5. Pre-Mortem: It's September 2028 and SafeLog Is Dead
Three most likely causes of death, ranked by probability. They compound: #1 usually causes #2.
Cause #1: No way in (probability ~45% of failure cases)
What happened. The founder built a clean evidence tool in eight weeks and then spent a year trying to get Alberta contractors to look at it. Consultants were polite and kept their Word templates. The certifying partner didn't answer emails. SEO pages ranked, but a market of ~11,000 companies produced 40 visits a month. By the second audit season there were 9 paying accounts, most of them one consultant's clients.
Early warning signs
- By 30 Oct 2026, fewer than 5 paid audit sprints and no consultant pilot.
- Every warm lead comes from one person.
- No certified COR auditor has reviewed your element mapping.
- You're writing code on weekdays in Mountain Time instead of being on calls.
Fixable? Yes, but only at the start. It takes a paid insider (an auditor or consultant as advisor or partner) before building. Without one, the right move is a plain no, which is why the odds halve without condition 1 (WHAT-TO-DO-AND-AVOID.md).
Cause #2: A once-a-year tool (probability ~30%)
What happened. Employers signed up in November, used the product hard for six weeks, submitted their audits in January and cancelled in February. Monthly plans made this easy. Use from March to September was near zero, because the paper habit on site never changed. Annual renewal ran at 45%. Every autumn the founder had to resell the same customers.
Early warning signs
- Monthly active use below 40% of paying accounts in April–August.
- Fewer than 60% of accounts open the monthly gap report.
- Most records are uploaded in batches (scanned paper) rather than captured on site.
- More than 30% of employers on monthly plans.
Fixable? Partly. Annual plans tied to the audit year, certificate-expiry alerts, the monthly gap report and consultant accounts (which don't churn after one audit) all help. If renewal is still below 70% at the second season (month 16), the business is a seasonal service, not a subscription. Price it that way or stop.
Cause #3: Someone else owned the mapping (probability ~25%)
What happened. One of three things. (a) SiteDocs, Safety Evolution or another Canadian tool already had, or shipped, element-by-element evidence mapping and sold it to the same consultants with a bigger team. (b) The certifying partner released a new audit tool, and the mapping had to be rebuilt in the middle of audit season. (c) The certifying partner objected to a third party referencing its audit instrument, or to the use of "COR" in marketing.
Early warning signs
- In discovery calls, consultants say "SiteDocs already does that" and can show you.
- A new audit tool version is announced for the next cycle.
- No written answer from ACSA on what third parties may reference.
Fixable? Partly. Research (a) in week 1, before building (10.4). Store the mapping as data, so a new audit tool is a config change, not a rebuild. Get (c) in writing before launch. If (a) is true and the incumbent is good and affordable for small firms, stop.
Honourable mentions (less likely, still real)
- A serious injury at a customer's site where your records were incomplete. Even without legal fault, one story in a small market hurts. Clear terms: the employer owns its safety program; the product organizes records.
- Program change. WCB-Alberta revises PIR refunds, and the money reason to hold a certificate shrinks. Bids still require it, but urgency falls.
- The seasonal crunch breaks the founder. Oct–Jan is when every customer needs help at once. At ~15 hrs/week, onboarding more than ~5 new accounts a month in that window isn't realistic.
- Data-residency objection. A consultant's large client asks where the data lives and gets the wrong answer.