Public app analysis
Duolingo: language learning and a broader product
An outside-in analysis of public listing text. This is an independent learning example, not a client case study or a report of measured performance.
What is observable
Checked 29 September 2026 on the US Apple App Store. Title: Duolingo: Language Lessons. Subtitle: Languages, Math, Music & Chess.
The title leads with language lessons, while the subtitle introduces three additional subjects. The listing describes short lessons across languages and these other learning areas. This creates a useful question about which benefit to lead with for different prospective learners.
View the source listing. Store text can change after this observation. Screenshots were not evaluated in this text-focused analysis.
A hypothesis to investigate
Does a language-focused acquisition journey need a narrower first message than a journey aimed at people exploring multiple learning subjects?
This is a proposed question, not an identified conversion problem. Public text cannot reveal keyword coverage, visitor intent, acquisition cost, conversion or subscriber retention. We have no access to this app’s private analytics or experiment history.
What I would test
Compare one focused language-learning proposition with a broader learning proposition for a defined acquisition source. Keep the product promise truthful in both versions. Start with the audience whose intent can actually be identified rather than assigning intent to every visitor.
Write the hypothesis and primary metric before creating the variants. Define eligible traffic, the comparison period, changes that must stay fixed and the decision rule. Review the current platform’s experiment capabilities before selecting a method. A dedicated landing experience can help maintain message continuity, but a before-and-after comparison alone cannot establish causality.
A change in downloads could reflect campaign targeting or seasonality rather than the breadth of the message. Record the source mix, locale and subscription offer alongside the listing version.
Evidence needed before a decision
- Comparable listing traffic and downloads, with the store’s metric definitions.
- Source, territory and device context for the target audience.
- Activation and downstream value where consent and measurement permit.
- A release and campaign log to identify competing explanations.
Keep an inconclusive result visible. If the test cannot separate the variants reliably, retain the learning and change the next question instead of declaring a winner from a small numerical difference.
Apply the method to your app
Create your experiment brief, then read the related strategy guide and execution guide. For a review using your own evidence, explore an ASO audit.