Skip to content
AppDues
Log in
Contact salesSign up

JTBD paywall optimization: sell the job, not the feature list

Most paywall problems are not really paywall problems. If retention is healthy but few free users subscribe, the app may be describing its features accurately while missing the job users hired it to do. This guide turns Jobs to Be Done research into a practical sequence: spot the signal, interview users well, interpret the findings, improve the path before the paywall, then test the paywall itself.

Why JTBD often beats feature-focused paywalls

A feature list answers “what does the app contain?” A JTBD message answers “what will this help me do or feel?” Those are different questions, and the second one is closer to the moment a user decides to pay.

Growth advisor Léa Samrani, who has worked in product roles at Bumble, Uptime, and Passion and later advised more than 45 apps, told interviewer Daphne Tideman that the largest return came from JTBD work rather than new features, new designs, or retention tweaks. The reason is simple: a good product can still fail to communicate the problem it solves.

JTBD is not a sales trick layered on top of the product. It is a way to reconnect the product’s real value with the people most likely to pay for it.

Captured results from three education apps

The figures below are the outcomes reported in the original 2025 article for Léa Samrani’s client work. They are dated, unverified captured evidence rather than a benchmark you should assume for your own app. Treat them as an illustration of what a sustained JTBD programme produced in those specific cases, and check your own baseline before setting targets.

The three apps sat in the same broad education category, yet each surfaced a different job. That is the central warning: a competitor’s JTBD cannot be copied into your product.

Why founders resist the research

Founders often built the app around a problem they personally know. Being told to step back from that experience and study what paying users actually want can feel like a rejection of the product’s origin.

The second objection is time. Interview programmes can take many hours, and growth pressure makes that feel expensive. The counter-argument is that the learning compounds across the funnel: better onboarding, better timing, better copy, and sometimes better packaging all draw on the same insight.

A useful reframe is that JTBD research does not replace the founder’s problem insight. It tests whether the market describes that problem in the same language and values the same outcome.

Step 1: recognise the signal that points to a JTBD problem

The first step is diagnostic. Look for a pattern where the product is clearly useful but the commercial front end underperforms.

Smart Tales, an educational app for children, had strong retention and, at the time of the work, a 4.6 Apple App Store rating from more than 11,000 reviews. Its front-of-funnel conversion was lagging. The team was emphasising lesson and content quality, but the interviews later showed that the core job was different from “teach maths, science, and reading.”

A practical signal checklist:

  • Retention is solid, so the product delivers something people keep using.
  • Free-to-paid conversion is weak relative to that retention.
  • Click-through or paywall engagement is also soft.
  • Marketing and paywall copy lean heavily on features and content volume.

If several of those are true, the problem may be message-market fit rather than product quality.

Step 2: practise interviews before running the real ones

Support logs, reviews, and usability tests are useful, but they tend to be feature-shaped. They rarely reveal the underlying job. Well-run user interviews are the hardest part of the workflow, and the recommendation is to rehearse before speaking to real customers.

A rehearsal method that avoids bias:

  1. Do not interview close friends or family, whose context you already know too well.
  2. Interview a colleague or acquaintance about an unrelated purchase.
  3. Pick something you know nothing about, so you cannot fall into assumption mode.

In one rehearsal, Samrani was interviewed about her living-room table. What looked like an ordinary purchase turned out to carry meaning about buying her first home. That is the depth a good interview can reach when the interviewer has no stake in the answer.

Step 2.5: write questions that reach past the literal answer

Samrani’s favourite question is: “What do you think the app was going to enable you to do?” The aim is to learn what the user expected to feel and why they bought, not to confirm a feature list.

A cautionary example: when asked why they bought a VPN, one customer answered “security.” That is the surface answer. The useful follow-up is what security was supposed to make possible, such as peace of mind about personal data.

If an interviewee struggles, shift to the buying journey. Ask when and where the decision happened, what they were worried about, and what they hoped would change. Context helps memory.

A small field map for turning notes into a testable job:

  • Trigger: the moment the user decided to look for a solution.
  • Expected outcome: what they thought the product would enable.
  • Emotional payoff: how they wanted to feel afterwards.
  • Alternative considered: what they would have done instead.
  • Evidence quote: the user’s own words, kept verbatim.

If you are nervous, run the interview with a second person who can take notes or prompt you to go deeper.

Step 3: analyse interviews together, not by handing them to a tool

It is tempting to paste interview notes into an AI tool and ask for the JTBD. The recommendation is to resist that shortcut. Interviews contain subtleties and app context that automated summaries can miss.

The most important nuance: the right JTBD is not necessarily the one mentioned most often. It is the one your app can solve better than competitors can.

For Smart Tales, the finding was unexpected. The job was not teaching maths, science, and reading. It was helping parents cope with the screen switch-off moment. The app gives children non-addictive screen time, so ending screen time is less painful, and the time spent is enriching either way.

As a job statement: a parent wants screen time that guarantees their child is learning something, so the child avoids mindless scrolling and screen addiction.

The problem being solved was screen addiction, not curriculum coverage. That reframing became the basis for the paywall work.

Why you cannot copy a competitor’s JTBD

The other two education cases had different jobs. For the teenage audience, the job was passing exams and building confidence about reaching a dream university. For the community learning app, the job was belonging to a community rather than learning itself.

Similar categories can hide very different motivations. The differentiation comes from the nuance of your language and from showing precisely how the app helps a user complete their job.

Step 4: improve the path before the paywall

Paywall experiments often focus only on the paywall screen. In Samrani’s experience, the biggest impact frequently happens earlier, where the largest volume of users passes through, provided retention is already healthy.

Onboarding is the main lever. Smart Tales and the teenage app had both sent users straight to the paywall with little context. The teams tested adding positive friction: helpful context and questions that guide users to see how the app will help them complete their job.

Positive friction versus negative friction:

  • Positive: questions and context that help the user connect the app to their goal.
  • Negative: questions that serve the business, such as “Where did you hear about us?” or “How old are you?”

One technique Samrani likes is showing a review that reflects the core job, then asking subtly, “Can you relate?” It builds trust and signals understanding, especially when the same message continues through the journey. Headway, a book-summary app, uses social proof in onboarding to speak to the problem of not having enough time to read full books.

Timing matters too. For the community learning app, the aha moment arrived after six lessons, enough for users to feel they could learn consistently and rely on the app to structure progress. Showing the paywall later produced better results even though fewer users saw it early.

Expect a series of experiments rather than one big test. Combining larger changes, such as onboarding friction, with smaller ones, such as App Store copy and screenshots, produced the strongest results in Samrani’s work.

Step 5: optimise the paywall itself

Once the earlier stages carry the message, the paywall can become simpler. The community learning app tested changing paywall copy from feature-focused to JTBD-focused. Many paywalls list endless features instead of explaining the value behind them.

A counter-example from the original article is Uxcel Go, a UX learning app whose paywall is entirely feature-focused. Its closest line to a job, “Invest in your design career today,” stays vague: invest why, and to achieve what?

Smart Tales’ original headline, “Accelerate your kid’s growth,” was closer to a job but still not the core job found in research. After multiple experiments, the final paywall carried much less copy because onboarding did the explanatory work. The visual also shifted from functional to emotional.

A quick paywall copy audit:

  • Does the headline name an outcome or a feature?
  • Would a user recognise their own words in the copy?
  • Is the social proof tied to the job rather than generic praise?
  • Does the visual support the emotional payoff?
  • Is the message consistent with onboarding and the store listing?

JTBD research and your pricing and packaging

Sometimes the research shows that pricing or packaging needs to change, not just the message. Adding a package is usually easier to approve internally than changing or removing one. Samrani recommends starting with a lower-friction experiment that can win buy-in quickly, then moving to harder, longer tests.

Two considerations:

  • You may be underpricing. A higher price can be tested earlier, because a JTBD message can support more value without changing the product.
  • Not every test wins. JTBD-led experiments have tended to outperform average tests, but the first attempt may miss. Avoid tests that are too small: a result with no significant difference teaches less than one that shows a slight drop.

Keep testing and keep researching. The case examples took roughly 10 and 20 experiments respectively to reach their reported gains.

What sustained commitment looked like

The hard part is usually the research, not the experiment mechanics. Samrani’s advice is to do one form of JTBD research well rather than many forms superficially.

Smart Tales committed to the process. The team later credited “tens of customer interviews to truly understand the needs of parents and children” and “hundreds of A/B tests to optimize every aspect of the app’s performance” among the reasons for reaching $1M in ARR, a milestone the article says only 0.5% of apps reach.

If you want to optimise a paywall with JTBD, start before the paywall. Prepare for strong interviews, analyse them deeply, then work back from the front of the funnel, testing and refining as you go.

Frequently asked questions

How do I know my paywall problem is a JTBD problem?

Look for strong retention alongside weak free-to-paid conversion and often weak click-through. That combination suggests users value the product but the commercial message is not connecting with the job they hired it for.

How many user interviews do I need before changing the paywall?

The original article does not give a fixed number. It stresses quality over quantity: rehearse first, interview people outside your existing circle, and analyse the notes together. Smart Tales later credited “tens of customer interviews” as part of its process.

Can I use AI to analyse interview transcripts?

AI can help with some analysis, but the recommendation is to discuss findings as a team. Interviews contain subtleties and app context that automated summaries can miss, and the right job is the one you can solve better than competitors, not simply the most-mentioned theme.

Should I show the paywall earlier or later?

Show it when users begin to feel their job is being met. In the community learning app, the aha moment came after six lessons, and moving the paywall later produced better results even though fewer users saw it early.

Can I reuse a competitor’s JTBD messaging?

No. The three education apps in the article each had a different job: screen-time management, exam confidence, and community belonging. Similar categories can hide different motivations, so the language has to come from your own interviews.