Skip to content
Idea Atlas
Sign in

TripDesk

As described:StopNarrowed:Narrow

No as consumer. Go only as the B2B tool for travel pros.

travel advisors · $58 blended · last reviewed (next review by 28 Dec 2026) · sources not yet verified

Save
Share

← All research

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

This pre-mortem is for the idea as described (consumer, end to end, with guides). Causes are ranked by probability.

Cause #1: Nobody came back (probability ~50% of failure cases)

What happened. A launch push brought 30,000 signups. People planned a trip, some booked through affiliate links, and they left. Their next trip came 5–8 months later, and by then they planned it in ChatGPT, Google or MakeMyTrip, or on Instagram. Monthly actives decayed to 5% of signups. Commission revenue per trip was about $10–30, and paid acquisition cost $3–15 per install with a small share booking. CAC never paid back.

Early warning signs

  • 30-day retention below 10%, and 90-day retention below 5%.
  • Under 20% of planned trips produce any affiliate booking.
  • Commission per active user per year below your cost to acquire them.
  • Growth only when you spend on ads.

Fixable? Not as a consumer product. Frequency is set by how often people travel, and you can't change that. It's fixable only by changing the buyer to someone who plans trips daily (advisors, agencies, operators).

Cause #2: Free AI planners made it pointless (probability ~30%)

What happened. Google put Gemini trip planning into Search, Maps and Gmail. ChatGPT added booking apps from Expedia and Booking.com. Booking, Expedia, Tripadvisor and MakeMyTrip shipped their own assistants tied to loyalty points. Your itineraries were no better, your data was worse, and users couldn't earn points with you.

Early warning signs

  • Users say "ChatGPT does this" in interviews.
  • Your conversion to booking falls as incumbents launch features.
  • Differentiation relies on the AI model rather than on data or workflow you own.

Fixable? Partly. Own things the giants don't: a professional's workflow (proposals, supplier library, branded client app), niche constraints (dietary, accessibility, pilgrimage), and in-trip service tied to a human advisor.

Cause #3: A trust or liability event (probability ~20%)

What happened. One or more of:

  • the AI sent a family to a closed restaurant, a permanently shut attraction, or a wrong ferry time, and it went viral;
  • a guide matched through the app behaved badly with a solo traveller;
  • bundling flight and hotel made you legally the package organiser when an airline collapsed;
  • stored passport scans and travel histories leaked.

Early warning signs

  • Support tickets about wrong or outdated information above 2% of trips.
  • Guide complaints, no-shows, or any safety incident.
  • Legal advice says "you're a package organiser" or "you need seller-of-travel registration".
  • Storing documents without encryption, retention limits or a DPDP/GDPR basis.

Fixable? Yes, by design:

  • ground AI in verified data, and put a human advisor in the loop;
  • link out to guide marketplaces instead of running one;
  • deep-link to booking sites, and never bundle-sell;
  • store the minimum personal data.

Honourable mentions

  • API and data costs. Flight status, places and hotel content APIs charge per call. Unmetered free users make costs grow faster than revenue.
  • Seasonality. Travel demand is seasonal (holiday peaks). Revenue and cash can swing 2–3x across the year.
  • Founder-market fit gap. Without someone who has worked at an agency, OTA or tour operator, you'll build for how travel should work, not how it does.