What if Duolingo optimised for fluency?

Redesigning a motivation system for progress instead of daily obligation

Duolingo promises fluency. The current system is a carefully layered set of mechanics.

But almost every mechanic inside it — the streak, the XP, the leagues, the owl — is answering a different question: how do we get you to come back tomorrow?

For a while, those two goals travel together. Then they diverge. I have a 661-day streak and I still can’t hold a conversation in any of the 6 languages I have attempted, which tells you roughly which one the system is actually optimising for.

I unpacked how that happens in **Part 1.** This piece is about what you’d build instead.

Because fluency and retention — those aren’t opposing goals. A product that helps people become fluent should also retain them. But when the two come into conflict, today’s Duolingo usually chooses retention.

What happens if you flip that?

There’s a useful idea from Self-Determination Theory here.

The theory argues that long-term motivation isn’t driven by rewards alone. Instead, people stay engaged when three psychological needs are supported:

• Autonomy — feeling like you have meaningful choice.
• Competence — feeling yourself genuinely improving.
• Relatedness — feeling connected to other people or to a larger purpose.

Looking back at Duolingo through that lens: The app does a fantastic job of making you return. It does a less convincing job of making you feel autonomous. And it often measures performed competence (XP, streaks, league position) instead of actual language ability.

But what’s a different way of designing that answers these questions we are after? I explore it here.

📕 Reframing the Goal: The Learning Momentum System

As the name suggests, I believe this is about reframing the goal rather than redesigning the product.

Right now, Duo is effectively (and very angrily) asking:

Did you do your lesson today?

What if it asked:

Are you moving forward?

While those sound similar, they’re not. One measures attendance; the other measures progress.

So thinking in terms of momentum instead of obligation.

💡Core idea

The central idea is simple.

Replace the daily obligation model with a learning momentum model.

Momentum isn’t about never missing a day.

It’s about continuing to move forward over time.

That means rewarding depth instead of attendance, helping people recover after breaks instead of punishing them, and being honest about what they’re actually getting better at.

Here’s what that looks like.

1. 🔄 Consistency score (replaces the streak)

The first thing I’d remove is the binary streak. Not because consistency doesn’t matter — it absolutely does. But binary systems create strange behaviour. Once you’ve invested enough into them, the goal changes from learning Spanish to protecting a number.

Instead of asking whether you’ve maintained a perfect streak, I’d show a rolling consistency score based on the last month of learning. Missing one day barely changes; taking a week off has a bigger effect and coming back starts rebuilding it immediately.

The important difference is emotional.

A streak says: Don’t lose what you have.

A consistency score says: Keep building.

You’re still encouraged to develop a habit, but missing a day no longer feels catastrophic.

Why this works: It preserves the motivating signal (your engagement pattern matters) without the loss aversion trap. The reference point shifts from “don’t lose what I have” to “keep the momentum going.”

2. 🎛️ Let people choose the kind of learning they have capacity for

One thing Duolingo rarely acknowledges is that not every day looks the same. Some days you have 20 minutes and genuine focus. Some days you have 3 minutes on a bus back home from a social event you never wanted to go to.

Today, both behaviours count equally.

I’d rather ask the user what they’ve got capacity for. Maybe today’s options are:

⚡ Quick Practice (<5 mins): “I’ve only got a few minutes.” — review and reinforcement, low cognitive load

🧠 Deep Learning (5–15 min): “I’m ready to tackle something new.” — new material, harder exercises, more XP reward

🚀 Challenge Me (open-ended): “I actually want to push myself today.” — for days when you’re genuinely interested in pushing further

The important part isn’t the labels — but the choice.

Research around Self-Determination Theory consistently shows that people stay motivated for longer when they feel they have meaningful autonomy over how they engage with a task. Giving people options acknowledges that learning isn’t one-size-fits-all.

It also makes the app feel like it’s working with you instead of on you.

3. 🔁 Reward coming back instead of punishing absence

This is probably the change I feel strongest about. Current Duolingo treats absence as failure.

Your streak disappears, you get guilt-driven notifications. Sometimes you’re offered a streak freeze to soften the blow.

I’d invert that entirely because we already have enough anxieties in life, I am not sure we need more. So if someone comes back after a few days away, that’s a success — they’re returning. Help them. Instead of highlighting what they lost “you lost your 47-day streak,” give them a short “getting back into it” session that revisits the concepts they’re most likely to have forgotten. Same behaviour (returning after a gap), completely different emotional framing.

Celebrate the return. Reduce the friction.

4. 📊 Make progress honest

XP is satisfying. It’s also incredibly easy to mistake for learning. I’d make mastery the primary measure instead.

Imagine opening Duolingo and seeing something like:

Vocabulary: ~850 words Grammar: Lower Intermediate Listening: Improving

It’s not as flashy as a giant XP number. BUT it’s much more useful.

It answers the question I actually care about every time I finish a session: Am I getting any better? Because that’s why I downloaded the app in the first place.

🧱 What Actually Changed?

I realized while putting this together: None of these ideas are particularly radical.

I’m not suggesting Duolingo remove gamification, or stop caring about retention, or delete the owl (Oh please don’t!).

The mechanics are mostly familiar. What’s different is the direction they’re pointing.

Instead of asking every feature to maximise tomorrow’s login, they’re all trying to answer a different question: How can we help someone keep making meaningful progress?

I know that feels like a small shift. But I don’t think it is. Everyone learns differently, and there is no one-size fits all, but if there is a common thread I believe works for everyone, that is empathy.

I feel the most valuable thing a learning system can do is make itself unnecessary — to build the internal motivation to learn to the point where you don’t need a Duo scowling at you to come do your lesson.

References

Deci, E. L., Koestner, R., & Ryan, R. M. (2001). Extrinsic rewards and intrinsic motivation in education: Reconsidered once again. Review of Educational Research, 71(1), 1–27. https://doi.org/10.3102/00346543071001001

Deci, E. L., & Ryan, R. M. (2000). The “what” and “why” of goal pursuits: Human needs and the self-determination of behavior. Psychological Inquiry, 11(4), 227–268.

Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47(2), 263–292.

Mekler, E. D., Brühlmann, F., Tuch, A. N., & Opwis, K. (2017). Towards understanding the effects of individual gamification elements on intrinsic motivation and performance. Computers in Human Behavior, 71, 525–534.

Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist, 55(1), 68–78.

Tversky, A., & Kahneman, D. (1991). Loss aversion in riskless choice: A reference-dependent model. The Quarterly Journal of Economics, 106(4), 1039–1061.

van Roy, R., & Zaman, B. (2018). Need-supporting gamification in education: An assessment of motivational effects over time. Computers & Education, 127, 283–297.