Each week I'll assess a well-known B2C app through the full-funnel lens: what's excellent, what misfires, what I'd test. All public data, and no cheap shots - building products and companies is tough.

Duolingo's headline set: 56.5 million daily users, growing 21% a year. Revenue of $292 million last quarter, up 27%. Profitable, on just $183 million of venture capital ever raised. Share price: down about three-quarters from its 2025 peak. The market isn't pricing the trophy cabinet. It's pricing the machine. And the machine has a funnel story.

What is excellent: onboarding. You complete a real lesson before anyone asks for an email address. Value first, commitment second. Ten years on, and it's still the reference implementation.

Misfire one: the energy system charges for effort, not failure. Every question costs energy, right or wrong, and free users run on a daily cap. Hearts punished mistakes; energy taxes volume. And it taxes the wrong users hardest. For the committed learner the cap is an irritation. For the on-the-fence user, whose habit hasn't set yet, it's a wall that ends the session before the habit can form. Those users don't leave in a churn spike. They slow-bleed out, one capped session at a time. In a reader poll on a Duolingo fan site, 17,000 votes in, about six in ten actively dislike the system.

And the relief valve is the tell: for many free users, the fix is watching an ad. Good monetisation sells access to value you came for. This sells relief from friction the product just manufactured.

The fix: price the premium on added value, not removed pain. And segment it. Casual learners will mostly never subscribe, but they will make one-off purchases with real value in them. Serious learners want serious tools: better learning loops, not fewer roadblocks. Free tier: frictionless core loop.

Misfire two: the stated strategy and the shipped mechanics disagree. Management is spending 2026 chasing 100 million DAUs, absorbing over $50 million in foregone bookings to strip out conversion friction. Meanwhile the energy cap squeezes the exact users that plan needs to win. Misfire one isn't separate from misfire two. It's the mechanism by which it ships. And DAU doth not a successful business make. Just ask TV Time. More on that in the next teardown.

The deeper question: why does the market discount a profitable, growing company this hard? Because what Duolingo sells is being commoditised in real time. Any AI model now teaches a language, adapts, corrects. What AI hasn't cheaply copied is the habit engine and the brand. Duolingo's durable moat is behavioural, not pedagogical. Which makes the energy tax worse than a margin leak: it taxes the one wall still standing.

None of this threatens the empire tomorrow. All of it compounds. The companies that own the joints between engagement, monetisation and retention grow cheaper than the ones that manage them in separate meetings.

If you run growth in a consumer app: where does your monetisation mechanic tax the exact behaviour your retention depends on?

Steelman the bull case

Because it is real: maybe Duolingo becomes the distribution-and-habit layer on top of commoditised AI teaching. Brand plus streak plus a 56-million-user habit is the moat, and AI just cuts the cost of making content. That is probably management's actual bet, and the margin guidance says as much: gross margin is guided down to roughly 69% by year-end as AI-powered features roll out to everyone.

But the bull case only tightens the knot. If the moat is the habit, taxing the habit is self-harm. Even Duolingo's best case argues against the energy system.

One more data point that cuts either way: in January they made Explain My Answer, a marquee paid AI feature, free for everyone. DAU growth stayed on trend. Read it generously and a headline AI feature no longer moves the needle. Read it defensively and the company is commoditising its own paid features before someone else does. Both readings put the value in the same place: not in the AI. In the habit.

Fix the leaks. Fix your world.

Monetisation & Pricing is one of eleven modules in the Revenue Engine diagnostic, and the collision between pricing mechanics and retention loops is precisely the kind of thing it catches. If any of this resonates, our diagnostic maps your entire growth funnel and shows you where the real leverage sits. No deck. No pitch. Just clarity on where the engine is misfiring.

Book a discovery call

Jonathan Stanton-Humphreys is Founder and CEO of Revenue Engine. He spent a decade as a commercial executive in B2C and B2B tech, and built Revenue Engine after watching revenue leak through strategic and implementation dysfunction, and the joins that nobody owned.

FAQ

Why is Duolingo's share price down if users and revenue are growing? Markets price the durability of the machine, not this quarter's metrics. What Duolingo sells is being commoditised by AI in real time, its monetisation mechanics tax the habit loop its retention depends on, and its stated growth-first strategy contradicts its shipped mechanics. Users up and confidence down can both be true.

What is wrong with Duolingo's energy system? It charges for effort rather than failure, and it taxes the wrong users hardest: learners whose habit hasn't yet formed hit a session-ending wall before the habit can set. In a fan-site reader poll with around 17,000 votes, roughly six in ten actively dislike it.

What should consumer apps monetise instead of the core loop? Added value, segmented by intent. Casual users rarely subscribe but will make one-off purchases with genuine value; committed users pay for capability. The free tier's core habit should stay frictionless, because that habit is what everything else compounds on.