INDEPENDENT PRODUCT STRATEGY CASE · 2026
PRODUCT STRATEGY / SPEAKING PRACTICE
Duolingo↗
Speaking Confidence Path
A product hypothesis for moving from short lessons to more confident speaking practice.
Independent caseProduct discoveryUnit economics
PORTFOLIO CASE · NOT WORK FOR DUOLINGO● ● ●
03
MISSION
SPEAKING PATH · 03:00DAt a café
Order a coffee and ask whether oat milk is available.
“Hi! Could I have a cappuccino, please?”
✦ Strong attempt
Conceptual scenario visual, not a real Duolingo interface.Your response is clear. Try adding a polite follow-up.
PORTFOLIO CASE · NOT WORK FOR DUOLINGODISCOVERY / EXPERIMENT / ECONOMICS01 / 08
COMPANY CONTEXT · Q3 2026
Duolingo's reported figures. Company metrics are not outcomes of this product hypothesis.
01 / PROBLEM01
Lots of practice in an app, but can learners speak without prompts?
The case starts with a working hypothesis: some learners may complete short structured exercises yet struggle to produce an answer independently in a real situation. The frequency and severity of this gap must be validated before development.
Motivated A2-B1 learners with an existing study habit and basic vocabulary but limited confidence in spontaneous speech. This is a research hypothesis, not an official Duolingo segment.
“When I need to speak a language I’m learning, help me practise in manageable steps and understand what to improve so I can feel more confident in a real conversation.”
02 / DISCOVERY
First find where the path to speaking breaks down.
This case has no access to Duolingo’s internal data, completed user interviews or launched experiments. Personas and barriers are hypotheses; the plan below is for validation, not research findings.
Interviews
Semi-structured conversations about a recent real speaking experience, workarounds and reasons for skipping practice.
Behaviour
Review starts, completions, retries, retention and differences across learner cohorts.
Synthesis
Identify frequent, meaningful barriers and the segment for whom this is a priority.
03 / USER JOURNEY
Design a short path from entry to another attempt.
Trigger
I want to speak more confidently but don’t know where to start.
Entry
Show the topic, goal and time needed.
Practice
Give a short context and a chance to answer out loud.
Feedback
Highlight a strength and 1-2 specific next steps.
Retry
Let me try again and see what changed.
Return
Suggest a next step and make progress visible.
04 / PRODUCT SOLUTION04
Speaking Confidence Path
A short guided scenario on a familiar topic: hear a prompt, respond out loud, receive specific supportive feedback, retry and see a next step. The first version tests the value of this loop instead of building an open-ended AI tutor.
Scenario
A familiar topic from previously studied material.
Voice response
Record an answer with a simple exit and retry path.
Feedback
One strength and a concrete suggestion.
Another attempt
Retry or finish with a clear next step.
05 / METRICS
Measure completed practice and quality, not microphone taps.
NORTH STAR HYPOTHESIS
Weekly Meaningful Speaking Learners
Unique learners per week who complete at least two speaking missions and receive feedback. The two-mission threshold is an operational proposal for testing.
Share of users who repeat a speaking mission within 7 days of their first completed mission.
Core lesson completion and overall retention.
Recognition errors, feedback usefulness and drop-off after feedback.
Response latency and AI cost per completed mission.
06 / PRIORITIZATION06
Start with a guided mission and a low-cost demand test.
RICE compares initiatives over a consistent horizon. The scores below are illustrative assumptions from the case, not Duolingo analytics or a final decision.
INITIATIVERICERATINGS R / I / C / E
RICE = Reach × Impact × Confidence ÷ Effort
07 / EXPERIMENT
Check repeat use before scaling.
Current learning experience.
A short path with a prompt, voice answer and feedback.
A proposed A/B test randomised at user level. The 2-4 week period is a planning assumption; the actual duration depends on eligible traffic and power analysis. Define the primary metric and guardrails before launch.
08 / UNIT ECONOMICS08
The base scenario does not break even yet.
Scenario model: 100,000 eligible users, 30% adoption, 7% paid conversion among adopters and $5 incremental monthly revenue per payer. This yields $10,500 revenue against $75,000 included costs and -$64,500 modeled monthly contribution.
Revenue$10.5k
AI cost-$40k
Other variable-$20k
Fixed cost-$15k
Contribution after cost-$64.5k
Recommendation
Do not move straight to a broad launch. First validate the problem and test value with a fake door, prototype or concierge approach. Build an MVP only if users return, feedback helps, guardrails hold and cost per completed practice supports viable economics.
This is an independent academic portfolio case based on public materials and product analysis. No interviews, A/B test or launch of Speaking Confidence Path were conducted inside Duolingo. User segments, RICE, timelines, product effects and financial model are hypotheses or assumptions, not company data.
SOURCES: DEEP RESEARCH, PRACTICAL SUPPLEMENT, PRESENTATION AND UNIT-ECONOMICS MODEL