Performance marketing for AI-era growth – backed by models, not gut feeling.
TL;DR
  • The plan was built bottom-up, not top-down: a unit economy model per platform (iOS, Android) and geography set the benchmark cost of install, trial and subscription that ROI allowed – and only then was the scaling budget written.
  • Cost per subscription trial was the proxy metric, chosen because it is fast to reach and optimizable in every channel; actuals were compared to the forecast weekly, per platform and geo, including how much room for scale was left.
  • 50+ influencer collaborations worldwide produced the creative that Meta and Google then scaled – the ROI-positive publications also lifted trust, which fed back into performance.
  • Revenue grew 5× while the profit KPI was exceeded, and the round closed at $30 million in under 3 months.

Situation

Praktika is an EdTech startup that uses AI avatars to create a personalized language-learning experience. It focuses on speaking practice for non-native English speakers through an accessible mobile app. Praktika has already found a product-market fit as a startup and was preparing for the Series A round for scaling.

As Senior Marketing Manager at Praktika, I was responsible for the development of unit economy model benchmarks and establishing user acquisition at scale to achieve ROI-positive growth.

Task

At that startup stage, it's vital to establish new student acquisition via marketing channels while meeting ROI KPIs. This will be the base of revenue growth and lead the company to close the Series A round successfully.

Thus, the challenge was to turn overall requirements into numbers, set the benchmarks, and plan and execute the scale while maintaining the ROI KPIs.

Action

To meet these challenges, I've decided to divide the tasks into the following steps:

Thus, based on the goal decomposition, I've formed a bottom-up plan for scaling. After justifying this plan to the founders, the team and I began to put the plan into action.

We've picked the cost per subscription trial as the proxy metric because it was fast to achieve and easy to optimize across all channels.

Starting slow, we were looking for communication approaches that would resonate with users. We began by collaborating with influencers and testing these creative ideas for performance channels. After several iterations, we found ROI-positive options that allowed us to scale.

Further, we scaled all performance channels week by week and compared the weekly actual results to our initial forecast to see:

It is also worth noting that the ROI-positive publications with influencers supported brand awareness and users' trust, boosting the performance of the Meta and Google channels.

Result

Results in numbers

MetricTargetResultHow measured
Revenue×5Against Praktika's profit KPI targets, which were exceeded
Series A round$30M, closed in under 3 monthsCompany funding round
Influencer collaborations50+ worldwideCampaigns run across platforms and geographies
Scaling controlWeekly plan vs factCost per action, trial-to-subscription conversion and remaining headroom, per platform and geo

Table 2. Outcomes as stated in this case. The scaling plan was built bottom-up from the unit economy model and checked against actuals every week.

Written by Maria Atamanova – Senior Growth & Performance Marketing Manager and applied mathematician; twelve years growing businesses – apps, e-commerce, EdTech, AI data platforms – on numbers, not gut feeling. The unit economy model behind this scale is public as an interactive tool – the Unit Economics Estimator.

More cases