How LTV forecasting can help you scale your mobile game 10X faster

How can LTV forecasting change your growth curve? In a word: massively. Think orders of magnitude faster.

Just 1 example from about a decade ago, a very tasty mobile game used cohort-based LTV forecasting to boost marketing spend from double-digit millions to about $400 million, and in the process grow to a $2 billion juggernaut. I’m talking, of course, about King’s massive hit, the match-3 game Candy Crush Saga.

This game is an absolute beast:

• Launched in 2012
• Still getting hundreds of thousands of installs a month
• Passed $20 billion in lifetime revenue in 2023
• Still making over $1 billion/year
• Has inspired spinoffs like Candy Crush Soda Saga and Candy Crush Jelly Saga

Anything’s possible, but I’m not sure that kind of absolute explosion is something that you can plan on or engineer, even with the best cohort-based forecasting. But it is an example from mobile growth history of what is possible.

To dive in more to what forecasting can offer, I wanted to talk to an expert. So in a Growth Masterminds episode released just this week, I chatted with Kohort’s Nathan Ceulemans about how LTV forecasting is the key that unlocks growth like this.

Watch the episode here:

Oh, and don’t forget to subscribe to our YouTube channel plus the podcast platform of your choice.

Growth without LTV forecasting …

What does growth look like without LTV forecasting?

Pretty much as you’d expect: simple, linear, slow.

1. You invest $100 in advertising and promotion
2. Your investment has a 200-day payback period
3. After 200 days, you’re revenue-positive on your investment
4. Rinse and repeat: you turn around and reinvest that acquired revenue

(Of course, there are plenty of details and complexities I’m not diving into here. You’re clearly making some percentage of your revenue back right on D1 and through D100, and past there, for example, and revenue that you recognize may not actually appear in your bank account instantly due to delayed store or ad partner payments.)

So this is a safe, easy-to-understand strategy that results in arithmetic growth: you only invest new money that you have already literally recouped. And perhaps you grow an overall 10% during this period after taking into account churn. Or maybe you get lucky and grow 25%.

No one’s gonna write you up in PocketGamer for your explosive market-smashing jump up the App Store/Google Play rankings, but hey, it’s growth.

The question is: how can you accelerate that growth cycle 2X? Or even 10X? 

You’re gonna need really, really good LTV forecasting.

Growth with LTV forecasting … hello Color Bomb and Lucky Candy!

So in Candy Crush, like many games, you can get a bunch of different power-ups. Perhaps the best is the Color Bomb, which removes all candies of a certain color from the board. Or maybe it’s the Lucky Candy, which matches more and helps you clear levels faster.

That’s like growth with forecasting.

If you get good enough at forecasting to predict future earnings with a high level of accuracy, you can potentially access capital via companies like PVX Partners or Braavo to spend earnings that you haven’t yet realized.

And that exponentially increases the rate at which you can grow.

The critical component, of course, is having enough faith in your LTV forecasting to spend as-yet-unrealized earnings. And Ceulemans says that, done well, you can get “single-digit accuracy.” That means your forecast is accurate plus or minus 9% … pretty good. If you can achieve that level of accuracy, you can safely spend up to 90% of your estimated future earnings on growth very quickly, and instead of having a 200-day cycle, maybe you have a 100-day or even 50-day cycle.

Again, there are caveats, and you’re going to want to validate assumptions and accuracy regularly: do all future cohorts behave the way existing cohorts do? Does my LTV forecasting take into account that existing users or players were low-hanging fruit and future cohorts might be less profitable? Not every cohort will be a ‘golden cohort,” perfectly self-selected to love your game.

But all things considered, growth can proceed much faster.

Forecasting results: 1 case study

Ceulemans can’t talk about every client and reveal all their private data, of course, but 1 case study that’s public is Lotum, a German mobile gaming company that is both a Singular and Kohort customer. Lotum offers games like The Test, with over 100 million global players, and 4 Pics 1 Word, with over 400 million players.

(In other words: this is not a tiny studio, and their marketing spend was already significant.)

Cohort-based LTV forecasting helped Lotum boost investment and results very quickly in new markets:

• 55% increased UA spend
• 15% boost in revenue
• 10% boost in profits from 4 Pics 1 Word

The key: being able to infer long-term retention and monetization trends for Lotum’s new segments. More accurate forecasting provides the confidence to reinvest continuously at higher levels, speeding up growth.

More learnings

Of course, all this needs to be exceptionally rigorous. LTV forecasting is notoriously complex, and it also can be impacted by externalities that you have zero control over.

And your game needs to be mature enough as a baseline.

“Essentially it’s an extrapolation of historic current behavior,” says Ceulemans. “And so what goes into that is you need to have enough historic current behavior to base it on.”

And if you’re scaling user acquisition, there’s by definition going to be some level of variance in future cohorts.

That’s also true in a positive sense, of course:

“Sometimes a game update, no matter how inconsequential it might seem, just has completely outsize effects,” Ceulemans says.

The classic example is Clash of Clans introducing the Clan Wars feature in 2014. While the game was already popular with perhaps 30 million players and almost $900 million in revenue in 2013, by the next year that had increased to almost 40 million and well over $1 billion of revenue.

Much more in the full podcast!

Check out our complete conversation for much more, including:

• 00:00 Introduction to Growth Masterminds
• 01:03 The Importance of Forecasting in Scaling Games
• 01:32 Understanding User Acquisition and Financials
• 02:24 Forecasting Retention and Monetization
• 04:15 Challenges in User Acquisition and Scaling
• 07:31 The Role of Product Teams in Forecasting
• 13:17 Accuracy in Forecasting with Neural Networks
• 17:51 The Impact of Accurate Forecasting on Growth
• 22:35 Conclusion and Final Thoughts