leads on Install & Attribution at 19%: nearly 2.5x the overall rate, which is consistent with a high-CPI vertical where every install is scrutinized and attribution accuracy has direct budget implications. When acquisition costs are high, marketers ask sharper questions.

AI Assistant questions split by type among subscription marketers *

* Based on analysis of over 4,000 questions from subscription apps in AppsFlyer’s AI Assistant

05 key takeaways

Skip numbered cards section

Follow the Android money east

The next subscriber is more likely to come from India, LATAM, or the Middle East than from North America. Build Android-first acquisition strategies for emerging markets: different creatives, different price points, different monetization expectations. The playbooks built for Western iOS users won't transfer directly.

Audit your trial funnel before scaling spend

Gaming is spending heavily on trials that convert at 19%. Before increasing UA budget, establish whether your trial is generating purchase intent or just free usage. Education and Lifestyle show that lower trial volume with higher conversion is a more efficient path.

Don’t assume pure subscription is enough

Pure subscription is gaining ground in OTT & Live Streaming, but Short Drama is layering in ads and Gaming is trading ad revenue for IAP. Generative AI is adding IAP as a fallback for users unwilling to commit to recurring billing. No single model is winning across the board. Audit your current mix against your category benchmark.

Compete on differentiation in a concentrated market

Short Drama and OTT & Live Streaming top 5 apps control 90%+ of spend. Health & Fitness is consolidating fast. In these categories, incremental UA budget without a genuine product or content advantage will not move the needle. Photo & Video shows that AI-powered differentiation can break concentration but it requires a distinct angle.

Ask your AI tools harder questions

The data shows subscription marketers using AI assistants almost entirely for retrieval: rankings, lookups, top-line performance. The diagnostic questions like why did ROAS drop, what's driving churn in this geo, and which campaign mix is underperforming are largely going unasked. The tools can answer them.

About the authors

Shani Rosenfelder

Director of Global Content Strategy & Market Insights at AppsFlyer

Shani is the Director of Global Content Strategy & Market Insights at AppsFlyer. He has over 10 years of experience in key content and marketing roles across a variety of leading tech companies and startups. Combining creativity, analytical prowess and a strategic mindset, Shani is passionate about building a brand’s reputation and visibility through innovative, content-driven projects.

Follow Shani Rosenfelder:

Yuval Hay Hirsh

Data Analyst

Yuval Hay is an Industry Data Analyst at AppsFlyer, with 4 years of experience specializing in marketing data exploration and research. Passionate about uncovering insights and identifying new data opportunities, Yuval Hay helps drive strategic decision-making through deep data analysis and exploration.

Follow Yuval Hay Hirsh:

Ready to start making good data driven choices?

Request a demo Sign up free

The post [Report] The State of Subscriptions for Marketers – 2026 Edition appeared first on AppsFlyer.

The State of Finance for Marketers in APAC – 2026 Edition

01 KEY findings

17%

decline in APAC finance installs in 2025, the first drop on record, even as iOS hit a record 16% share

iOS grew 8% year-over-year while Android installs fell 20%, reinforcing platform mix as a central performance lever. As acquisition budgets tighten, install distribution is increasingly focused on premium-device segments to improve monetization.

2x

year-over-year surge in the number of paid installs among Investment apps in Indonesia

Across finance apps more broadly, paid installs also increased in growth markets such as Vietnam, rising 75% over the same period. Meanwhile, mature ecosystems moved in the opposite direction, with Australia’s paid install rate declining 70% as acquisition budgets tightened.

47%

of SEA Android UA spend came from Chinese apps in 2025, up from 38% the year prior

As overall budgets contracted, spend concentrated in fewer markets. In ANZ, UK-based apps dominated with 47% of iOS UA spend, while India’s Android spend remained 92% domestic.

193% year-over-year surge in SEA remarketing spend, signaling the start of a structural shift toward lifecycle investment

The acceleration spanned Indonesia (+144%), Thailand (+339%), the Philippines (+241%), and Vietnam (+161%). Android Day 30 retention in SEA rose from 2.35% to 3.86% over the same period, suggesting that lifecycle investment is translating into measurable engagement gains.

39% of APAC Android finance IAP revenue came from Indonesia in 2025, with India second at 27%

South Korea posted the fastest growth in the region, with IAP revenue up 52% on Android and 63% on iOS. Nearly 40% of global finance IAP revenue originated outside the top contributors, underscoring the geographic breadth of monetization across APAC.

22% fraud rate across APAC finance in 2025, down by nearly half from the year before

The improvement was seen across categories, with Investment dropping from 26% to 8% and Personal Loans from 31% to 14%. Elevated risk persisted in specific markets. However, Vietnam recorded 46% and Bangladesh 35%, requiring continued active monitoring.

02 introduction

Finance pivots from hyper-growth to sustainable value

The APAC region remains the global heavyweight for digital economic expansion, forecast to grow 4.7% in 2026, despite global trade uncertainties. For millions of users, the smartphone has evolved from a simple communication tool into the primary engine of financial participation, powering a digital economy that is set to exceed $300 billion in gross merchandise value by the end of 2025.

However, the “growth at all costs” era has ended. Capital is no longer deployed broadly across expansion markets. It is being reallocated toward higher-efficiency segments, premium device cohorts, and lifecycle depth. In the first half of 2025, fintech funding in the region moderated to approximately $4.3 billion, a multi-year low that signals a decisive move away from speculative expansion.

This capital discipline is forcing a structural maturation across the sector. Established platforms are no longer rewarded for vanity metrics; the new mandate is profitability. We are already seeing this play out in Southeast Asia (SEA), where the digital economy has grown profits by 2.5 times over the last two years, reinforcing a broader regional shift from land-grabs to unit economics.

For marketers, this redefines success. Performance is now measured by how efficiently budgets are spent, platform mix quality, retention depth, and monetization durability.

Our report provides a data-driven perspective on how this strategic pivot is unfolding on the ground. Our analysis of app behaviors offers insights into where the market is correcting, where hyper-growth opportunities persist, and where the next wave of high-value users is emerging.

Data sample *

5.31B

finance app installs in 2025

$5.7B

in UA ad spend in 2025

$29M

in remarketing ad spend across analyzed APAC finance markets in 2025

*All results are based on fully anonymous and aggregated data. To ensure statistical validity, we follow strict volume thresholds and methodologies and only present data when these conditions are met.

03 Overall Key Trends

Finance installs fall for the first time, even as iOS hits record 16% share

Finance app installs in APAC declined 17% year-over-year in 2025, the first broad correction the region has seen, even as iOS reached a record 16% share, up 33% since 2024.

iOS accounted for 16% of total APAC finance installs in 2025, its highest share on record. Over the same period, Android installs declined 20% year over year. As overall volume contracted, a greater share of installs came from premium-device cohorts.

This shift carries economic implications. Finance apps consistently record stronger in-app purchase intensity and monetization depth among iOS users. As acquisition budgets tighten, install distribution is concentrating in segments with stronger revenue characteristics.

Geographic performance reinforced the pattern. India remained APAC’s largest volume contributor, accounting for over 40% of total installs, though installs declined 22% year over year. Thailand fell 40% and Vietnam 22%, reflecting maturing adoption in markets where financial app penetration is already widespread.

Pakistan grew 61%, consistent with earlier-stage digital finance adoption. South Korea expanded by 9%, standing out among mature ecosystems. Growth in Korea reflected continued engagement expansion within an established financial services infrastructure rather than first-time access dynamics.

In 2025, APAC finance was increasingly shaped by tighter, more focused spending decisions. Platform mix, device economics, and monetization intensity are becoming primary determinants of where growth persists.

Overall install trend (normalized)

Share of installs by platform

---

21% paid install rate: Indonesia rises as Australia declines

While overall finance installs declined year over year, the share of installs driven by paid campaigns varied widely across APAC markets. Growth ecosystems expanded paid acquisition, while mature markets pulled back amid budget tightening.

Indonesia recorded one of the clearest increases in paid install rates across finance apps during 2025. The paid install rate rose steadily through the year, reaching roughly 21% by Q4, nearly double the levels seen in early 2024. Vietnam followed a similar trajectory, with paid install rates climbing from 12% to about 21% over the same period.

Other markets showed more stability. The Philippines remained broadly consistent throughout the year, maintaining a balanced mix between organic and paid acquisition.

In contrast, mature ecosystems moved in the opposite direction. In Australia, the paid install rate declined sharply, falling from 49% in early 2024 to roughly 15% by late 2025, reflecting tighter Android budgets and a stronger focus on acquisition efficiency.

At the category level, the divergence becomes even more pronounced. In Indonesia’s Investment segment, paid install share surged from 22% in Q4 2024 to 45% in Q4 2025, more than doubling year over year and highlighting continued expansion in markets with strong monetization potential.

Across APAC, the pattern is clear: growth markets continue to lean into paid acquisition, while mature ecosystems are prioritizing efficiency and lifecycle engagement.

Paid Installs Rate (normalized) – Finance Overall

---

UA spend declined 27% across APAC as Android budgets contracted

User acquisition investment contracted sharply across APAC finance in 2025, reinforcing the capital discipline theme observed throughout the year.

Across analyzed APAC markets, total UA spend declined approximately 27% year over year. The Indian Subcontinent recorded the steepest contraction at 38%, driven primarily by a 43% decline in Android spend in India. Southeast Asia declined 27%, with Android spend falling 25% and iOS contracting 48%.

In contrast, Japan and South Korea recorded modest overall growth of 3%, reflecting more stable acquisition environments. Australia increased 16% year over year, driven entirely by iOS growth of 86% while Android declined 32%, signaling a clear platform shift rather than broad expansion.

The contraction was uneven across markets. India fell 42%, Vietnam declined 48%, and Thailand declined 55%, confirming widespread budget compression in scale-driven ecosystems. Pakistan and Bangladesh expanded, reflecting earlier-stage growth markets with smaller budget bases.

Contraction hit hardest in large Android ecosystems, while iOS and mature markets proved more resilient.

YoY % change in the UA Ad Spend (2025 vs. 2024)

UA ad spend by platform in 2025 (USD)

---

47% of SEA Android UA spend came from China-based apps

As UA budgets tightened across APAC finance in 2025, spend concentrated in fewer markets rather than being evenly distributed.

In Southeast Asia, China-based apps accounted for 47% of Android UA spend, up from 38% in 2024. Even as overall SEA spend declined 27%, their relative share expanded, indicating that a greater proportion of remaining budgets flowed through a single dominant hub. Much of this allocation was directed toward Indonesia and the Philippines, reinforcing Chinese apps’ structural influence within the sub-region.

A comparable pattern emerged in ANZ. UK apps accounted for 41% of Android inbound spend and 47% of iOS inbound spend in 2025. UK-origin budgets increased year over year despite broader regional contraction, solidifying its position as the primary outbound driver into ANZ.
The Indian Subcontinent followed a different configuration. Android allocation remained overwhelmingly domestic, with India accounting for 92% of spend within the sub-region. In contrast, iOS investment was more diversified across China, Israel, Russia, and the UK, with no single external hub exerting dominant control.

China’s Android budgets were heavily concentrated in Southeast Asia, with minimal allocation to India — targeted concentration rather than universal dominance.

YoY % change in the share of finance app UA spend of China-HQ apps

Inbound UA spend by source market, 2025

---

SEA remarketing spend surged 193% as lifecycle investment intensified

While user acquisition budgets contracted across APAC finance in 2025, remarketing investment accelerated in selected markets, an early but meaningful shift in spending from user growth toward engagement depth. UA budgets still far outpace remarketing overall, but the gap is narrowing.

Southeast Asia recorded the strongest expansion. Overall remarketing spend increased 193% year over year, driven by a 202% rise on Android and an 84% increase on iOS. Indonesia grew by 144%, Thailand by 339%, the Philippines by 241%, and Vietnam by 161%. The magnitude and breadth of these increases suggest structural reprioritization rather than short-term optimization.

Japan and South Korea followed a steadier trajectory. Remarketing spend rose 74% overall, with South Korea up 85% and Japan up 56%, reflecting sustained lifecycle investment in mature ecosystems.

The Indian Subcontinent diverged. Total remarketing spend declined 8%, as Android, the dominant platform, fell 9%. Although iOS remarketing increased 101%, it expanded from a much smaller base, limiting its impact on total regional spend.

Unlike user acquisition, which remained shaped by cross-border hubs, remarketing activity was largely domestically anchored. In Southeast Asia, Android remarketing was concentrated in Indonesia (30%), Thailand (28%), and the Philippines (22%), reinforcing the idea that lifecycle investment followed market scale rather than outbound capital flows.

Remarketing ad spend by platform in 2025 (USD)

Remarketing Conversion Trend by country (normalized)

---

Japan & Korea lead Day 30 retention as SEA closes the gap

Retention improved across multiple APAC finance markets in 2025, reinforcing engagement depth as a growing performance priority in a capital-constrained environment.

Southeast Asia recorded steady gains. Android Day 30 retention increased from 2.35% in Q1 2024 to 3.86% in Q4 2025, while iOS rose from 5.50% to 8.87% over the same period. At the market level, Indonesia improved from 1.92% to 3.37%, Thailand from 1.97% to 3.05%, Vietnam from 1.75% to 4.23%, and the Philippines from 0.98% to 2.95%.

India showed steady improvement. Android Day 30 retention increased from 2.48% in Q1 2024 to 4.36% by Q4 2025, while iOS rose from 6.29% to 9.65% over the same period. The improvement reflects deepening engagement within an established financial app ecosystem where usage extends beyond transactional entry into recurring financial management behavior.

North Asia showed more moderate but stable improvement. Android Day 30 retention in Japan and South Korea ranged from 6.72% in Q1 2024 to 7.24% in Q4 2025, holding steady at approximately 7% throughout the period. This is the highest Day 30 Android retention across all APAC sub-regions, consistent with markets where financial app engagement is already habitual and incremental gains reflect refinement rather than adoption expansion.

As install-led growth moderated, retention became a central indicator of performance durability heading into 2026.

Session trend: Finance – overall (normalized)

Average Retention trend: Finance – overall apps by sub-region

---

39% of APAC Android finance IAP revenue came from Indonesia

In-app purchase revenue patterns in 2025 reflected both concentration and geographic breadth across APAC finance markets. A small number of economies anchored platform-level monetization, while revenue generation remained widely distributed across the region.

Within APAC, Indonesia accounted for 39% of Android IAP revenue, followed by India at 27% and South Korea at 14%. Globally, Indonesia accounted for 12% of total Android finance IAP revenue, ahead of India and the United States at 8% each. Brazil contributed 7%, reinforcing Android’s monetization strength in large-scale emerging ecosystems.

On iOS, the story is similar. Indonesia led within APAC at 33%, while South Korea outperformed its Android share at 18%, signaling stronger monetization intensity on premium devices.

Although Indonesia and India remained APAC’s dual-platform anchors, their relative global revenue shares declined between 17% and 19% year over year across both platforms. The shift reflects faster revenue growth in other markets rather than outright revenue contraction.
South Korea recorded the strongest year-over-year expansion in the region, with finance IAP revenue increasing 52% on Android and 63% on iOS. The acceleration narrows the gap with the region’s established leaders and signals rising monetization depth in premium-device ecosystems.

Revenue remained broadly distributed. Nearly 40% of global finance IAP revenue originated from markets outside the top contributors, underscoring the geographic diversity of monetization across APAC and beyond.

As scale-led growth moderated, monetization depth, not install volume, became the defining performance metric heading into 2026.

YoY % Change in the share of IAP Revenue

Share of in-app purchase revenue by country

---

Fraud rates down 22% as traffic quality improves

Install fraud across APAC finance declined materially in 2025. The regional average fraud rate fell from 41% in 2024 to 22% in 2025, reflecting strengthened detection frameworks and a broader shift toward higher-quality traffic sources. However, with more than 1 in 5 installs identified as fraud, vigilance should remain high.

The improvement was most visible in historically high-risk verticals. Investment fraud rates declined from 26% to 8%, marking the lowest exposure among finance categories in 2025. Personal Loans improved from 31% to 14%, while Mobile Banking remained the highest-risk vertical at 29%, despite a meaningful reduction from prior-year levels.

Platform exposure converged. iOS recorded an average fraud rate of 21%, compared to Android’s 20%, narrowing gaps observed in previous periods when platform-specific spikes were more pronounced.

At the market level, outcomes diverged. Thailand and South Korea recorded fraud rates of 3% each, positioning them among the lowest-risk ecosystems in the region. The Philippines reached 9%, and India 11%, reflecting sustained quality improvement.

However, elevated exposure persisted in specific markets. Vietnam recorded 46% and Bangladesh 35%, while Australia and ANZ remained materially above the regional average. Fraud has compressed, but it hasn’t disappeared;  risk remains concentrated in specific markets and verticals heading into 2026.

Fraud Rate – Finance Overall

Fraud Rate – Investment

Fraud Rate – Mobile Banking

Fraud Rate – Personal Loans

---

In-app event queries reach 12% in Mobile Banking

Analysis of more than 2,000 questions submitted by Finance marketers to AppsFlyer’s AI Assistant shows that AI is primarily used for performance interpretation, with attribution emerging as the dominant use case.

Personal Loans recorded the highest share of attribution queries at 20.7%, nearly double Mobile Banking. The longer conversion cycles and higher CAC typical in lending funnels increase the need to understand which sources actually drove installs and downstream activity.

Usage patterns vary by vertical. Mobile Banking generates the highest share of In-App Event questions (12.2%), reflecting onboarding-heavy funnels where activation milestones determine early performance. Investment apps display a more observational usage pattern, producing the highest share of Chart Summarization queries (9.4%) and the lowest concentration of Media Source questions (24.8%), suggesting greater emphasis on interpreting dashboards rather than actively adjusting acquisition channels.

Prompt behavior reveals a second pattern. Attribution queries consistently produce the highest share of advanced prompting (34–40%) across all finance subcategories, indicating that marketers provide more context when the analytical stakes are higher. A similar dynamic appears in channel analysis, where Personal Loans show significantly higher advanced prompting for Media Source queries (32%) than Mobile Banking (14%).

Together, these patterns suggest that Finance marketers increasingly rely on AI to interpret complex performance signals, while prompting sophistication rises when decisions directly impact acquisition spend.

AI assistant queries split by type among Finance marketers*

*Based on analysis of over 2,000 questions by Finance marketers in AppsFlyer’s AI Assistant

Share of queries with advanced prompting

---

Omnichannel fragmentation persists despite high digital engagement

Mobile apps have become the primary banking touchpoint across APAC and globally. According to Accenture’s Global Banking Consumer Study 2025, customers average 150 app interactions per year, the highest frequency among service channels. Mobile apps also receive the highest satisfaction scores at 60%, reinforcing their central role in daily financial activity.

However, engagement depth does not translate uniformly across the customer journey. Most interactions remain transactional. Checking balances accounts for 45% of weekly activity, while money transfers represent 31%. When customers move beyond routine actions into more complex needs such as loan applications, dispute resolution, or service escalation, continuity across channels often weakens.

Accenture’s research also highlights a perception gap. Only 21% of banking service executives view customer service as a value driver. Separately, in a cross-industry study on customer service, 32% of consumers believe service quality has improved in the past five years, and just 18% feel that technology has enhanced their overall experience.

The findings point to a structural challenge. Digital engagement is high, but omnichannel continuity remains uneven, particularly during high-friction or high-value interactions. Deep linking plays a key role here, connecting touchpoints seamlessly and ensuring users land in the right place regardless of where they started. As financial institutions deepen digital investment, seamless coordination across app, web, paid, and physical touchpoints remains a critical area for improvement.

Accenture Data: Channel Usage and Satisfaction

---

04 Experts’ corner

Rohan Dhanuka

Sector Lead, SG & MY Core, Google

Many finance brands are shifting focus from cost-per-acquisition toward activation, engagement, and lifetime value. From what you see globally, which signals best indicate high-quality growth, and how should platforms support optimization toward those outcomes?

Volume is a vanity metric in a challenged macro environment; value is the reality. The finance world is overcrowded. The average customer splits their finances across 5+ relationships. The goal isn’t more customers; it’s capturing more value from the ones you have. High-quality growth is measured by Product Density, the proxy for “Share of Wallet.” e.g., Can you turn a mortgage lead into a wealth relationship? A good starting point is to focus on value metrics rather than simple CPA metrics.

• The Shift: We are seeing a global move toward Value-Based Bidding (VBB). According to Think with Google research on the e-Conomy, the top 30% of spenders account for 70% of total digital spend – those are the relationships to win.


• Case in Point: Zoe Financial optimized their bidding for the value of each lead rather than just the number of applications. By identifying what made an “ideal client” in their CRM and feeding that signal back to their campaigns, their most valuable client segment rose to 60% of total sales.


• Strategic Tip: Stop reporting on “Total Leads.” Start reporting on “Value Delivered”. That is where the real margin is hidden.

Customer journeys in financial services increasingly span web, app, paid, and owned environments. As data in these environments is fragmented and customer journeys become more complex, how should marketers evaluate performance beyond individual channel metrics to ensure they are optimizing for true business impact?

If your data is in silos, you’ve likely hit your ceiling. Stop optimizing for individual channels and start optimizing for a multi-modal customer. The modern consumer doesn’t live in “owned” or “paid” silos. They shift between high-intent search, immersive video, bank apps, and product websites, sometimes within a single decision journey. Channel-level metrics can’t capture this. The question isn’t “how did this ad perform?” — it’s “where is this customer in their journey, and what’s the next best action?”

To answer that, you need a unified data stack that connects signals across every touchpoint — not just your CRM, but the full funnel. As highlighted in Think with Google’s Digital Transformation playbooks, the critical link is between your First-Party Data and external ecosystems — so that every signal, from every touchpoint, is visible, connected, and actionable.

• Case in Point: Agicap moved beyond “casting a wide net” by feeding their HubSpot CRM data directly into Google Ads, shifting from channel volume to “prospect life cycle” bidding, and nurturing high-intent leads at the right moment.


• The Architecture: A centralized warehouse (e.g., BigQuery) that ingests signals across the funnel lets AI like Performance Max find your next customer, whether they’re scrolling, streaming, or searching.

Trust is central in financial services, where relevance must align with compliance and brand integrity. How are leading brands using AI and automation to deliver contextual messaging across environments without compromising consistency or customer trust?

In finance, trust IS the product. Brand isn’t a billboard; it’s a trust-transfer mechanism. Trust is the single most critical variable in the FSI equation. In periods of macro-volatility, consumers don’t look for the cheapest rate; they look for reassurance. Leading brands recognize that their brand equity is a “future demand” generator that keeps them above the “demand ceiling.”

• The Balancing Act: According to the Google/WARC whitepaper on brand value, emotional bonds guide 95% of consumer decisions. For a bank, trust is that emotional bond.


• Consistency at scale: Trust erodes when a consumer sees a personalized wealth offer on mobile that doesn’t match the in-branch experience.” Leading brands use AI to ensure contextual messaging stays consistent across 70+ environments without losing that “human” touch.


• The Insight: Trust and performance aren’t trade-offs. An ad that creates brand interest and trust is statistically more likely to drive short-term sales. Think of your brand as a “Trust Balance sheet” — every insight-driven, secure interaction is a deposit.

AI is increasingly deployed in campaign optimization, targeting, and creative optimization. In highly regulated sectors like financial services, what governance and compliance practices should marketing teams put in place to use AI responsibly — and how do you build consumer trust in the process?

AI is increasingly deployed in campaign optimization, targeting, and creative optimization. In highly regulated sectors like financial services, what governance and compliance practices should marketing teams put in place to use AI responsibly — and how do you build consumer trust in the process?

In FSI, security isn’t a feature; it’s the foundation that enables responsible growth. In a highly regulated sector, being “secure in your data” is the prerequisite for innovation. You cannot unlock meaningful growth without Responsible AI foundations. This isn’t just about privacy; it’s about Explainability — the ability to tell a regulator why an AI made a specific decision.

Google’s Responsible AI Principles (Safe, Secure, Fair, and Accountable) are critical here.

• Security by Design: Use technologies like Data Clean Rooms or Confidential Computing to process sensitive PII without exposing it.


• The “3Rs” of Responsible AI: Implement the 3R framework (Regulation, Reputation, Realization). You comply with acts like the EU AI Act not just to avoid fines, but to preserve your brand’s reputation for fairness.


• Practical Next Step: Audit your AI pipeline for bias. If your model training data is skewed, your outcomes will be too. High-fidelity, clean data is the only way to build an AI engine that regulators will actually let you switch on.

As marketing, product, and data functions become more interconnected, how do you see the operating model of high-performing finance marketing teams evolving over the next few years, particularly in measurement, budget allocation, and AI-supported decision-making?

The wall between the “creative” and the “cloud” has fallen. If your marketing team can’t talk to your data team, you’re already behind. The operating model of high-performing FSI teams is evolving into a unified intelligence stack. The line between IT, Product, and Marketing is blurring. In this new world, the “generalist” — someone who understands data architecture and creative storytelling — is the most valuable asset.

• The CMO-CFO Power Couple: High-performance teams have a tight alliance between the CMO and CFO, using shared KPIs such as CLV (Customer Lifetime Value) rather than siloed marketing metrics.


• The “Orchestrator” Model: We are moving away from fixed budget caps to demand-led pacing. AI agents handle the routine campaign setups (the “grunt work”), while the marketing team focuses on “Moonshot” strategy and creative distinctiveness.


• Evolution: Measurement will move toward MMM-Calibrated Attribution (like the Meridian model), providing a single source of truth that the whole C-suite can agree on. The future belongs to the teams that can pivot from “buying media” to “orchestrating business outcomes.”

05 key takeaways

Skip numbered cards section

Shift from expansion to disciplined capital allocation

APAC finance exited its hyper-growth phase in 2025. Installs declined 17% year over year, UA budgets compressed across major regions, and capital deployment became more selective. Growth is increasingly defined by allocation efficiency, platform targeting, and corridor prioritization rather than regional scale expansion.

Capture value through platform mix

Premium device segments reached a record 16% install share as overall volumes cooled. iOS continued gaining relevance even amid contraction, reinforcing platform mix as a performance lever. Monetization intensity remains strongest in premium-device ecosystems, including South Korea and Indonesia.

Prioritize lifecycle as acquisition efficiency tightens

Remarketing spend increased by 193% in Southeast Asia and 74% in Japan and South Korea, while UA contracted. SEA Android Day 30 retention rose from 2.35% to 3.86%, and Japan and South Korea held the highest Day 30 Android retention in APAC at approximately 7%, ranging from 6.72% to 7.24%. Markets that intensified lifecycle investment recorded stronger post-install engagement outcomes.

Deepen penetration within proven verticals and markets

Performance diverged across markets and subcategories. Investment paid install share doubled in Indonesia, Digital Wallets expanded from 9% to 34% in Southeast Asia, and South Korea recorded the fastest IAP revenue growth in APAC. Returns in 2025 were concentrated in markets and verticals with established monetization depth.

Maintain vigilance as fraud exposure concentrates

The regional average fraud rate declined from 41% to 22% in 2025, reflecting improved traffic quality and stronger filtering. However, elevated exposure persisted in specific corridors and high-risk verticals, including Mobile Banking. As budgets tighten, risk management remains central to performance durability.

Ready to start making good data driven choices?

Request a demo Sign up free

The post [Report] The State of Finance for Marketers in APAC – 2026 Edition appeared first on AppsFlyer.

Mobile banking is now a primary channel — but growth isn’t just about driving installs. Users expect frictionless journeys across web, app, and branch, while regulators and customers demand uncompromising privacy and security.

In this guide from AppsFlyer, you’ll get a practical, measurement-first playbook for scaling acquisition and re-engagement without losing trust (or ROI visibility) along the way.

What’s inside:

• How banking apps became essential digital channels
• Mobile banking user acquisition: the key metrics and how to set them up
• Banking app acquisition: 7 best practices to improve performance and efficiency
• Re-engagement: the key metrics and attribution setup to measure what works
• Remarketing best practices to bring high-intent users back (profitably)

The post Winning the mobile banking growth race – Your ultimate guide to trust and growth in the age of privacy  appeared first on AppsFlyer.

TL;DR

• The theme was “Where Do We Go From Here?” — and the answer wasn’t a single destination. It was about how to measure, where to invest, what to automate, and what to protect.

• The day’s central tension: AI is changing how consumers discover, research, and buy — but the brands with the strongest owned customer relationships (especially through apps) are the ones best positioned to thrive through the disruption.

• Sessions spanned the full spectrum from Fortune 500 strategy to a 14-day AI app-building experiment, with practitioners from KFC, Snap, Home Depot, Moloco, TickPick, Fubo, Reddit, Soundcloud, Hopper, Poshmark, Mistplay, and more sharing real frameworks and real numbers.

• Five themes emerged across the day: apps as economic infrastructure, the shift toward incrementality, AI that removes friction vs. AI as theater, creative as a strategic lever, and the evolution of the marketer’s career itself.

The role of “mobile marketer” is evolving in real time. At MAMA SF 2026, the practitioners managing multi-million dollar budgets came together to work through what it all means — and what to do about it.

The theme this year was Where Do We Go From Here? A question that felt particularly pressing with AI reshaping how consumers discover and buy, measurement models evolving rapidly, and the marketer’s role itself in transition. Here’s what emerged.

The Bottleneck Has Moved

Brian Quinn, AppsFlyer’s President and GM of North America, opened with a framing that set the tone for the entire day. 

As marketing continues proliferating across more surfaces — growth strategies on mobile, web, CTV, retail media, and now answer engines and LLM interfaces — Quinn argued that apps are the one property brands truly control: logged-in, consent-rich, and where identity, loyalty, and personalization converge. In that fragmented world, the quality of your measurement layer determines whether you’re set up for success or failure. 

And as AI agents begin making autonomous budget decisions (reallocating spend, adjusting bids, rotating creative), the integrity of that measurement layer becomes existential.

“The future of marketing will not be defined by who generates the most impressions,” Quinn told the room. “It will be defined by who allocates capital most intelligently across fragmented environments.”

The integrity of your measurement layer, he argued, becomes the difference between scaling intelligence and scaling mistakes throughout your organization. That framing (apps as owned infrastructure, measurement as strategic lever, AI as amplifier) became the connective tissue for the entire day.

Four Themes That Defined the Day

1. Your App Is Economic Infrastructure, Not a Channel

Stephane Godoy, Head of Digitalization and E-Commerce at KFC, showed how his team grew to over 7 million active loyal users in less than a year and demonstrated that most companies capture only a fraction of their app’s economic impact when they measure direct revenue alone. 

His framework broke total app value into four components: direct revenue, influenced revenue, operational savings, and lifetime value lift, with influenced revenue often running 2–4x the direct number. 

“Translating into the CFO’s language is so important,” Godoy said. “Language determines funding.” (More insights and frameworks from Godoy are incorporated into this strategic playbook.)

Karissa Corbitt from Home Depot reinforced this, showing how their app has become the biggest driver of momentum in how customers engage and shop, from in-store navigation to room visualization to loyalty. 

Beth Berger of Moloco, presenting the AI Disruption Index developed with BCG, arrived at the same conclusion from a different angle: customer relationships are the durable asset of the AI future, and the mobile app is where those relationships live.

2. Measurement Is Evolving, and the Leaders Are Pulling Ahead

Sam Mulinder, Head of Marketing Science at Snap, presented a framework that resonated across the room: the Three E’s: Execution (daily optimization), Experimentation (incrementality testing), and Evaluation (longitudinal models like MMM). His point: no single methodology tells the full story, and the teams combining multiple measurement approaches are the ones making smarter allocation decisions.

He illustrated the stakes with a memorable analogy: giving a new channel 3% of your budget and declaring it ineffective is like giving Timothée Chalamet five minutes of screen time in Interstellar and deciding he can’t act. (He checked: Chalamet’s screen time was exactly 3%.)

Eran Dunsky, AppsFlyer’s Head of Product for North America, shared data showing that across hundreds of advertisers, 30% of campaigns are being significantly undervalued by last-touch models, in some cases by up to 10x. His exercise comparing two campaigns where last-touch and incrementality told opposite stories drew visible nods.

The takeaway: “Stop asking for budget. Start proving you’ve earned it.”

3. AI That Works vs. AI as Demo

Berger’s keynote reframed the app’s role for an AI-disrupted world. Presenting joint research with BCG, she made the case that as search becomes commoditized by AI overviews and brand discovery shifts to LLM interfaces, apps are uniquely positioned as a brand’s primary competitive advantage — the one surface that delivers a logged-in experience, leverages first-party data for personalization, and provides paths to diversify acquisition away from increasingly volatile search and web channels.

Her data backed it up: the Moloco/BCG research found that a third of US adults now discover brands through personal AI agents, 80% of AI-powered Google searches end without a click, and 45% of consumers already feel comfortable letting AI make purchases for them. The implication was clear: the channels being disrupted are the ones you don’t own.

But the most visceral AI demonstration came from Bobby Sayers, who built, launched, and marketed a mobile game called Quack the Planet using AI tools in 14 days, from Cursor for development to AI agents running campaign checks every three hours via AppsFlyer’s MCP. It went from concept to ROI-positive in under two weeks. 

Other presenters complemented this with real daily workflows: natural language data queries replacing SQL, creative QA tools catching errors before launch, and autonomous agents flagging broken events in Slack. The throughline: AI is most powerful when it removes friction from work practitioners already do.

4. The Marketer’s Role Is Evolving Fast

Atif Rafiq, who was the first Chief Digital Officer at McDonald’s before becoming a C-suite operator at Volvo, MGM Resorts, and beyond, delivered the day’s most personal keynote. His message to the room: stop delivering tasks and start owning recommendations. 

“Human contribution, alongside AI, is going to be here for a long time,” Rafiq said. “Future jobs are based on the ability to make recommendations work end to end.” 

Susan Ho echoed this from the founder’s side, sharing her path from $50 in the bank to a multi-million dollar exit to Hopper, complete with a “resume of failures” that earned one of the most candid applauses of the day.

What Sports Taught the Room About Marketing Moments

The sports marketing panel featuring Matt Ferrel (TickPick), Vincent Eterlet (Fubo), and DJ Capobianco (Reddit) became a masterclass in capitalizing on cultural tentpole moments. Ferrel delivered a memorable line about how he advises his marketing team to capitalize on mainstream trends: “It’s easier to ride a wave than to create one,” describing how gold medal hockey immediately spiked NHL ticket sales and how TickPick talks to fans through the lens of a fan.

Capobianco revealed how 20,000 Reddit communities were actively discussing the Super Bowl on game day and takeaways for brands eager to participate in those conversations. Eterlet described real-time campaign monitoring in 30-minute increments during high-profile moments, with the 2026 World Cup as Fubo’s next major opportunity. 

The broader takeaway: find the human moments of connection in your category, and market to them on a human level.

The Afternoon: Where Theory Became Practice

Attendees chose their own path across Growth Showcases, Leadership Circles, and Innovation Labs.

Highlights included Mistplay’s Brett Jones presenting a channel-by-channel creative playbook showing that the same creative attribute can drive 3x lift on one network and fail on another. 

Poshmark’s Eric Kovalkoski shared how his team moved from reactive silos to a systemized omnichannel model with podded teams and shared signal dashboards. Mackenzie Quinn argued that most retention outcomes are shaped before lifecycle marketing begins, and that aligning acquisition to high-intent behaviors lifted value per install by more than 6x for one client.

The Leadership Circles, meanwhile, tackled topics that rarely make it onto conference stages: career evolution frameworks with Atif Rafiq, loyalty through discovery-led experiences with Ian Dewar from Anthropologie and The North Face, gaming marketing insights with Jen Donahue of Deconstructor of Fun, and fandom marketing playbooks with Jonathan Yantz of M&C Saatchi, with fresh insights from AppsFlyer’s latest report with Sensor Tower and Saatchi: “Scoring Big: The Complete Marketer’s Guide to the World’s Top Soccer Event.” 

The Session Nobody Expected to Be Their Favorite

Lakshmi Rengarajan studies dating culture, an unusual choice for a mobile marketing event. 

But her research on how we form first impressions hit a nerve. We’ve replaced the language of realization and development with the language of consumerism, she argued: we swipe, filter, and optimize in dating and in professional relationships alike. 

Her interactive exercise had attendees pair up and share three names: someone who shaped them professionally, someone they grew to like over time, and a childhood crush. The room transformed. As Rengarajan put it: “They started to look different to me.”

Her closing question, “Who do you need to date?” landed as a challenge about professional relationships we underinvest in.

Four Takeaways Worth Keeping

• The marketer’s role is expanding. The practitioners pulling ahead are the ones who own recommendations end-to-end, speak the CFO’s language, and treat their career trajectory as something they design, not something that happens to them
  
• Your app isn’t a channel — it’s economic infrastructure. Brands are proving a 5–6x ROI when they measure total app value beyond direct revenue alone.
  
• Measurement is the new competitive advantage. 30% of campaigns are being significantly undervalued by last-touch, and the teams combining incrementality with modern attribution frameworks are making smarter allocation decisions.
  
• AI’s edge right now is eliminating the gap between knowing and doing. From natural language data queries replacing SQL, to a mobile game built and marketed to ROI-positive in 14 days, the biggest AI wins at MAMA were about accelerating your capabilities and collapsing the distance between insights and actions.

MAMA will be coming to other destinations around the world in 2026.

The post Where Do We Go From Here? At MAMA San Francisco 2026, Mobile & App Marketing Leaders Navigated What Comes Next appeared first on AppsFlyer.

TL;DR

• Web-to-app is not about opening the app, it’s about preserving customer intent as they move from mobile web into the app
• Web-to-app continuity is one of the most powerful levers for improving installs, onboarding, and in-app conversion.
• Most brands underestimate how much revenue they lose when web-to-app journeys quietly break.
• Because the leak happens in the handoff between web and app, most analytics setups never capture the loss
• Brands like AirAsia, Tata CLiQ, and Apartment List have documented measurable improvements after strengthening their web-to-app flows with AppsFlyer’s Deep Linking Suite.
• Fixing this hidden leak turns web-to-app from a fragile handoff into a predictable source of growth

You earned the click. And then you lost them

You’ve done the hard work. You ran the campaign, drove traffic to your mobile site, and earned the click. A real customer, with real intent, tapped “Open in App.” And then, quietly, you lost them.

Not because your product isn’t good. Not because your creative underperformed. But because somewhere in that split-second handoff between your mobile website and your app, the journey broke – and your customer had no idea why. Neither did you.

This is the web-to-app leak: one of the most common, most costly, and least visible conversion problems in mobile marketing today. It doesn’t trigger an error message. It doesn’t show up as a crash. It just looks like a drop, a bounce, a customer who “wasn’t ready.” But in most cases, they were ready – your routing infrastructure just wasn’t

What is web-to-app?

Web-to-app continuity is the ability to move customers from your mobile website into the right in-app experience without losing what they were trying to do. It preserves intent (what they clicked on and where they expect to land), context (what they were viewing or configuring), and what their state is (installed, logged in, returning) so the journey continues instead of restarting.

The “Intent Drop” pattern: Meet Emma 

Emma is browsing your mobile site.

She finds something she wants – a product, a flight, a hotel, an offer, a loan she’d like to take, a reward she wants to redeem.

She taps your “Open in App” banner (often a Smart Banner – a website banner that opens the app or sends customers to install it) because she knows the app experience is faster, easier, and more personalized..

She’s ready to continue.

This is the Intent Drop – when a customer clicks with intent, but the handoff into the app drops the context and resets the journey.

Instead, here’s what happens to customers like Emma every day:

• The app opens, but drops her on the homepage instead of the item she tapped on
• Or it sends her to the App Store even though she already has the app installed
• Or she hits a login page before seeing any relevant content
• Or the banner takes her to a generic mobile web page in a different tab
• Or the fallback behaves differently on iOS and Android
• Or she clicks, but nothing happens…

Emma wants to buy, but her journey has hit a dead end..

This scenario is not a bug or an edge case – it’s the predictable outcome of treating web-to-app only in a channel redirection lens instead of a seamless continuation of customer intent.

Multiply Emma by thousands of daily visitors, and you start to realize just how big the leak is. 

Why web-to-app matters more than most teams realize 

Some teams are comfortable keeping customers on mobile web, but the performance gap between web and app is simply too large to ignore.

Regardless of which industry you’re in: 

• Customers convert better in the app which is the native mobile device environment
• They complete authenticated actions more reliably
• They complete purchases and other key actions at higher rates
• They return more frequently
• They experience  more personalized experiences
• They interact more deeply with loyalty programs, wallets, saved preferences, and push notifications

Mobile web is where discovery happens. But the app is where value compounds.

And the step between the two, that seemingly small moment in time when a customer tries to move from mobile web to app, is one of the highest-intent points in the entire experience. But it’s also one of the easiest to break.

Emma didn’t abandon the journey. The journey abandoned her.

Internally, you test your Smart Banners and they work. But these tests usually happen under controlled conditions: one device, one browser, one app version, one login state, and one clean session.

But in the real world, customers arrive from many different entry points and contexts that change how web-to-app behaves and how it should be optimized:

• Search and social feeds (including in-app browsers)
• Paid ads and affiliate/influencer links
• Email, SMS, and push
• Saved tabs and returning visits days later
• Logged out sessions, expired sessions, or private browsing
• Older app versions and mixed OS versions
  

When anything breaks in that handoff, it rarely shows up as a clear error. The customer just lands somewhere generic, gives up, and disappears. Your web metrics still show a click, your app metrics show a drop, and the gap in between gets misread as “low intent” or “weak creative” – even though the customer was ready to convert.

This is the hidden leak.

What’s actually causing the leak?

At its core, the leak exists because web-to-app is treated as a channel-specific implementation instead of a shared intent-resolution layer.

The problem isn’t your campaigns.
It isn’t only your banner design.
It’s that the routing behind the tap isn’t built to preserve intent and context consistently.

It’s your routing infrastructure.

Web, app, CRM, product, growth, and engineering teams each own different parts of the journey. They often make different assumptions about:

• Where the customer should land
• How login occurs 
• What to show customers who have the app and to those who don’t
• How fallbacks should work
• How parameters should pass into the app
• How journeys are measured and optimized 

Every team creates its own version of “what should happen after the tap.”

The result is an inconsistent and fragile web-to-app experience that breaks for real customers in real conditions.

The real lesson: web-to-app is an intent problem, not a linking problem

Most teams approach web-to-app as a technical or channel problem.

They ask:

• Is the Smart Banner implemented correctly?
• Does the deep link open the app?
• Does the fallback send customers to the right store?
• Are they seeing an error?

Those questions matter – but they miss the point.

The real job of web-to-app is not to open the app.  It is to preserve customer intent across the transition.

When a customer taps “Open in App”, they are not asking to switch environments.
They are asking to continue the same action – with the same context, state, and expectation.

Web-to-app works when three conditions are  consistently met:

• Intent is preserved: the customer sees exactly what they tapped on in mobile web.
• Context is resolved: login state, install state, device, and platform differences are handled automatically.
• The transition is invisible: the customer never feels the “jump” between web and app.

Most web-to-app journeys break because teams solve only the first part – opening the app – and leave intent and context to chance.

The impact of fixing your web-to-app leak

Just as importantly, fixing your web-to-app leak builds trust in your data – because teams can finally see what happens after the tap and optimize the journey with confidence

These companies boosted their web-to-app continuity using AppsFlyer’s Deep Linking Suite and our OneLink Technology and saw measurable improvements

AirAsia: Strengthening Smart Banners and web-to-app routing

Industry: Travel

Problem: Drop-off between mobile web browsing and app installs

Intervention: OneLink-powered Smart Banners with unified routing

Measured outcome: 5% increase in total installs driven by web-to-app paths

Tata CLiQ: Improving conversions by making web-to-app consistent

Industry: Ecommerce
Problem: Mobile web customers restarting journeys after moving into the app
Intervention: OneLink-enabled Smart Banners preserving context across the handoff
Measured outcome: Increased conversions after improving web-to-app consistency

Apartment List: Increasing early engagement by fixing entry-point continuity

Industry: Real estate marketplace

Problem: High-value mobile web customers failing to continue into onboarding in the app

Intervention: OneLink deep linking to stabilize entry-point continuity during onboarding redesign

Measured outcome: 2× increase in Day 0 logins, 10× increase in transferring highest-value web users to the app, 15% decrease in CPI, and 30% increase in user LTV

In all three cases, the common shift wasn’t more campaigns or new creatives, it was treating the web-to-app moment as a continuation of intent, not a redirect

The turning point: one routing system, one logic, one customer experience

Top-performing teams across retail, travel, finance, marketplaces, and lifestyle apps all do the same thing:

They unify routing with one engine, one set of rules, and one source of truth.

In practice, that means every entry point – regardless of where it comes from – follows the same logic:

Whether a customer is redirected from mobile web, a Smart Banner, an in-page CTA, an in-app browser, a referral flow, or even a help center article, the same routing engine (like the one behind AppsFlyer’s Deep Linking Suite) determines:

• Should the app open or the store?
• Which screen should the customer land on?
• What context should persist?
• Should login interrupt or not?
• What analytics need to be attached?

This is infrastructure, not an add-on.

And once routing is unified:

• CRM performs better
• Product teams see fewer continuity bugs
• Developers stop handling routing edge cases
• Customers stop abandoning before they ever reach the app

It’s the difference between “Emma bounced” and “Emma continued and converted.”

Key takeaways

• Most web-to-app journeys silently break, even when internal tests appear fine
• Web-to-app failures aren’t a banner problem – they’re an intent-preservation problem
• The leak happens in the handoff – before the app measures anything
• Brands like AirAsia, Tata CLiQ, and Apartment List have documented meaningful improvements after fixing routing consistency
• Unified routing transforms web-to-app from a fragile step into a reliable growth engine
• Teams that win are the ones that treat routing as customer experience infrastructure

The teams that get this right don’t just see better conversion numbers – they stop losing customers they already earned. If web-to-app is a meaningful entry point for your owned media journeys, see how AppsFlyer’s Deep Linking Suite helps teams preserve intent and context across web, email, QR, and deferred install flows – without requiring a paid attribution package. Explore the Deep Linking Suite or talk to an expert about your web-to-app leak

The post Your customers are trying to convert but your web-to-app journey is stopping them appeared first on AppsFlyer.