We built Ktrl to end the era of hoping your campaigns pay back.
Today, Kohort announced the global launch of Ktrl, a platform to optimize ROAS by injecting pLTV signal engineering into every team's UA workflow, powered by industry-leading ROAS prediction.
Today, Kohort announced the global launch of Ktrl, a platform to optimize ROAS by injecting pLTV signal engineering into every team's UA workflow, powered by industry-leading ROAS prediction.
Introducing Ktrl
Every day, UA teams at mobile game studios wake up to the same problem. They're spending hundreds of thousands of dollars across campaigns, and are left hoping it will pay back, vulnerable to the ad network algorithms.
The ad networks want a signal. A target ROAS, a CPI bid, a cost-per-event goal. And they want it updated constantly. Get it right, and the algorithm works for you. Get it wrong, and you burn money or lose scale.
The best UA teams solve this with data scientists, custom models, and hours of manual campaign review. Others use ratios or excel models. Both groups can be slower and less accurate than they should be.
That's why we made Ktrl.
Ktrl makes three promises to game studios and UA teams:
- Never waste a dollar of UA spend. Every campaign gets a daily, data-driven verdict: scale it, cut it, and how to adjust the signal.
- Manage hundreds of campaigns in minutes. Not hours. Not with a spreadsheet. With a single dashboard that tells you exactly what to do and all the integrations to do it in one click.
- Make data-scientist quality decisions, without being bottlenecked by your data science team. The models run automatically, for every campaign, every day. All validated through accuracy backtests.
Ktrl has been in beta since November '25, and since then it's grown to manage $1 billion of annual UA spend. Its models are trained on $6 billion of historical UA spend across hundreds of games, and it delivers predictions with 95% accuracy, updated daily, for every campaign.
The studios already using Ktrl in beta are managing more campaigns, with leaner, more focused teams, at better ROAS than they were 5 months ago. It's the future of UA, and it's available today.
Our UA team have been using Kohort's products for over a year now, to guide day-to-day decisions on where to spend and what to expect from each campaign. Ktrl takes that to another level. It's like having a senior data scientist reviewing every campaign, every morning, before we've finished our coffee. The accuracy of the predictions has fundamentally changed how we allocate budget.
How we achieve this
Ktrl is built around three capabilities that work together. Individually, each one replaces a workflow that currently takes hours or doesn't exist at all. Together, they give UA teams a continuous, automated optimization loop, across every campaign.
Precise signals for every ad network
The core problem in UA optimisation isn't just knowing your ROAS. It's knowing what to do about it. Every ad network runs its own bidding algorithm, and each one responds differently to the signals you send it. Even with identical ROAS curves, a tROAS target that works on Google might be wrong for AppLovin.
Ktrl generates network-specific bidding strategies and targets for every campaign, updated daily. It supports ROAS, CPI, and CPE/CPA campaign types, and it accounts for the way each network's algorithm interprets the signals you give it. The result is a precise, actionable recommendation: not just "this campaign is underperforming," or "lower spend here", but "set this campaign's tROAS to 140% on AppLovin and 160% on Google, and here's why."
This is the signal that turns ad network algorithms from black boxes into tools that work for you. And it's generated automatically, for every campaign, every day, with 95% accuracy.
Your personal data scientist
Behind every recommendation is a model trained on your game, campaigns, and players — but boosted by $6 billion of historical UA spend across hundreds of games.
Every campaign gets its own model that accounts for ad, IAP, and subscription revenue. It predicts LTV and ROAS across every time horizon that matters, from D1 to D365. Then, it converts it between net and gross so you can have a holistic view on your app no matter what data you send to your MMP. And finally, Ktrl offers two incrementality methods to get a blended ROAS estimate to make decisions on.
For most studios, this kind of capability means taking considerable time away from the data science team, as well as maintaining data pipelines, and waiting for a model to mature. With Ktrl, it's automatic. Models are trained, validated, and updated continuously. And because we believe you should never trust a prediction you can't interrogate, every forecast comes with confidence intervals and model accuracy transparency, so you always know how much weight to put on a number.
All the while, your data science team gets time back to spend on LiveOps, segmentation, and everything else that it takes to run a modern, data driven studio.
The best data scientists tell you the answer, how confident they are, and what the range of outcomes looks like. Ktrl does exactly that, for every campaign, at scale.
Ktrl significantly reduces the guesswork around early performance signals. UA Recommendations acts like a coworker you can consult for a second opinion, the confidence rate of the recommended action is very helpful when it comes to reliability.
Always-on agentic alerting
The hardest problems in UA aren't the ones you know about — they're the ones you miss: a campaign that starts overspending on a Tuesday, a network that silently shifts traffic quality, a geo that stops converting. By the time you spot it in a weekly review, the money is gone.
Ktrl monitors every campaign, continuously, against statistical baselines derived from your own data. It doesn't flood you with alerts every time a metric twitches. It watches for the patterns that actually matter, the ones that separate noise from a genuine problem. And when something triggers, it tells you what's going wrong and how to fix it.
All signal, no noise. That's the difference between an alerting system you ignore and one you trust.
Why we built Ktrl
The UA industry has a paradox. Studios spend millions acquiring users, but the tools they use to manage that spend haven't kept up with the complexity of the problem. Ad networks have gotten smarter. The number of campaigns has grown. The signals they need have gotten more nuanced. But the workflows haven't. The spreadsheets, the manual reviews, the delayed data. All of it has stayed the same.
We built Ktrl because we believe UA teams deserve the same quality of tooling that high frequency traders have had for decades. Real-time signals. Automated decision support. Predictions you can actually trust. And a system that watches your portfolio around the clock.
Get started
Ktrl is now generally available. Visit kohort.io to see it in action and start optimising your campaigns today.