
Some takeaways:
- Streaks are an engaging tool, but they can’t make up for a weak product. If users don’t find inherent value in your product, streaks won’t change that. Build something that users actually want to come back to regularly, then layer in engagement mechanics like streaks to amplify retention.
- Test copy relentlessly: A small copy change can dramatically affect user behavior. Duolingo’s simple switch from “continue” to “commit to my goal” helped users feel more invested and significantly increased retention. Don’t assume you’ve found the best wording—run tests and keep refining. Copy testing is cheap and can lead to massive improvements in user engagement.
- Duolingo initially overcomplicated streaks by tying them to XP points, which confused users. Simplifying the system to just completing one lesson a day made the streaks more effective. Your streak or goal feature should be easy to understand at a glance and fit seamlessly into the user experience—don’t make users jump through hoops to engage.
- For Duolingo, retention is most fragile in the first seven days. They found that if they could get users to stick around for that critical week, their chances of long-term retention went way up.
- Duolingo’s “opt-out” feature, where users could choose their streak duration, gave users a sense of ownership and significantly increased retention. When users make an intentional choice to commit to a goal, they’re more likely to follow through.
- Perfecting every detail before shipping can lead to paralysis. Duolingo didn’t wait for everything to be perfect before launching new features—they tested ideas quickly and iterated based on user feedback. Your MVP doesn’t have to be flawless. Get it out there, see what works, and improve from there.
- Align teams around metrics: Structure your teams around the key metrics that matter—daily active users (DAU), retention, etc. Duolingo’s teams were all focused on outcomes, not individual features. When your team is laser-focused on driving the right metrics, you avoid getting lost in feature creep and can prioritize the work that drives growth. Always ask, “Is this moving the needle?”