Skip to content
Idea Atlas
Sign in

BhashaCity

As described:StopNarrowed:Continue

No as described. Go as work English for Hindi/Bengali-medium learners.

language learning · ² At a blended ~$2.50 a month per payer: ₹199/month and ₹1,199/year plans, plus some ₹399 Pro, at ₹88 = $1. Institute seat licences (from month 9) reduce the number of individual payers needed; see plans/11. · last reviewed (next review by 28 Dec 2026) · 60 cited sources

Save
Share

← All research

5. Pre-Mortem: It's September 2028 and the Language App Is Dead

This pre-mortem covers the idea as described: three languages, official and casual goals, a gamified city map and a voice companion, for everyone. Causes are ranked by probability.

Cause #1: "Duolingo with a map". Nobody paid for month 2 (probability ~35% of failure cases)

What happened. The city looked good and the first scenes were fun, and signups were fine. Then learners noticed that the free Duolingo on their phone already had scenes, characters and speaking tasks. Speaking Adventures became free in 2026, and Video Call moved into the cheaper tier. Learners couldn't see their own speaking improve, so after a week or two there was no reason to pay ₹199 again. Education apps have the weakest annual renewal of any category, 24% (RevenueCat 2026), and the app sat right in that pattern.

Early warning signs

  • Day-7 return below 20%, day-30 below 8%.
  • Month-2 renewal below 45%.
  • Median voice minutes per payer fall below 15 a week after week 3.
  • Reviews say "fun but I'm not improving" or "same as Duolingo".

Fixable? Yes, by reframing. Make the paid core work situations (interview, customer call) that Duolingo doesn't cover. Place by speaking. Show a weekly speaking score with before/after recordings. The failure mode is shipping the map and the voice without the proof of progress.

Cause #2: Customer acquisition cost higher than lifetime value at Indian prices (probability ~35%)

What happened. Search terms like "English speaking app" were owned by Duolingo, SpeakX and Stimuler. The category buys growth: Seekho spent ₹134 Cr on ads for ₹142 Cr of FY25 revenue, and Kuku FM spent ~₹285 Cr on marketing. A ₹199 subscriber who churns in 3 months is worth about ₹500 after fees, and a paid install that converts at ~2% cost more than that. Without an audience or an institute channel, growth stalled below a few hundred payers.

Early warning signs

  • Paid CAC above 2 months of ARPU.
  • Under 20% of new users from shares or referrals.
  • Organic signups flat after month 4.
  • No institute willing to pilot by month 9.

Fixable? Partly.

  • Make the speaking-score card and the scene challenges shareable.
  • Run a daily short-video routine in Hindi or Bengali, generated from the scene library.
  • Lead with Bengali-medium learners, where competition is thinner.
  • Add institute seat licences, which bring 50–500 learners per sale with no ad spend.

Cause #3: The build ate the timeline, and voice cost ate the margin (probability ~30%)

What happened. One or more of these:

  • Six months went on the city, avatars and three language curricula before anyone paid.
  • The voice companion used a premium real-time model. Daily users on a ₹199 plan cost ₹300+ a month each, so every engaged learner lost money.
  • Bengali speech recognition misheard learners, so the feedback was wrong, and Bengali-medium users (the chosen gap) left first.
  • Caching "whole conversations" didn't work, because every learner's replies differed.

Early warning signs

  • No paying user by day 45.
  • Voice and model cost above 25% of revenue.
  • Learners report "it didn't understand me" in more than 5% of sessions.
  • Bengali sessions have a lower completion rate than Hindi or English ones.

Fixable? Yes, by design.

  • Sell with a WhatsApp concierge first.
  • Build a 2D map, not a game.
  • Use the "scripted spine, live edges" scene engine with cached audio.
  • Cap voice minutes per plan and route to cheaper Indic-capable stacks.
  • Test Bengali accuracy on real learner recordings before promising it.

Honourable mentions

  • SpeakX ships Bengali well. It announced Bengali, Telugu, Tamil and Marathi in Oct 2025. The defence is depth in work scenes, the institute channel, and speed. Bengali-first alone isn't a moat.
  • Privacy or safety incident. Stored voice clips leak, a minor signs up and parental-consent rules apply, or a "companion" chat drifts into personal territory. Delete audio after scoring, allow 18+ only, and use a coach persona.
  • Outcome promises. "Get a job in 30 days" marketing draws refund demands and complaints. Promise practice and a score, not a job.
  • Missing teaching expertise. Without an experienced trainer, the levels and feedback feel arbitrary to learners who've been to coaching classes.