John Koetsier is a journalist and analyst. He's a senior contributor at Forbes and hosts our Growth Masterminds podcast as well as the TechFirst podcast. At Singular, he serves as VP, Insights.
It’s probably a controversial statement, but we’re entering a golden age of marketing measurement right now. Not an easy age, necessarily, but a golden age.
I’m not sure you had it on your Bingo card for 2025. I didn’t. And it’s something probably no one thought would happen 4 years ago.
But we haven’t really clearly laid it out until Singular CEO Gadi Eliashiv talked about this in his recent keynote at ChinaJoy.
Masses of data, from multiple sources, tied together with smart AI-driven analytics.
Check it out here:
Thanos and the golden age of marketing measurement
Of course, we have to start with a point in time, 4 years ago.
It’s hard to talk about marketing measurement without revisiting the moment that Apple’s privacy moves changed everything. ATT and iOS 14.5 dropped like Thanos snapping his infinity-stone-enhanced fingers … and half the marketing measurement universe kinda just disappeared.
On iOS it felt like 90% of our user‑level signals disappeared overnight.
(Mostly because that’s exactly what happened.)
Naturally, there was a lot of doom and gloom.
Was performance marketing over?
Would cohort analysis, ad optimization, and ROI modeling go away
Was reliable attribution over?
And in the years that followed, it looked like Android would follow a similar pattern with Privacy Sandbox. (Which has been very muted lately!)
But now we have our answer.
Yes, the rules changed. Yes, tactics evolved. But also, the ecosystem adapted. Platforms opened new pipes, analytics got smarter, and marketers learned to stitch together many signals instead of leaning on just one.
And the end result of Thanos’ snap is a new golden age of marketing measurement.
We went from a data drought to signal abundance
What looked like scarcity — and originally was scarcity — has become a kind of abundance.
But only if you know where to look. If you can catch all the signals. And if you can intelligently combine them:
Open internet & device signals On iOS, you still have SKAN, IDFV, and meaningful first‑party data. On Android, GAID, referrer, and first‑party data continue to matter. And don’t forget the humble IP address and user agent when used responsibly.
The big platforms’ data Meta, Google, TikTok, Snap, X and others now share more aggregated and modeled performance signals than ever before. Meta’s AMM is back. Google’s ICM is here. Deterministic joins might be rarer, but probabilistic and modeled insights are richer.
From the big platform side, it looks like this:
Platform
Modeling
Data sharing with MMPs
Unique aspects
Meta
Yes
Touchpoints, claimed installs via AEM, Advanced AEM, AMM
From the open internet, ad network, and device signals side, it looks like this:
iOS
Android
IDFA
GAID
IDFV
Referrer
SKAN
Ad network data
Ad network data
IP address
IP address
First-party data (in-app, requested account data including for messaging, etc)
First-party data (in-app, requested account data including for messaging, etc)
The result: more sources, more coverage, and more opportunities to triangulate truth.
This all adds up to a sturdier foundation for growth
Thanks to that broader signal set, growth teams can now assemble a durable measurement base:
They can unify data Use clues from platforms, partners, and your own first-party data into a single reliable attribution model … Unified Measurement.
They can answer core questions Get CPI, CAC, LTV, and ROAS dependably across both Android and iOS.
They can measure at the right level of granularity Drill down when you want to, and roll up when you need the big picture.
They can layer on prediction and modeling Build high‑resolution views where user‑level data isn’t available.
With that foundation, analysis is less a random archaeology dig and starts feeling like product engineering again.
And arguably, it’s a more contextualized, nuanced, balanced, and reliable view of attribution and marketing measurement than the previous model, which essentially relied very heavily on just 1 signal: IDFA on iOS and GAID on Android.
Golden age of marketing measurement, meet AI
All of that foundation is what our increasingly capable AI models can now feed on and generate ever-more-powerful insights.
This isn’t a bolt-on: it’s now embedded in how marketers ideate, operate, and optimize. Three practical examples:
Creative intelligence in Creative IQ
In a world where bidding levers are increasingly automated, creative is the one you can still pull yourself.
(For now: creative will increasingly get automated at the platform level.)
But whether you make it or you offload that to generative AI, you still need to know how it performs.
The Singular approach, Eliashiv says, pulls every asset from every channel into one system, then uses AI to understand them: audio, visuals, text, colors, characters, elements, even language. Creative IQ maps those attributes to performance so you can see not just what wins, but why.
“Now we can use AI to understand why this creative is working so well. Is it the language? The characters? The tone, the colors, the soundtrack?” – Gadi Eliashiv, Singular CEO
The upshot: faster iteration cycles, smarter briefs, and a shared creative language across UA, design, and leadership. Because now you don’t just know which individual creatives are winning. You also know why. And that knowledge allows you to make more like it.
AI that helps you set up measurement: Singular Copilot
Ad networks and platforms are powerful. With that power comes complexity. And measurement platforms like Singular have to live in that complexity and make it manageable.
How?
With AI help.
Configuring partners, tracking links, postbacks, re‑engagement windows, cross‑device settings … there’s a lot both to know and to do as a growth marketing expert. So we’re embedding a context‑aware assistant directly inside the product: Singular Copilot.
You can ask, “How do I set up a tracking link” and it won’t just answer. It will take you to the right screen, highlight the right fields, and guide you through every step of the setup. Essentially, it’s a smart agent that understands both your question and what is happening on your screen.
This is a bridge to a future where less is configured manually and more is orchestrated by intelligent systems.
Optimization engines that never sleep
Our new ELT product, Extract, moves raw marketing and store data and much more into your warehouse quickly, easily, efficiently.
Some of the largest game publishers use it to query the App Store and Google Play every 15 minutes, check installs, analytics, and rankings, and automatically adjust their spend and bids. Paired with modeled conversion signals from ad partners, these loops create near real‑time allocation engines for UA and re‑engagement.
“They’ve started automating their user acquisition and re‑engagement work … choosing where to allocate budget, which creative to generate, writing their own text for creatives, and choosing bids.” – Gadi Eliashiv, Singular CEO
The common denominator across all three of these? A strong, complete, and accurate data foundation.
AI multiplies your impact only when the inputs are consistent, rich, and trustworthy.
What this means for marketers right now
Measurement is back. The golden age of measurement is here.
But only for those who are ready for it, and are collecting the multiple data points you need to unify attribution from all the different signals.
So what should you do?
Audit your signals Inventory what you have (SKAN, GAID/referrer, first‑party, partner‑provided modeled events) and what you’re missing. Meta’s AMM, for example, is only available if you agree to it explicitly, and you can only get it via an approved MMP like Singular. Close the gaps.
Instrument creative learning Invest in smart tools like Creative IQ to boost your creative results. AI will help you tag systematically, manage data governance, extract insights, test hypotheses, and feed learnings into generation.
Embed AI into operations Start with Singular’s integrations with ChatGPT and Claude. Use AI assistants to understand what’s happening, and then reinforce your actions. There are many more tools coming …
A final word: the golden age of measurement?
The last few years have reshaped measurement.
What began as turbulence now feels kind of like a tailwind … if you bring the right data, analytics, and AI to the table. And have some fun.
As Gadi said on stage:
“We are in a golden era right now and I think it’s exciting.
GPT-5 is here and that’s great news for marketers, especially if you’re using our just-released MCP integration with ChatGPT. In fact, GPT-5 might just be a game changer for marketers who aren’t 100% sold on the AI Kool-Aid yet. (Even better, because 5 is much more processor-intensive and might hit your rate limits soon, OpenAI is bringing back 4.0 in case you prefer it.)
But 5 should be better for most things:
OpenAI CEO Sam Altman says it’s like the change from the early iPhones to the retina display. He also said GPT-5 is a “PhD-level expert.”
That might be a bit hyped, but at minimum, 5 is incrementally better than GPT-4o, which was already pretty ridiculously good. Note: you may have to manually select Thinking Mode, which ensures the full power of GPT-5 is unleashed on your query, to get the best results.
We’ll get into all the details below, but briefly here are a few places where I see GPT-5 could be super helpful for UA and performance marketers. The new GPT-5 …
Makes higher-quality content
Generates much better computer code (yeah, this matters to marketers … keep reading!)
Improves data analysis for large datasets
Customizes customer-facing comms with new “personalities”
Improves decision-making
Boosts safety and compliance on sensitive topics
Reduces costs over previous pricing models
Automates multi-step workflows
Marketers are super-hungry for this. In fact, our industry-first MCP integrations with Claude and ChatGPT have seen some of the fastest uptake of any Singular product, ever.
“This is a watershed moment for so many of our customers,” says Singular CTO & co-founder Eran Friedman. “Now any marketer can talk to their data exactly like they talk to a teammate.”
There’s so much here I’m going to present this in numbered points for easier digestibility. This is a major new release for OpenAI and ChatGPT, and it’s massive.
Specifically for marketers, here’s what’s new in ChatGPT with GPT-5:
GPT-5 is faster Ever waited in ChatGPT for the text to slowly … crawl … down … the … window? Now Altman says the model is so fast that he found himself wondering whether it had thought enough before answering (but it had).
It’s more accurate Everyone hates LLM hallucinations. We’ve actually invested significantly in reducing their likelihood in our AI integrations. According to OpenAI, the system delivers ~45% fewer factual errors than GPT‑4o in real‑world scenarios and up to ~80% fewer when its thinking mode is triggered.
You can make software on demand In a demo video, an early beta tester revealed he had created an app where a camera mapped his hands’ locations in real-time and allowed him to control on-screen artifacts … all from a prompt. Just think what you can do for custom little apps that do parts of your work for you. Another test showed GPT-5 making a working French-language learning site within seconds. What websites or landing pages do you need made … this is the ultimate no-code development, and will unlock so many possibilities for busy marketing teams.
GPT-5 now has a single unified model You don’t have to pick a model anymore. ChatGPT will get your query and automatically assign more thinking and processing resources where needed.
The system is more honest (and humble) GPT-5 will now admit more readily … “I don’t know.” This is huge, and will build more trust in ChatGPT when it does actually answer.
It’s more powerful and can take more inputs GPT‑5 dramatically expands the context window to 256,000 tokens, roughly equivalent to a 600‑800 page book. So you can feed it huge prompts … think tons of data … and it will crunch through it all and supply answers.
ChatGPT can now connect to Gmail, Google Calendar, and more Drowning in email? Need help answering customer requests? ChatGPT can act more as an agent for you with GPT-5. Maybe you get support requests via email … now ChatGPT can connect, read, analyze, and report on common issues. And maybe even learn how to respond appropriately.
Better charts (!!!) I didn’t see this in any of the claims that OpenAI made, but we’ve been playing with GPT-5 behind the scenes here as we add it to our ChatGPT integration, and the charts are much better. This has been a place where Anthropic’s Claude has had a huge advantage over OpenAI.
This is all impressive and great, but what does it mean for marketers?
GPT-5 for marketers: what can it do?
OK, let’s get to the important stuff for marketers: what can GPT-5 do for UA pros and performance marketers?
In a word: lots.
“Speed wins in marketing,” Gadi Eliashiv, CEO and co-founder of Singular, said recently about our ChatGPT integration. “By connecting our analysis‑ready data to world‑class LLMs via MCP, we’re giving marketers a new level of intelligence and agility.”
Given that GPT-5 is faster than previous models, that’s a good thing.
A caveat: this has just been announced, so we have to wait to see how it works in the real world. But OpenAI has proved they’re pretty freakishly good at AI, so I’m willing to provisionally trust them. (And you can always roll back to 4 as well, if you prefer.)
But here’s what should be better in 5:
High-quality content generation GPT‑5’s improved reasoning and writing abilities will provide marketers with a powerful co‑creator. It’ll make copy more coherent, context‑aware, and aligned with brand voice. So creating text for A/B testing emails or headlines will be easier and better. And writing blog posts should improve too. (But … yeah … keep a human tone, huh?)
High-quality code generation I said it above, but if you want to prototype landing pages, build microsites, or create interactive tools straight from natural‑language descriptions … GPT-5 is gonna be better. (Now we need tools like WordPress and others to offer easy ways to integrate that kind of AI-generated code.)
Better data analysis This is near and dear to my heart: I just finished up Singular’s next Quarterly Trends Report. GPT-5 can ingest larger datasets and make sense of them. Think … piecing together customer journeys, understanding and predicting behavioural data, or merging multi‑channel campaign logs.
Better customer-facing comms You can now tell ChatGPT to adopt a custom personality: cynic, robot, listener, or nerd. You can bet more will be coming soon. That means you can build brand personality into just about anything where you want AI help in conversing with customers or users.
More confidence in AI-assisted processes Up to 80% fewer errors is a big deal, and that’s going to translate into greater trust over time. My daughter recently told me that ChatGPT got basic things wrong about the F1 summer schedule and racing calendar, which diminished her trust in AI across the board. As we experience GPT-5 and see if it’s less error-prone, we’re going to trust it in all kinds of marketing processes.
Higher-level work with less handholding A junior employee needs lots of handholding. Senior ones, less so. GPT-5 should be a more senior assistant that can do bigger jobs with less human intervention. GPT‑5 has improved multi‑step reasoning and agentic capabilities, which could enable things like analysing campaign performance, recommending budget reallocations, generating creative assets, and maybe — in the future — even implementing those changes.
Using AI for marketing … it’s time
If you’ve resisted so far, it’s time to start using AI for marketing. And GPT-5 is a good place to start.
Start by trying our integration with either Claude or ChatGPT: whichever you prefer. You can find the instructions right in our Help Center. If you need help, there’s an AI-powered Singular Assistant to walk you through it, step by step:
I promise you’ll find it interesting, useful, and maybe even mind-blowing. Just like, perhaps our latest AI-powered creative optimization product, Creative IQ. There’s more coming there too.
Food delivery apps have come a long way, baby. Just 7 years ago in 2018 they took in $90 billion in revenue. But by last year, that exploded to $289 billion, and it’s still growing. In fact, the food delivery sector is estimated to hit over $500 billion by 2030.
A third of this market is currently North America, but APAC is the fastest growing, while LATAM and EMEA continue to see annual double-digit growth.
That’s … impressive.
So who’s going to win?
Which apps and companies are best-positioned to grab the most share of this massive and quickly growing market? Covid set the stage for the current battle, when food delivery downloads increased by 20% year-over-year. Now, as the competition continues to heat up, the emphasis is shifting to q-commerce: quick commerce, with promised delivery times as low as 10 minutes.
In this post I’m going to look at the top food delivery apps in 4 key global markets:
APAC
EMEA
LATAM
NA (North America)
I’m looking at both app downloads and time spent in app, and I’m weighing time spent in app about twice as much as downloads, just because I think usage and engagement — and the revenue which is much harder to directly measure — is more important than simple installs.
Key takeaways
Food delivery is growing fast: global revenue hit $289B in 2024 and is projected to surpass $500B by 2030
APAC is the fastest-growing region; North America still leads in overall revenue
The most-used apps globally include Grab, Zomato, Swiggy, DoorDash, Uber Eats, and Just Eat
Rankings are based on both app downloads and user engagement, with heavier weighting on time spent in-app
Regional leaders vary widely across APAC, EMEA, LATAM, and North America
Restaurant brands like McDonald’s, Pizza Hut, and Taco Bell show strong app engagement, especially in North America and LATAM
The next wave is q-commerce: ultra-fast delivery is setting the pace for competition
Drone delivery is emerging as a critical innovation for speed, cost savings, and scalability
First off: global top 10
Below, I’ll look at each of the individual markets. But if we look at the global top 10 food delivery apps, here’s what the model shows:
Grab Dominates Southeast Asia
Zomato Leads in India
Swiggy Leads in India
Foodpanda Strong is Southeast Asia and parts of Europe
iFood Leads in Brazil
DoorDash Strongest in the USA
Deliveroo Strong in the UK, Ireland, and parts of mainland Europe
Just Eat/Takeaway.com Leads in Western Europe
Uber Eats Strongest in the USA
PedidosYa Leads in Latin America
Grab has huge engagement and usage in Southeast Asia across multiple verticals, including but not limited to food delivery and groceries, while India-focused apps like Zomato and Swiggy rank high thanks to their own deep engagement and frequent usage. In the United States, we see more DoorDash and Uber Eats, while Deliveroo and Just Eat dominate in the EU.
PedidosYa is strong in Argentina, Uruguay, and Chile to lead in LATAM, along with iFood, which dominates in the huge economy of Brazil. And while Rappi doesn’t make this list, it’s also a significant player here.
OK: now let’s look at each region …
Top food delivery apps: APAC
Here are the top 20 food delivery apps for APAC right now.
What we see is that regional leaders like Swiggy, Zomato, and Foodpanda stand out with very high engagement in specific markets.
Grab
Zomato
Swiggy
Zepto
foodpanda
Pizza Hut India
EatSure QuickiES Food Delivery
LINE MAN: Food Delivery & more
DoorDash
magicpin
Coupang Eats
SNACC: 10-Min Food Delivery
Yango
EatClub
Wolt Delivery
PedidosYa
McDonald’s India Food Delivery
ShopeeFood
Milkbasket
Deliveroo
Swiggy and Zomato are big in India, a large and fast-growing market. Foodpanda and Grab are stronger in Southeast Asia. We see high engagement levels, lots of repeat usage, and strong loyalty in these markets.
Only 2 restaurant brand apps, Pizza Hut and McDonald’s, make the list. All the others are pure-play delivery apps, and most of them deliver food plus other goods and services.
Interestingly, some of these brands are so big and so popular that their delivery personnel or service apps also rank in the top 20, and had to be filtered out by hand.
Here are the top 20 food delivery apps in EMEA right now.
Again, we see strength in specific markets, with Deliveroo leading in the UK and Ireland, Just Eat stronger in Spain, Italy, Denmark, and the UK, and Wolt focused on the Nordic and eastern European countries.
Yandex Go
Zomato
Wolt Delivery
Swiggy
foodpanda
PedidosYa
Glovo
DoorDash
Grab
Deliveroo
Yango
Zepto: 10 Min Grocery Delivery
HelloFresh:
Keeta
Just Eat
EatSure QuickiES
DiDi Food
Domino’s Pizza Delivery
Coupang Eats
Rappi
Yandex Go is surprisingly strong in EMEA, with Zomato and Swiggy showing up here as well as APAC. Glovo is strong as well, and HelloFresh makes an appearance as a meal prep (and delivery) app.
For restaurant brands, we see Domino’s … and that’s it.
Top food delivery apps: LATAM
Here are the top 20 food delivery apps in LATAM right now.
iFood
PedidosYa
McDonald’s Offers and Delivery
Zomato
Swiggy
Yango
foodpanda
Zé Delivery de Bebidas
Rappi
DiDi Food
DoorDash
Zepto: 10 Min Grocery Delivery
Wolt Delivery
EatSure QuickiES
Coupang Eats
SNACC: 10-Min Food Delivery
Deliveroo
HelloFresh
LINE MAN
aiqfome: delivery de tudo
iFood is a clear leader with high engagement scores, but PedidosYa is doing very well also. McDonald’s makes a surprisingly strong showing in LATAM: the highest any individual restaurant brand shows up anywhere, globally.
Top food delivery apps: North America
Dominated by the United States, the North American market is about 27% of global food delivery revenues right now. Here, DoorDash, Uber Eats, and Instacart dominate both downloads and engagement.
Top 20 food delivery apps in North America:
DoorDash
Instacart
Pizza Hut
Grubhuby
Taco Bell
Papa Johns Pizza
Postmates
Skip
Gopuff
Grab
Factor
HelloFresh
7NOW
Slice
Shipt
foodpanda
Zomatog
McDonald’s
Sheetz Food
PedidosYa
Factor joins HelloFresh here as food prep delivery services, but there’s an even more significant difference between North America — and particularly the United States — and the rest of the world.
And that’s the number of individual restaurant brands showing up:
Pizza Hut
Taco Bell
Papa Johns
McDonalds
Another difference: a food-specific delivery app in Slice (for pizza).
Next step in the delivery business: drones
Food delivery apps are growing fast.
“Takeout used to be a convenience, now it’s a culture,” says Food and Wine magazine.
In the US, consumers average 4.6 food app orders per month, with younger users even more hard-core users, at about 5.1 orders per month. A staggering 75% of restaurant traffic now involves takeout or delivery, and speed is one of the most important factors.
“94% of all consumers say speed is critical,” says the National Restaurant Association.
That means 10 minute delivery might soon be too slow, if you can believe it.
The only solution here will be drones. That’s super-fast, becoming more and more available as regulations and laws change all over the world, and costs 90% less than driving big steel and glass cars or trucks out to customers’ homes.
It’s coming soon: Uber Eats has been testing it for years.
At that point, the 75% of restaurant traffic stat might get a lot closer to 100%.
How leading food delivery brands optimize for growth
As food delivery apps continue to scale and evolve, it’s clear that user experience, engagement, and cross-platform performance are becoming just as important as download volume.
Want to optimize every touchpoint, reduce acquisition costs, and serve up better results for your food or drink brand? Explore the Singular choice for Food & Drink businesses and learn how top marketers are using cross-device attribution, customizable links, and data-driven insights to build loyalty and drive growth.
San Francisco, CA – August 5, 2025 – Singular, a leading analytics, measurement, and data management partner for mobile app marketers, today announced it is the first Mobile Measurement Partner (MMP) to go live with a Model Context Protocol (MCP) integration with ChatGPT. This ChatGPT integration follows Singular’s earlier first‑to‑market launch with Anthropic’s Claude. Additional support for other MCP-enabled LLMs is coming soon.
“Speed wins in marketing,” said Gadi Eliashiv, CEO and co-founder of Singular. “By connecting our analysis‑ready data to world‑class LLMs via MCP, we’re giving marketers a new level of intelligence and agility.”
Because ChatGPT has a free tier and Singular also has a free tier of service, this powerful service is now available at zero cost even to those mobile app and game developers with the tiniest of budgets.
And it’s unlocking an unprecedented richness of data for all mobile app and game makers.
What this ChatGPT MCP integration enables
An MCP integration is a deep integration between the intelligence of major LLM-based AI engines and the super-rich marketing and advertising data in Singular.
Which ad partners return the highest ROI for least cost
Which ads are working
And much more …
App developers can also get instantly generated visual charts and tables right in their conversations. Non-technical staff don’t have to navigate dashboards or write SQL.
Ultimately, this unlocks growth for game and app developers who don’t have deep data or marketing technology knowledge. Now ChatGPT (or Claude) can help them find, understand, and apply data-driven insights for new marketing campaigns, ad partners, ad creatives, and marketing offers.
How marketers are already using Singular’s MCP integrations
Global growth teams have made Singular’s MCP integration with Claude one of the company’s fastest-growing product features and are rapidly adopting it for daily reporting, and we expect the same for ChatGPT.
The most common questions fall into six buckets:
Campaign performance and return on ad spend
Trend analysis and time-based reporting
Revenue and spend comparison
Creative performance and A/B testing
Install and conversion tracking
Geo and platform-based segmentation
Early adopters report dramatic speed gains in getting data, but also unexpected access to gamechanging insights providing data points they didn’t previously consider important.
Limiting hallucinations
LLMs are subject to occasional hallucinations, which might be OK for casual searches but are definitely bad for marketers making investment decisions. So Singular has worked hard to use strict schemas, break every question into tiny, individual tasks, force LLMs to admit ignorance, keep prompts super lean, and provide full observability and logging so that hallucinations don’t happen or are quickly caught.
“We design each model call to do just 1 small job: nothing more,” says Eran Friedman, Singular CTO. “Clear, single-purpose calls prevent prompt injection, keeping outputs predictable.”
Within the MCP integration, each LLM’s output is tightly scoped to numerical data and visual analytics returned from Singular, minimizing faulty responses. And making both Claude and ChatGPT helpful, reliable partners in growth.
First of its kind
MCP is a very recent integration technology introduced by Anthropic in November of 2024. It defines how LLMs can securely query structured data sources in real time. The MCP protocol keeps data access safe, secure, and auditable, and allows Singular to enforce user-level permissions at query runtime.
Think of it like a USB-C port for AI.
It’s only recently enabled deep integrations with rich datasets like our clients’ data in Singular, and the integrations for ChatGPT (OpenAI), Microsoft, and Google are brand-new.
While other marketing analytics companies are still testing their integrations in limited betas, Singular’s is openly available to all clients.
About Singular
Singular is the only end-to-end marketing analytics platform that uncovers true ROI by unifying cost, performance, and engagement data across every marketing channel. Trusted by growth leaders at LinkedIn, Rovio, EA, and Nike, Singular helps teams optimize smarter, scale faster, and eliminate wasted spend.
Tilting Point just raised a new $150 million fund for user acquisition. They’re the money behind SpongeBob, Star Trek, Warhammer, and Zombieland games. And they scaled monthly ad spend on Match 3D from $500K to $11M (!!!) helping them explode revenue 30X. In other words, these guys know what they’re talking about … and they might just have some UA lessons for the rest of us.
(Plus, who knows, they might even have some cash to help scale your game or app.)
I recently chatted with Asi Burak, Tilting Point’s Chief Business Officer, about the fund. But we also talked about improving monetization and retention in games, how to kickstart growth, how to scale growth, his best advice for indie developers, and the future of game monetization.
Sure, growth starts with a good product. But Burak can’t count how many developers he’s seen that want to kick off UA spend after the next version, or the next big feature, or the newest mechanic …
UA lesson 1: start.
“If you can put money and bring users and bring players and quality players, do it,” Burak says. “Don’t wait for permission … don’t wait for the next version that’s going to be so much better. If you can scale, scale.”
There is no mythical perfect product. Don’t delay growth. If your metrics show retention and monetization are strong, lean in.
That said, scaling requires discipline.
And a few doctor’s visits.
“I always tell developers … check the data, check the data, check the data. Test, test, test. Find the model and go get the doctor’s opinion … when they come to us, we’ll give them a very objective look.”
In other words, talk to smart people who have been there and done that. They’re the doctors who can diagnose growth problems or limiting factors.
UA lesson 2: but don’t start too early
Growth is pretty scientific these days, sure, but it’s also an art. And part of the art is in knowing when the hit the gas.
UA lesson 2 is all about remembering to focus on retention and monetization before kicking off big UA spending.
One of Tilting Point’s best-known success stories is Match 3D, which scaled from $500,000/month in UA spend to $11 million/month and achieved astonishing 30X revenue growth.
The reason? Exceptional retention and a unique genre.
“You saw a game that was super, super engaging, super sticky,” Burak says. “Day 90 retention was 5% … so we knew that we could start putting the gas on the pedal.”
If your product doesn’t retain well or monetize effectively, pouring money into UA just accelerates churn.
(Oh, and burns cash.)
UA lesson 3: payback periods are getting longer
Everyone wants the marketing spend back ASAP, but today you need to have a little more patience.
A few years ago, many mobile games recouped their ad spend in 30 to 90 days. UA lesson 3 is that now, payback often takes a year or more. And that’s a long time.
“Unfortunately … what happened to UA post‑ATT and some other things in the market … it became tougher,” Burak says. “We see objectively that paybacks extended … one year and a half, two years … and it’s not that they want to go beyond one year. They don’t have a choice.”
This means studios need a bigger financial cushion or external funding — like Tilting Point’s — to scale safely. Unless you have unlimited cash on hand.
UA lesson 4: don’t waste equity
Especially with payback periods getting longer, it’s tempting to sell equity for venture financing to fund user acquisition and growth.
Think twice.
UA spend is critical, but it’s not the best use of scarce equity or VC funding, says Burak.
“I want you to take your VC money … and help you to invest that money in your product, in the growth of your team, in your IP,” he says. “And take my money for the low‑quality investment [of funding user acquisition] … because it’s money that you recycle all the time.”
Marketing is essential, but it’s an ongoing cost. Investing equity into core product and long-term value drivers is smarter.
UA lesson 5: learn from others
UA lesson 5 might seem obvious, but we’ve all seen founders flush with cash burn it on things that didn’t matter.
Talk to people who are experts. Who have scaled mobile companies. Who know how to grow.
Tilting Point doesn’t just provide funding. They bring experience managing campaigns for over 80 games, which gives them a wide-angle view of UA and monetization trends:
“Because we work on so many … we can say, you know, for your genre, that channel might be something you should look at,” Burak says. “Or rewarded platforms have strong short‑term performance but long‑term retention challenges.”
That cross-genre expertise can identify new channels or strategies you wouldn’t see if you only focus on one title.
UA lesson 6: experiment with new revenue models
Most games are doing hybrid monetization now:
IAPs
Subscriptions
Ads
That’s not all there is, or at least not all there will be, says Tilting Point. Mobile gaming revenue is evolving beyond in-app purchases (IAPs) and ads. Burak sees growth in hybrid models, season passes, real-money rewards, and even light Web3 or e-commerce integrations.
(Could web3 be back? Maybe only in very specific use cases: check the whole chat for more.)
“I think we’ll see more diversity in revenue sources … subscription, the season pass … even hybrid things like earned cash elements inside a regular game,” he says. “You add a layer of revenue that nobody else has.”
For marketers and product teams, that means more ways to monetize, and more data points to optimize campaigns. Importantly, it means more revenue than your competitors, which means more money for UA, and the ability to bid higher for premium revenue.
All of those are massive competitive advantages.
Oh and … yeah … that $150 million funding round
We didn’t just talk about UA lessons. We also talked money, because one of Tilting Point’s biggest differentiators is non‑dilutive funding. That means developers get growth capital without giving up ownership of their company.
“Developers were shocked, you know, they’re like, wait, wait, wait. So I’m not giving my equity for this? You’re not going to take my company away or something?” Burak says. “We want to give them an offer, a proposal, a model that doesn’t risk it at all. And in fact, in many cases we don’t even touch the corporate level … we collect revenue directly. Then we don’t even need to do any security agreement on the corporate level. So it’s not even getting to be a loan.”
Instead, Tilting Point’s funding focuses on user acquisition and performance marketing: investing millions of dollars a month into acquiring new players for promising games.
That’s worth checking out, and it’s the focus of this $150 million funding round.
Best-case scenario
It’s super rare, but the best case scenario for a new game or app is that you a) invent a new genre, and b) it turns out to be a massive hit.
(Yeah, crazy hard and unlikely on both counts, but we can dream, right?)
That’s the case with Match 3D:
“You see a guy that invented a genre … he did it with five guys in a room,” Burak says. “He later sold it for $200 million because his EBITDA was so crazy just for the fact he had such a lean team and made it so quickly.”
If you’re super-successful like that, Tilting Point doesn’t put a hard cap on the funding. (Especially because the payback period for this particular game was really fast.)
Much more in the full podcast!
You know you need to listen to or watch the podcast to get all the most juicy learnings, right?
When you look at the state of modern marketing analytics in 2025, you need to take a few deep breaths. It’s growing massively — as usual — but where it’s growing is most impressive: AI, AI, AI. But the big question is: what do marketers need most?
The answer just might be: even more AI.
Key Takeaways
Massive martech growth: 14,108 tools in 2025, up 9,304% since 2011
AI dominates marketing analytics: 3,068 new AI-native tools launched last year alone
Top AI use cases in marketing: Content ideation, personalization, creative testing, and campaign analysis
Marketers’ biggest needs: Unified data infrastructure, AI-enhanced attribution, and automated insights
Rise of AI agents: Custom-built tools helping marketers create, analyze, and optimize faster
Performance marketing leads AI adoption: Majority of UA pros already using AI daily
Singular’s innovations: Creative IQ and “Talk to Your Data” unlock instant, AI-powered analytics
The golden age of measurement: Rich, multi-source data + AI = deeper, more accurate attribution
Let’s dig in
Some of the big numbers defining the marketing analytics landscape, right off the bat:
14,108: total number of martech tools now
27.8%: growth rate last year (!!!)
9,304%: growth since 2011
3,068: number of AI-native martech products introduced last year
I am a big fan of Scott Brinker and his amazing, awesome, even frightening marketing technology landscape. Starting out in 2011 with just 150 logos, his martech landscape has grown massively.
Here’s the current state in 2025, with the data, marketing analytics, and attribution section highlighted by moi:
The current AI focus is no accident. It’s a trend caused by the collision of 3 main drivers:
The super-rapid progress of AI since ChatGPT launched in late 2022
The massive challenges marketers have in making sense of gigabytes of data
The ever-increasing workloads we’re all under
What that means is that trend is not slowing down. In fact, Brinker says both that AI is over-hyped (sure) and massively disruptive (absolutely). And, like many technologies, we’re tending to overestimate the short-term impact while at the same time underestimating its long-term impact.
That hype — and that disruption — is something that Singular is making a massive contribution to. We just recently released several significant AI-powered tools, and there’s much more coming soon:
So what is the state of marketing analytics in 2025? And what do marketers need most?
Marketing analytics landscape 2025
Martech and marketing analytics have both grown massively and astonishingly over the past decade.
Shockingly fast to shockingly large numbers, in fact. If we just look at Brinker’s martech landscape, we see a 9,304% increase in martech companies from 2011 to 2024 and beyond:
Importantly, 77% of the new tools introduced last year in 2024 were AI-focused: 3,068 new products for the marketing tech ecosystem that were AI-native.
If the current trends continue, we’ll see something like 17,000 total martech tools by the end of 2025, and perhaps another 3,500 AI-native tools built or released just this year alone. (Of course, with numbers like these, we’re going to see some consolidation, and some — maybe most — will fail.)
Where marketers are using AI in marketing analytics solutions
Right now, we’re mostly seeing the wider marketing community use AI for content-related needs, and many of the solutions are in that space:
69%: Content ideation
62%: Copy production
53%: Transcription, notes, summaries
49%: Content optimization and testing
45%: Personalization
I’d argue that the user acquisition and performance marketing slices of marketing professionals are probably more advanced in their AI use than the average marketer.
When we recently asked user acquisition pros about their biggest challenges in terms of creative, 34% were having challenges just knowing what’s working. A fifth were fighting creative fatigue, and maybe losing. Another quarter needed help generating new creative ideas.
AI can help with all of that.
Then, when we asked the same user acquisition experts about their use of AI, the results tracked to the problems:
27.2% are using AI for concept creation. 20.6% are doing visual generation via AI. Only .5% of these high-performance marketers were not using AI at all …
But even this use of AI is quickly expanding to much more challenging projects and processes. Every day, marketing experts have to set up campaigns, create tracking links, manage marketing data flows, set up attribution, understand analytics, and summarize campaign results. Look to Singular for help with all of it, very soon.
And these high-level users are expanding even more to AI agents. Some of them are for building creative (I’ve made one of those myself). Others are tackling key parts of our workloads, plus eventually AI agents building custom software for very specific needs and uses.
The modern martech landscape in 2025 is all about solving these problems with smart systems.
What marketers need most in 2025 from martech, attribution, and analytics solutions
Unified and composable data infrastructure
Marketers need a unified data layer, whether via cloud data warehouses or data lakes, or, for simpler solutions, your data in Singular. They also need a composable data infrastructure: modular, flexible, scalable, and accessible for AI and other analytics. That includes Singular’s Marketing ETL solution, to get your marketing data just the way you want it, or our brand-new, super-modern ELT platform, Extract, which is perfect for unifying the 20 different data sources you likely have. (Learn more about Extract here.)
AI-enhanced attribution and measurement
Attribution and marketing measurement are complicated. You need Unified Measurement attribution models blending all the available data, including probabilistic, SKAN, first-party data, extra data from Meta (AMM)and Google (ICM) that you can only get via an MMP. You also need modeled cross-channel conversions, with the ability to access incrementality and media mix modeling (MMM) capabilities.
Instant AI-powered insights and recommendation And you need AI, like our Claude (and other LLMs) integration for near-real-time insights and automated causal analysis. Natural language interfaces for querying marketing performance are huge for easy non-technical access to data, and sometimes you learn something or uncover some insight that you never would have thought to search for in a dashboard.
Scalable content and personalization management
Marketers also need GenAI-driven creative production and testing, maybe extending to brand-specific LLMs for on-brand AI-generated content, and likely, eventually, dynamic personalization engines leveraging integrated data and content intelligence.
AI-driven SaaS services
Marketers increasingly need external AI-powered services for things like attribution, analytics, and marketing operations. You can’t build everything in-house while also running your business, and this frees teams to focus on product, marketing, and strategy while benefiting from AI-driven operational efficiency.
The goal is freeing marketers and UA experts from grunt work and giving them the tools to have almost instant rich data and deep insights which enable super-fast campaign creation, execution, and optimization.
Yeah, this is part of the golden age of marketing measurement
Briefly, what we’re seeing is more data from the big platforms like Meta, Google, TikTok, Snap, and more. We’re seeing more on-device data. And we’re building an increased ability to assimilate that data, enrich that data, and analyze that data in automated, quick, AI-driven ways.
That’s a golden age, in which the multiverse of analytics data from different angles and sources provides what might actually be better attribution and analytics results than the comparatively “perfect” tracking enabled by IDFA in the past and even GAID right now.
Why?
Because every buying decision, every install, every IAP purchase, every subscription converted is the result of a complex series of motivations and drivers, 1 of which is probably the last-clicked ad someone saw, but many of which might be hidden or non-digital. So getting more data from more sources shows more causality, including views, clicks, and other actions.
Singular can help
As a leading marketing analytics platform, we deliver new AI-driven and AI-enhancing solutions regularly. Talk to us to learn how you can get your data where it needs to go, and how to get the insights you need at the speed you want them.
Do you need an MMP if you just use one ad network? There’s more than a few mobile marketers and new app publishers that think so.
At first glance, it’s pretty understandable: there’s limited confusion, there’s only 1 ad partner to integrate, there’s no need to de-dupe attributions between multiple partners, and there’s the simplicity of just dealing with 1 partner, 1 dashboard, and 1 system.
Naturally, if you have 5 ad partners, or use 10 ad networks, you need an MMP. And we’ve discovered that more ad partners are generally better: using more ad networks is highly correlated with lower CPI, higher ROI, and higher retention.
But do you need an MMP for just 1 or 2? Or 3?
In a word, yes. (And it can be entirely free, by the way.)
I chatted with Owen Farmer, who’s been publishing apps since the age of 14, about why. Hit play and keep scrolling …
You need an MMP for more data
Farmer is the founder of Word Power, a vocabulary app that he started as a teen. He kicked off marketing with just one ad channel: Apple Ads. But he soon learned that the data Apple provided wasn’t really enough:
“There’s some people in the industry who advise against using an MMP in the beginning, which to be honest, blows my mind,” he told me.
“Apple Search Ads only tells you what happens up to the download, but it doesn’t tell you how many people install, how many people start a trial, or how many of those trials convert,” he added. “Having that data is essential. And if we’d been flying blind using just cost per download as our primary KPI, we’d be well underoptimized.”
A key thing you need is to see how different keywords perform in terms of user behavior after the click and after the install.
So Farmer uses Singular to measure all of his campaigns and sees far beyond the install event. And that’s exactly what most apps, even early ones, need.
But there’s more, much more
It’s pretty easy to kick off marketing with 1 partner, and then expand. Maybe you start with Meta or Google, but TikTok looks so tempting for a bit of influencer marketing juiciness.
With an MMP, adding partners is incredibly easy. And now you’ve got 1 place to look, not 2 or 3, and if there are conflicting claims, you can de-dupe them.
But there are plenty of reasons you still need an MMP even if you just use 1 ad partner:
See more of the user journey As Farmer noted above, you need to see the whole user or customer journey, including what happens post-install. An MMP like Singular gives you that activity, that engagement data, any purchases, and more.
Get holistic cohorts and LTV analysis When you get purchases and engagement data, you get cohort analytics that tell you which campaigns were profitable long term. Now you’ve got LTV and can make smarter marketing, bidding, budgeting, and optimization decisions.
Measure organics Meta isn’t going to measure organics for you. But based on your vertical, organics can be a major source of installs and revenue. That’s SEO, that’s web, that’s AI recommendations on ChatGPT, that’s social, and more …
Measure owned media Depending on your vertical, skills, size, and interests, you’re likely to have owned marketing channels including your website, social media, email and maybe even real-world or out-of-home options. With Singular you’ve got weblinks, deeplinks, QR codes, and much more, all with analytics hooked up to them.
Better attribution with Unified Measurement As hard as it may be to believe, Meta regularly under-reports its own effectiveness. View-through and brand awareness that results in later activity account for most of that, while some probably is just due to the fact that no 1 measurement methodology is perfectly accurate. So Singular’s Unified Measurement that looks at many different datapoints including SKAN, store data, first-party data, and more, shows you a more accurate picture. (Importantly, Unified Measurement also shows you all the component datapoints that add up to its decision, giving you valuable insights from multiple perspectives.
Fraud protection Yes, fraud exists. What used to be FAN (Facebook Audience Network) and is now MAN (Meta Audience Network and GDN (Google Display Network and AdMob bring in advertising on millions of publishers’ sites and apps, and guess what: not everyone is an angel. Singular’s fraud protection looks at 51 different parameters to judge the safety and reliability of reported clicks, installs, and conversions.
AI-powered creative analytics It’s no secret that we just dropped Creative IQ, which gives you a creative-to-ROI visual gallery to instantly spot winners and losers. In the case of some fairly black box platforms that mix-and-match your creative elements on the fly, Singular also reports effectiveness data that you literally can’t see on the platform. Which means you get more data in Singular than in their dashboards … not to mention the data governance that helps you keep track of all your creative and campaigns easily.
Deeper, more granular data access For some programs, like Meta’s AMM or Google’s ICM, you simply can’t get additional granular data any other way than via an MMP. Meta and Google rely on MMP’s like Singular to safeguard the data and ensure that it’s used safely and properly. This is huge, because now you have privacy-safe access to data that gives you 2020-like powers, similar to when IDFAs were easily accessible.
Future-proofing for growth When building a startup, it’s a good idea to set up the foundations of future growth, as long as you can do so without over-indexing on things you don’t know for sure that you’ll need. Setting up an MMP — perhaps on the free tier — is a good way to future-proof for growth. As you grow, you won’t always just have 1 or 2 ad partners, and with Singular, it’s super-easy to add new partners with a few clicks.
AI assistance Singular recently released MCP integration for major LLMs, and marketers are already asking Claude for deep insights on their marketing data. ChatGPT is around the corner. This is a super-simple conversational way to get all the data and insights you need to grow.
Alerts & notifications Have you overspent on a campaign when you’ve put in guardrails against it? I have. When campaigns break parameters, Singular can provide alerts and notifications.
Record of bids and budgets Singular captures and stores all campaign execution information such as historical bid changes, creatives trafficked, and more, establishing a system of record. When you want to know why something happened or changed in your marketing results, you have a place to look.
Deep linking, smart links, and measurement You want a great user experience as people come into your app from multiple places. Firebase is shutting down its Dynamic Links solution next month, so you’re going to need a deeplinking solution. Singular’s is built in, and best in the industry. And yes … it even works for branded QR codes.
Web banners Need smart banners for your website (mobile or desktop) that incorporate website-to-app measurement and work with your publishing solution? Done.
Better reporting tools and interface Ever need to share reports? Manipulate or customize a dashboard? Do simple pivots with your data? Use custom dimensions that align with your KPIs that you know drive revenue? It’s all in your Singular MMP dashboard.
That’s 15 reasons why you need an MMP, even if you’re just starting out. And even if you have precisely $0 budget for it.
Custom KPIs can be super helpful
I know I buried it in #15 above, but custom KPIs unlock huge opportunities.
UA all-star Hannah Parvaz calls it your north star metric. It’s the 1 key thing that if you hit it, you know you’re going to get revenue, be cash-flow positive, and make money. It takes some time and deep understanding of your app and users and their journey to know, but once you know it, you can optimize all your marketing campaigns around it.
As she told me recently:
“This is a metric that represents both what the customer wants — they’re getting value, so they’re coming back — and it also represents what the business wants: they want that usage. So your North Star metric, it’s really important that you’re balancing both of these.”
For Farmer, as a bootstrapped founder without millions in VC funding, his north star metric is oriented around early profitability.
“Day 14 ROAS is our main success metric because we’re a seven-day free trial … as a leading indicator, we’re always looking at cost per trial.”
A key reason you need an MMP is because you can set up custom dimensions in your reporting to orient all your dashboards around that key thing, making it easier to know when you’re approaching it, meeting, and hopefully, eventually, exceeding it.
So much more in the full podcast
Yeah, you need an MMP. But you also need much more.
I think you’ll enjoy this episode which follows a solo founder (well, with some freelancer help) on his journey to significant success at a pretty young age.
Here’s a preview of what you’ll find:
00:00 Do you really need an MMP?
01:46 Owen’s Journey: From Timer App to Vocab Success
04:59 Marketing Strategies: From Organic to Paid
06:15 The Role of Meta Ads and Campaign Optimization
08:22 Subscription Models and Monetization
15:34 Balancing Product Development and Marketing
20:51 The Importance of Focus and Product-Market Fit
For 99.9% of apps, mobile user acquisition is the foundation of app growth. Scaling mobile user acquisition is critical for all but the most viral of apps.
But in 2025, the rules are changing.
Yes, there’s privacy still
Yes, media costs continue to rise (the good news for app publishers who monetize via ads is that CPMs are going up too)
And yes, there’s challenges in making great creative and driving profitability
But cash isn’t (nearly) free anymore, and marketers can’t just spend their way to growth like it’s 2022.
To scale efficiently, you’ll need smarter mobile user acquisition strategies that include diversification, AI-powered everything (including you), first-party data, creative intelligence, and an acceptance (however reluctant) that the ecosystem is fragmented and will change dramatically every few months.
Key takeaways
Scaling UA in 2025 demands smarter strategies, not just bigger budgets.
AI, creative optimization, and diversified channels are your edge.
Better data is back—combine it to navigate the fragmented ecosystem.
Retention drives sustainable growth; acquisition alone won’t cut it.
Data and the state of mobile user acquisition in 2025
UA runs on data. Always has, always will.
Good data drives optimization and growth; bad data — or unavailable data — results in confusion, chaos, and ultimately stasis. You can’t invest if you don’t know what’s working.
That said, marketing measurement data is getting better, bigtime, in 2025. Which is a big statement, given that IDFAs are still rare thanks to ATT, and SKAN hasn’t really delivered on its promises.
The improvement is coming from other sources.
SKAN and Privacy Sandbox are still around, but we’re getting more and more data from ad networks big and small in 2025.
On Meta, AMM is back, baby … Advanced Mobile Measurement offers much more granular data on acquisition, offering better ROAS modeling, cohort analysis, and creative optimization (plus Meta also offers Advanced AEM for more measurement richness)
TikTok has Advanced SAN
Snap has its own version of Advanced SAN
And all the other non-SAN networks offer probabilistic measurement solutions that are good enough for attribution and optimization, and are working on modeling solutions too
Meanwhile, first-party data is still great. Zero-party data is useful. SKAN data can complete a picture. If Privacy Sandbox ever happens, it’ll fit in somewhere too.
Scaling mobile user acquisition may not be as simple as it used to be, though.
Without IDFAs on iOS, post-truth marketing measurement is about triangulation and bread crumbs rather than rock-solid certainties derived from a single datapoint. But marketing measurement is FAR better now than in the first few years after iOS 14.5, and maybe — just maybe — better than ever.
And on Android, of course, we still have GAIDs, plus referrers.
Scaling mobile user acquisition in 2025
If all of that is true, maybe that’s 1 reason why mobile ad spend is on track to hit $362 billion globally this year, according to recent projections.
That sounds good, but all those billions of dollars drive trillions of ads, and that’s a lot of noise. Our already super-competitive market is now hyper-competitive.
So what’s your playbook for scaling mobile user acquisition in 2025?
Adapt to post-truth marketing measurement
One of the biggest adjustments for marketers in 2025 is adapting to the new measurement reality. When SKAN tells you 1 thing, Meta another, and your internal app data yet another, Singular’s Unified Measurement brings them all together.
Enlist AI to scale YOU
Money is scalable. Machines are scalable. Data is scalable. Your biggest challenge in 2025 is scaling you: your attention, your analysis, your knowledge, your abilities. And that takes AI. Start with Creative IQ from Singular for creative optimization. Add our ability to connect your data with LLMs to ask natural language questions of your data so you both see and understand more. And use all the other tools that make sense … your competitors certainly are.
Diversify to lower CAC and boost ROI
Often you need to KISS, keep it simple stupid. Sure. But scaling mobile user acquisition requires using an MMP to KISS many frogs, because scaling ad partners is key to boosting ROI, reducing CPI, and earning higher retention rates. We have the data; we can show you the receipts. We’re seeing almost a 3X jump in ROI from those using 5 or fewer ad networks to those using 6 or more.
Understand and use incrementality
I’m not talking big fancy studies. Nor am I talking about insane levels of analysis or special tools. But check if adding an ad partner is helping you to fish in OTHER streams, not just dropping yet another fishing pole in the same stream. There are simple ways to do so, as Jonathan Reich, the CEO of an app publisher with 750 million app installs, recently told me.
Of course, as you do all this scaling and work, don’t forget about the Pareto Principle: 80% of your results will still come from 20% of your efforts.
Of course … growth isn’t just about acquisition
UA costs are going up. I’m working on Singular’s next Quarterly Trends Report right now, and CPIs are up steeply in some cases.
That means mobile user acquisition strategy isn’t just about acquisition.
Don’t be so focused on UA you forget about UX, value, engagement, and ongoing product development to boost retention.
Retention is as important if not more important than acquisition.
Don’t underestimate its critical importance. When you keep a high percentage of your existing users, you spend less time and energy fighting to fill a leaky bucket at the top. High retention means you can be profitable, even at a high CAC. Of course this is an ongoing strategy, not a quick fix, but every bit you’re able to improve retention and engage users results in huge gains in LTV, ROAS, and ultimately break-even.
This all supports a broader shift in mindset.
In 2025, mobile user acquisition is about more than the number of installs.
It’s quality, it’s onboarding success, engagement, and the full lifecycle including hybrid monetization with in-app purchases, subscriptions, and ad monetization.
And yeah … creative matters in mobile user acquisition
As you know …
Creative remains one of the most important levers for performance.
Creative can’t be underestimated, even as marketing teams are asked to do more and more with less. It’s the single most important factor (and yes, there are more.)
What are the elements of an ad that contribute to success, according to Mckinsey?
Targeting: 9%
Deliverability/viewability: 22%
Creative: 49%
Spend time where you need to to build great campaigns. But don’t neglect creative.
Because when you find a 10X creative … a unicorn creative … you change your mobile user acquisition picture. 10X creatives allow you to scale profitably, when you’re doubling your profitability goals at low volumes.
Can’t find the unicorn creative?
Keep looking.
Until you do, it’s incredibly hard to scale ad spend profitably. And if you can’t scale ad spend profitably, you can’t grow sustainably.
Need some insider tips on creative? Singular’s newest Creative Optimization Guide will be a valuable resource.
Final thoughts
Scaling mobile user acquisition in 2025 means adapting to a world where privacy and performance coexist.
You and your team need to blend automation and AI with creative strategy and data fluency. Success today requires a shift from volume to value, with a focus on acquiring the right users, nurturing them, and driving long-term engagement.
If you’re looking to scale efficiently and effectively, talk to us to see how the top app marketers are doing it.
It’s tough to win with creative today. Everyone is smart, everyone is spending, everyone has AI, and everyone is fighting like H E double hockey sticks to kick your metaphorical rear end.
So what do you do?
Watch our latest webinar with 6 creative experts on building a creative advantage. And shamelessly steal some of their best tips to shift into a higher creative gear.
While you’re watching the full webinar, check out some of the top insights from the webinar here …
Win with creative: what’s the problem?
First up: insight from attendee polls
Everyone wants to win with creative. The problem is that desire is not enough. It’s freakishly hard.
We asked the hundreds of webinar attendees about their core creative challenges, and what’s working:
When it comes to UA performance, what is currently your biggest creative challenge?
Which ad format has driven the highest ROAS for you recently?
Where are you using or planning to use AI in your creative process?
What’s actually working?
The biggest creative challenge: figuring out what’s actually working. (Sounds like it’s time for Creative IQ!) The second: coming up with new ideas.
Here are the results:
34.1% Figuring out what’s actually working
25.0% Generating enough new creative concepts
20.5% Fighting creative fatigue
15.9% Scaling production without losing quality
4.5% Localizing creative effectively
Marketers’ biggest creative challenge is understanding what’s working, followed by the pressure to generate new creative concepts. Creative fatigue and production quality are also significant concerns: it’s hard to maintain high-performing creative over time.
Understanding what’s working should not be the biggest challenge, but with siloed data all over your partners’ dashboards and black boxes proliferating, it’s understandable. As suggested above, get Creative IQ from Singular to clear the confusion.
What are the right formats?
To win with creative, you also need to use the right formats. We asked the webinar participants which ad formats have driven the highest ROAS for them recently.
Video and UGC ran away with the field.
Here are the results:
39.0% Short-form video
27.3% UGC
16.9% Playable ads
10.4% Influencer-driven
6.5% Static image
Short-form video dominates with 39% of respondents calling it their top ROAS driver, followed by UGC at 27.3%. Playables and influencer content have meaningful traction too. Static images are falling behind, reinforcing the move toward interactive and motion-based formats.
Where does AI fit in?
And finally, we asked webinar participants where they were using or planning to use AI in their creative process …
Here are the results:
27.2% Concept creation
20.6% Visual generation
18.6% Voiceovers or audio
17.5% Localization and translation
15.5% Creative tagging and reporting
0.5% Not using AI yet
Most marketers are already using AI to ideate and generate visuals. Concept creation leads, followed by visual and audio production. Very few said they’re not using AI, signaling that adoption is already mainstream in creative workflows.
Win with creative: magic and math
Our own CMO Stephanie Pilon hosted the session and she kicked it off with a kick in the pants highlighting just how important it is to win with creative …
“Targeting only makes up 9% of the success of an ad… deliverability and viewability make up 22%. But creative makes up a shocking 49%.”
Just to state the obvious because it’s so ridiculously critical, that means nearly half of your ad’s performance hinges on the creative you put in front of people.
No pressure.
There’s magic in creating creative, but there’s math in finding the ones that work best.
As I’ve heard from marketers before, it doesn’t have to hang in the Louvre. It does have to perform. It does have to demand the click. It does have to sell the next step in the customer journey.
How?
“It’s just about continuous iteration,” said Lisi Gardiner, Director of Product at Singular. “And I think a lot of marketers understand that it’s one of the best, most cost effective ways of driving performance.”
Structure matters too
Don’t misunderstand that continuous iteration. Testing creative isn’t just about throwing variations into the wild. It’s also about organizing ideas into frameworks and testing them with purpose.
Tinuiti’s Brent Hastie talked about using performance buckets to structure testing:
“These aren’t just themes, they’re concept archetypes … storytelling cohorts, if you will, that we use to ensure we’re telling distinct stories that can be tested and optimized.”
Those might be buckets of ad types focused on social proof, or product/service education, or brand introduction.
That helps with iterations, too. Now you can generate variations on the bucket themes easier, and structure them better. All of this will help you understand what themes resonate with specific audiences in specific geos.
Once you’ve dialed that in, you can dive down into more granular elements, fine-tuning what is working into what is awesome.
For example, if your UGC personal story bucket is doing well, maybe you try different version, like a skeptical weight loss veteran, or a first-time mom who’s looking to get back into shape, or an exec who just wants to fit those jeans again, or a retiree looking to be able to play more golf.
Win with creative by sweating the assets
If you’ve got it, flaunt it.
You’ve made a bunch of stuff already: brand images, product images, screenshots, videos. Rip, mix, burn all that stuff to make it work for different purposes.
From TikTok’s Dina Kirshner:
“We really try to emphasize making the most of what you already have… repurposing material, whether it’s a photo or clips that you have, and then making them feel like new again to create that robust volume you need.”
It’s always tempting to make new stuff, or to discount what you have.
But some of your best campaigns might come from materials you’ve already created.
So much more in the full webinar
There’s gold in this webinar. Here’s an overview of what you can expect to find …
00:00 Introduction and Session Kickoff
02:33 Importance of Creative in Advertising
03:43 Panel Introduction and Session Overview
06:10 Biggest Creative Challenges
07:16 Why Creative is Critical
10:21 Adapting Creatives for Different Audiences
15:37 Brainstorming New Creative Ideas
20:12 Best Practices for High-Value Creative Assets
Generative AI for marketing is getting really, really good in 2025.
Creative is of course the ultimate fantasy. Inspired by Star Trek’s Jean-Luc Picard, we’ve always dreamed of the ability to literally speak reality into existence. And while we don’t have a Holodeck and we’re not quite there yet for “Tea, Earl Grey, hot,” (atoms are hard!) we’re actually been given shockingly impressive powers with generative AI. Text, images, video, Python code, trip itineraries, sweet sentiments for significant others on Valentine’s day, entire apps … all of that is possible with generative AI today.
It’s probably not perfect, but it is available.
Sometimes it’s shockingly good. And we’re just getting started. And generative AI for marketing is entering prime time.
A storyteller I interviewed a few years ago used generative AI to create images and video and audio for interactive podcasting. At the time, she compared our present era to the first phones with onboard cameras … the 1.5 megapixel ones. (Remember, the same ones Kodak laughed and dismissed as toys.)
Today everything is different. We’re no longer at the crappy 1.5 megapixel smartphone camera era. We’re in the era when 100% straight up you can use generative AI for marketing and get frankly amazing results. Probably not as good as a great designer, let’s be honest, but still almost unbelievably good.
That’s kinda the scary part. It’s now as bad as it ever will be. It will only get better from here.
(How good can it get?)
There is a lot you can do today with the current state of the art in generative AI to help with marketing. I regularly use ChatGPT to auto-create a summary of a podcast episode. It takes 2% of the time that it would take to do manually. Most of you use ChatGPT, or Claude, or Gemini, to do similar things. This is now normal, basic, unsurprising, and expected.
Before I share a few tools that can help you, however, let’s see what the big machine in the cloud says first. Because … there simply has to be some generative AI in a post about generative AI, right?
ChatGPT: what generative AI says about using generative AI for marketing
Why yes, I did ask ChatGPT for its opinion on using generative AI for marketing. Here’s what ChatGPT “thinks” about the potential.
Generative AI, like GPT-4, has the potential to revolutionize the marketing landscape in various ways. While some of these implications have already been observed, more applications will likely emerge as the technology evolves.
Here are some ways generative AI can impact marketing:
Content creation: Generative AI can create high-quality, engaging, and contextually relevant content at scale, including blog posts, social media updates, and email campaigns. This can save time and resources for marketing teams, allowing them to focus on strategy and analysis.
Personalization: AI-powered marketing tools can analyze customer behavior and preferences to create personalized content and offers, enhancing customer experience and increasing conversion rates. This technology can make recommendations based on a user’s browsing history, previous purchases, or other interactions with a brand.
Ad copy and design: Generative AI can help create and optimize ad copy and design elements, testing various combinations to identify the most effective messaging and visuals. This can lead to better performing ads, higher click-through rates, and improved return on investment (ROI).
Chatbots and virtual assistants: AI-powered chatbots and virtual assistants can help brands engage with customers in real-time, answering queries, offering support, and providing personalized recommendations. This not only improves the customer experience but also frees up human resources for more complex tasks.
Market research and analysis: Generative AI can analyze vast amounts of data, identifying patterns and trends that can inform marketing strategy. This can help brands understand their target audience, stay ahead of competitors, and identify new opportunities.
Social listening and sentiment analysis: AI can monitor social media platforms, news articles, and forums to gauge public opinion and sentiment about a brand, product, or service. This information can help marketing teams respond to customer feedback and adjust their messaging accordingly.
Predictive analytics: Generative AI can help marketers forecast customer behavior and demand, enabling them to optimize pricing, inventory, and promotional strategies.
Influencer marketing: AI can help identify suitable influencers for a brand based on the influencer’s audience demographics, engagement rates, and other factors. This can help brands form more strategic and effective partnerships.
Creativity augmentation: AI can assist marketing professionals by generating creative ideas or providing inspiration for campaigns, allowing for more innovation and outside-the-box thinking.
A bit vanilla, sure. Maybe less personality that we might like, and perhaps a bit short on details and links, but not a bad summary. So: where’s the beef? What are the tools you can use today to help you grow your apps?
Here’s a short list that I intend to keep adding to over time … (with 6 brand-new tools just added!)
Generative AI tools for app growth and marketing
Generative AI for optimizing ad creative
Sorry, but we can’t ignore our own tools here! Singular recently released Creative IQ, an AI-powered creative optimization suite makes testing, tagging, and reporting creative results amazing and easy.
Creative IQ lets you analyze the results of your A/B tests better than ever before, automatically categorizes your ad creative with smart tags based on what’s actually in the images, and offers simple visual reporting: videos and images side-by-side with performance metrics.
Creative IQ gives marketing and creative teams the granular, visual, and actionable insights they need to deliver better creative that performs.
Generative AI for marketing measurement and reporting
Wouldn’t you like to just talk to your marketing measurement platform and get all your results served up on a platter?
Now you can, because we (yep, this is another Singular tool) have integrated with Claude (and soon other platforms) to get your marketing results, charts, insights, and suggestions right from a super-powerful but super-simple chatbot.
No SQL or Python needed: you can just talk to your attribution engine. Thousands of marketers already are, and they’re learning incredible insights that sometimes, you wouldn’t even think to check via your dashboard.
Simple and powerful … exactly what we want from AI.
Generative AI for objects or people in your game
You would typically need some skill for game asset generation. With PixelVibe, you let the machine do all the work for you. And yeah, that can be people, props, concepts, environments, weapons, skins or clothing … even graphical user interfaces.
“Rosebud is building the AI Roblox. We will make it 1000X easier for a community of game devs to create and enjoy the games of their dreams … we believe all media creation will be done via generative methods in a few years.”
All is a big word. But maybe …
Pick a “potion model” to select the types of objects you want, enter your prompt, choose an image size and the number of images you want. Here’s “ancient city crumbling into the desert” at 512×512 pixels. It’s not for a hi-res game, but it’s pretty cool. You can pixelate it, “re-vibe” it for a different look and feel, or remove the background:
And, if you want 3D worlds, check Rosebud.ai to generate them.
Generative AI for marketing copy
Writing copy for 500 different variations of my so good most awesomest app ever is soul-crushing. But only humans have souls, fortunately. Microsoft says AI can do it, thanks to its multi-billion-dollar investment in OpenAI and ChatGPT.
“Microsoft has begun to integrate models like GPT-4 and image generator Dall-E 2 from OpenAI, in which it has heavily invested, into its Azure enterprise cloud computing services. Microsoft has also partnered with Adobe to supply its language AI to help write marketing messages and photo captions within the latter’s marketing platform.”
In other words: there’s startups, and there’s enterprise. Both are building AI tools to automate marketing tasks.
Or you can just use OpenAI’s ChaGPT right from the source. Here’s ChatGPT in response to my “give me an exciting and persuasive marketing blurb for my spaceship shooter game” prompt:
“Prepare for liftoff, Captain! Step into the cockpit and unleash your firepower in [Game Name], the ultimate intergalactic spaceship shooter that will ignite your senses. Battle through explosive galaxies, face relentless alien forces, and customize your ship with cutting-edge weapons. Whether you’re seeking epic boss fights, immersive cosmic landscapes, or the thrill of becoming the galaxy’s greatest pilot, [Game Name] delivers heart-pounding action that keeps you on the edge of your seat. Ready to conquer the stars? Your mission awaits—download now and become a legend of the skies!”
It’s not horrible. And it’s perfectly easy to edit to exactly what you want … without the stress of staring at a blank sheet of paper for hours wondering what to say.
Generative AI for personalized calls to action
Getting some of your revenue via ads is great, but it also impacts your gameplay or app user experience. In-app purchases and especially subscriptions promise bigger rewards, but you need to appeal to people persuasively.
From Persado, the “Generative AI Platform to motivate every individual to engage and act:”
“Determine the messages that drive the highest levels of engagement and conversion through advanced decisioning and machine learning experimentation … uncover the nuanced narratives that resonate with your customers in specific scenarios across the entire journey … optimize the messages each customer receives with powerful set-and-forget adaptive technology.”
Almost scary, but very intriguing.
Persado will build an LLM just for you with the right language, tone, and emotion, which you can then feed into your mobile CRM or live ops solution.
Generative AI for new app and game ideas
Coming up with creative ideas for new games is HARD. I mean, how many times can you do a merge meets battle meets puzzle meets dating game mashup? (Once was more than enough, apparently.)
So ask ChatGPT for new ideas for a unique game.
I just did, and the hundreds of thousands of GPUs at OpenAI came back with this super-geeky nerd gem: The Quantum Quest, a “puzzle adventure game that takes place in a world where players can manipulate the fundamental forces of quantum mechanics to progress through the game.”
Sweet.
(Interestingly, GPT-4 thinks “cut scenes” should be “cutscenes.”)
Definitely nerdy, but not too bad, either. Rinse and repeat until you get an option you love, then get some help coding it from
Generative AI for user and customer journeys
Figuring out and mapping customer journeys is complicated and tiresome. Get AI to do it, with Adobe’s Sensei platform.
“Get AI-enhanced capabilities in Adobe Journey Optimizer that are trained on customer journeys, offers, and experience events to increase engagement with customers.
Perhaps more enterprise-y, but will certainly have applications for dev ops tools and certain to hit tools like Braze and CleverTap soon enough. Probably best for those who are cross-platform with web, apps, and other channels that they need to synchronize messaging across.
Generative AI for copywriting that makes money
There’s a lot of story in some games. And a lot of times, you want to encourage users or players or customers to do something, click on a button, and part with some juicy cash.
“Technology like Phrasee is to copywriting what Photoshop was to design.”
Plus, of course, you can use something like this for ad copy generation. The key thing you get from Jacquard: “hyper-personalized brand communications.
Generative AI for videos
Videos can be hard to make and they cost a lot of money and time and focus. Make them instantly from plain text with generative AI.
Sure, these could be for marketing. But they could also be part of your app: sarcastic congratulations from Master Chief in a space battle game, soothing, calming, gentle and positive support in a meditation app, in-your-face screaming yelling Tony Robbins for a tech bro motivation app.
“Our mission is to empower everyone to make video content – without cameras, microphones or studios. Using AI, we’re here to radically change the process of content creation and unleash human creativity for good.”
Deepfakes are some scary tech. But you can use them for good for characters you create in your apps.
That’s one option. I’ve seen HeyGen do some absolutely incredible things as well: train it on you, and it generates almost fully believable video with “you” speaking on-screen. Very impressive.
A game without a soundtrack sucks. But musicians are expensive, and copyright-free stock riffs are pretty much that: stock. So create some cool jams with generative AI for music.
Maybe you can even generate a soundtrack that has always unique style to your game or app, but also adapts to individual players’ or users’ progress, actions, and success or failure.
“Musico’s engines can generate infinite melodies, beats and harmonies, blending autonomy and responsiveness to the creator’s input … we’re exploring the relation between music and narrative to develop a next generation soundtrack plugin for storytellers, game, and cross-media developers.”
A unique sound for your app that isn’t just one track endlessly looping? Novel idea, and one whose time has come.
Generative AI for clothes in a fashion or social game
Wearing the same thing all the time gets old fast, even in the metaverse. Give your in-game avatars more choice and more creativity with generative AI that’s completely free.
“ClothingGAN is able to generate clothing images and mix these images. While mixing, you can control which structure or style that you want the clothing to copy. Additionally, you can edit the generated clothing with several given attributes such as dark color, jacket, dress, or coat.”
In-app purchases, anyone?
HeyGen also offer capabilities like this, as does Ready Player Me
Generative AI to create entire levels or worlds
There’s already procedural generation for infinite-scale games like No Man’s Sky. Upleveling that to full generative AI is the next step to making virtual spaces bigger, more detailed, more real, and more responsive to both game developer and user desires. That makes generative AI a key cog in enabling metaverse-style instant rooms, spaces, planets, or pretty much anything for MMORPGs.
Ready Player One isn’t going to build itself … or is it?
From a session at GDC 2023 with Roblox studio chief Stef Corazza:
“Roblox envisions a future where anyone can create anything, anywhere … the next step in this journey is leveraging AI to help professional game development studios push the limits of what’s possible on the platform—with high-fidelity experiences that reach a massive global audience of millions in seconds.”
As promised, this is a living list. And since the original post dropped, a few new contenders have joined the lineup — some brand new, others rapidly evolving.
So let’s keep it rolling with five more generative AI tools that are worth a little excitement:
Generative AI for explainer videos and product demos
Making launch videos that don’t suck is hard. Making launch videos that people want to watch? Even harder. With Pika’s generative video, you start with a text prompt and end with seriously professional-looking short-form animations. You can create walk-throughs, social snippets, or in-app explainers without a full creative team or a three-week timeline.
It’s still early-stage, but it’s already powerful enough to impress even the most over-it, anti-AI designer on your team. And it’s getting better with every release.
Generative AI for background removal and smart editing
Sometimes, you just need to zap an ugly background, drop in a clean replacement, or make the product pop on your ad creative. Clipdrop takes the drudgery out time-consuming, mundane design tasks, trained on massive vision models and built for fast, intuitive edits.
It does cleanup, relighting, upscaling, object removal, and even prompt-to-image generation. That can save a lot of back-and-forth on those final, miniscule edits.
Sure, Canva can remove backgrounds, but this does it at scale.
Generative AI for campaign-level creative strategy
Way more robust than Grammarly, Writer is a powerful platform for brand-safe generative content that’s tuned to your style guide and voice. That means anybody on the team, from the intern to the project manager, can quickly produce copy that aligns to brand standards and tone.
It’s like ChatGPT with guardrails, structure, and an editor that actually knows your brand voice. Plus, it plays well with office suites (like Google and Microsoft) and marketing platforms, making it easier to integrate with enterprise workflows and content pipelines.
Stock photos are, at worst, corny and, at best, boring. And AI-generated images can be overly slick and creepy, never mind the whole hands thing.
Good AI images, though? That’s another story. Ideogram turns short text prompts into visually unique, brand-right images, and it has a special knack for actually generating legible text inside images. (If you’ve ever tried creating an image with text in Canva or Dall-E … you know.)
The images are pretty awesome, and some are barely identifiable as AI. You could definitely lose a few hours creating memes that don’t look like they were made in 2010.
Looking for real-world feedback on your app or rebrand without actually booking interviews or focus groups? Synthetic Users lets you simulate interviews, tests, and user journeys using AI agents modeled on real demographics and behaviors.
This goes beyond wondering how users might interact with your product or campaign, it allows you to actually watch them do it, click by click, prompt by prompt. Marketers and PMs use it to test messaging, onboarding flows, landing pages, and more, long before the real humans get involved. Faster feedback, fewer wrong turns.
Overview of generative AI tools for marketing and app growth
Tool name
Category
Description
Creative IQ
Ad creative optimization
AI-powered creative tagging, A/B test analysis & performance insights by Singular
Claude + Singular
Marketing measurement
Chatbot interface for marketing analytics—get charts, data, insights without dashboards
PixelVibe
Game asset generation
Generate game characters, items, GUIs, and environments from prompts
ChatGPT / Microsoft / Adobe
Marketing copy
Automated generation of marketing blurbs and campaign messaging
Persado
Personalized calls to action
AI-powered message optimization using language and emotion models
ChatGPT
New app & game ideas
Generate original concepts like “Quantum Quest” with simple prompts
Adobe Sensei
Customer journey mapping
AI-enhanced journey optimizer across channels
Jacquard
High-converting copy
Hyper-personalized ad and in-app messaging for conversions
Synthesia / HeyGen / Runway
Video generation
Text-to-video tools for marketing, in-app cutscenes, or branded avatars
Musico
Music creation
Adaptive game soundtracks based on user actions
ClothingGAN
AI-generated outfits
Create in-game fashion items; mix & edit styles for social/fashion games
Rosebud.ai / Roblox
World & level generation
Create full 3D experiences and game environments via prompts
Pika
Explainer video creation
Fast, AI-generated animations and demo videos from plain text
Clipdrop
Smart image editing
Remove backgrounds, relight, upscale images, generate from prompts
Writer
Brand-safe content
Enterprise writing tool aligned with your brand voice & style guide
Ideogram
Image generation with text
Create visually striking images with actually legible embedded text
Synthetic Users
UX simulation & persona testing
AI-generated test users that simulate behavior across journeys and landing pages
More to come, I’m sure
That’s the latest crop. If you’re using something magical (or weird or brilliant) that should be on this list, let me know … I’m happy to add more.
The innovation is just getting started, and this list could (probably will!) keep growing.
