5. Pre-Mortem: It's September 2028 and RitualBot Is Dead
Three most likely causes of death, ranked by probability. They compound: #1 usually causes #2.
Cause #1: The Store never delivered installs (probability ~45% of failure cases)
What happened. The app went live in the Teams Store in month 3. It got 40 installs in its first month, then 15–25 a month after that. Half of the trial requests sat in IT approval queues and expired. SEO pages ranked for "anonymous questions in Teams" but that search is small. By month 12 there were 14 paying organisations and ~$900 MRR. The founder added polls, then kudos, then a Slack version, and none moved the install curve.
Early warning signs
- Fewer than 30 trials a month by month 6.
- Fewer than 50% of trial requests approved by IT within 14 days.
- Store listing impressions flat for 3 months in a row (Partner Center analytics).
- Most paying customers came from your own network, not the Store or search.
Fixable? Partly. You can improve the approval rate (publisher attestation, a security one-pager, a request template) and add a second surface (the Slack port). You can't make Microsoft rank a small app. If the Store doesn't work by month 9, the ceiling is the network you already have.
Cause #2: Episodic use, then churn (probability ~30%)
What happened. Organisations bought it for a reorganisation, a layoff round or one difficult all-hands. Question volume spiked, then fell to zero. Month-3 retention was 60%. Annual plans were rare because the buyer wanted to "see if we keep using it". The app churned 5–6% a month, and new trials only replaced the losses.
Early warning signs
- Fewer than 60% of paying organisations run a question window each month.
- Monthly logo churn above 4%.
- Cancellation reasons say "we don't need it right now".
- Fewer than 30% of new paying organisations choose annual.
Fixable? Yes, mostly. Tie the product to the monthly all-hands cycle (question windows, reminders, the answer archive). Push annual plans with a two-months-free discount. Offer per-event pricing as a pivot so episodic buyers still pay.
Cause #3: The platform or a trust event ended it (probability ~25%)
What happened. One of two things. (a) Microsoft added anonymous posting to Teams channels, or to Copilot or Viva Engage inside Teams, as it did with town-hall Q&A and Praise. Buyers asked "why pay?". (b) A customer's HR team demanded to know who posted a harassing question. A screenshot circulated. Whether or not identity was stored, the story became "the anonymous app isn't anonymous", and cancellations followed.
Early warning signs
- Microsoft 365 roadmap items mentioning anonymous posts in channels, communities or Q&A.
- Any support ticket asking "can you tell us who posted this?".
- Any design shortcut that stores user IDs next to content, "just for debugging".
- Moderation turned off by default to reduce friction.
Fixable? Partly. You can't stop Microsoft. You can pick the part of the job Microsoft does badly: an always-open, moderated, answered-in-public channel with an archive, across meetings, for organisations that don't buy Viva. You can fully prevent a trust event by never storing identity and saying so in writing, before anyone asks.
Honourable mentions (less likely, still real)
- Slack's rules changed again after the Slack port launched: stricter data-use terms, a longer review, or new rate limits. The 2025 changes show Slack will change terms quickly.
- Legal exposure from content. A defamatory anonymous post about a named manager. Moderation by default and clear terms that make the customer the publisher reduce this.
- Founder attention. At ~15 hrs/week, the first unanswered IT security questionnaire loses the deal. A slow reply in month 4 costs more than a missing feature.