How Babbel grows

When starting in a revenue role (Babbel, then Yousician), split subscription revenue into four buckets: new subscribers, upgrades between plans, renewals, and win-backs (churned users who return). Track how each bucket evolves over the years in Excel, find the big jumps or drops (e.g. new-subscriber revenue jumping from a given month), then ask what happened and map it to the decision (e.g. launching a lifetime subscription). Link or embed the documentation of each decision in the same spreadsheet and turn it into a short presentation to share. Prefer this over reading every entry of a long experiment log, since many experiments don't matter.

8fit, Blinkist and Babbel, big Berlin clients of Outbrain and Taboola, tried the networks' new app product that sent clicks straight to the store. They made creatives more and more obviously about an app (screenshots, 'download the app' text) to prepare users, but store-to-install conversion stayed very low. Readers of native ads don't expect to land on a store.

Two ways to buy radio: host-specific buys (a personalized read that reaches that host's audience, at a much higher cost) or broad national buys scattered across the whole country at a much lower cost. Babbel buys broad and national because its audience is broad; radio audiences skew older, less technical and higher income, who buy subscriptions. Targeting is limited, so the efficiency comes from buying broad and cheap.

Buy linear (cable) TV broadly: a large number of spots at very low CPMs. Measure each spot with spike attribution (the lift in web traffic and downloads right after the spot airs) and use it to optimize networks, times of day and channels, much like digital campaigns.

Ask all users a 'how did you hear about us' question and use the answers to explain organic sign-ups that came from TV or radio. Combine with web click attribution (Google Analytics), mobile attribution (Adjust), incrementality tests, media mix models and TV spike attribution; a data team aggregates everything into a proprietary model. Don't rely on one model or source. Start small: offline no longer needs a big upfront budget.

Instead of dropping users into a limited free product, run the sales 'discovery' step up front: a registration/sign-up flow asks many questions about the user's needs and uses the answers to explain how Babbel fits them. Then a hard paywall; only one free lesson is available. The goal is that by the end of the flow people feel compelled to buy without testing the app. Worth testing against freemium per app.

Offline viewers split about 50-50 between the website and the app, and Babbel could not change that, so it optimizes both flows. On the web, the quiz-like survey starts immediately with no download hurdle, collecting data and having the discovery conversation before any commitment; web also gives much better data than the app stores. Most web visitors are on phones anyway.

Tested out-of-home (billboards, subway ads) and direct mail. They cost more and reach fewer people; they suit geo-specific businesses, not apps with a nationwide audience, which do better sticking to TV, radio and podcasts.

Babbel invented a proprietary metric, 'learner success', tied to the customer's goal of learning a language, and optimizes for it; the growth team does not track DAU or MAU. Users get checkpoints showing they are on track, rather than usage or streak counts.

Podcasts are one of Babbel's three main offline-style channels with TV and radio. No details on how they are bought.