How Failure Forecast Analyses Business Ideas

What a business pre-mortem is

A pre-mortem is the opposite of a post-mortem. Instead of explaining a failure after it happens, you assume the failure has already happened and work backwards to explain it. The question changes from “could this fail?” — which invites a defensive answer — to “it failed; what was the most believable reason?”

That reframing matters because it removes the need to defend the idea. Once failure is assumed, the honest reasons surface quickly, and most of them turn out to be assumptions nobody had tested.

How your idea description is interpreted

The analysis starts from the description you write: what the product or service is, who it is for, how it makes money, and what stage you are at. That text is read as a set of implicit claims — about the customer, the problem, the price, the channel and the cost of delivery.

Detail changes the output. A vague description produces generic risks, because there is little to falsify. A specific description — named customer, named channel, real price, real cost — produces specific risks. Nothing is added from outside the description except general reasoning about how businesses of that shape usually fail.

The risk categories assessed

• Demand — whether the problem is urgent enough for anyone to change behaviour.
• Willingness to pay — whether the people with the problem are the people with the budget.
• Distribution — whether you have a repeatable way to reach buyers at a cost the price supports.
• Competition and alternatives — including the status quo, spreadsheets and doing nothing.
• Unit economics — whether the maths survives real delivery costs, refunds and churn.
• Execution and capacity — whether the plan fits the time, skills and money actually available.
• Legal, platform and dependency risk — rules, licences and single points of failure you do not control.
• Timing and market shift — including how quickly AI and no-code tools can erode a thin advantage.

How weak assumptions are identified

Each risk category is turned into the claim your idea needs to be true. An assumption is flagged as weak when it is load-bearing (the idea does not work without it), untested (no observed behaviour supports it yet), and cheap to check relative to the cost of being wrong. Assumptions that are true-but-irrelevant are deliberately not promoted, because chasing them wastes the time you have.

How likelihood and consequence set severity

Severity is not the same as scariness. Every identified risk is weighed on two axes: how likely it is to occur given what you described, and how much damage it does if it occurs. A likely annoyance ranks below an unlikely event that ends the business. Risks that are both likely and terminal are the ones surfaced first, because those are the ones worth testing before spending.

What the survival score means

The survival score is a summary of how much of the idea currently rests on untested, load-bearing assumptions. A low score means the idea depends heavily on things nobody has verified yet. A higher score means the critical claims are either evidenced or cheap to recover from.

What the survival score does not mean:

• It is not a probability that your business will succeed.
• It is not a valuation, a market size or a revenue forecast.
• It is not comparable across ideas as a ranking — two ideas with the same score can carry entirely different risks.
• It is not a verdict. It is a measure of how much you still do not know.

What warning signs are

Warning signs are observable events that indicate a risk is turning real — a landing page with traffic and no signups, interviews where everyone is polite and nobody follows up, a customer acquisition cost climbing above the margin. They are written to be noticeable early, while the decision is still reversible.

What kill criteria are

Kill criteria are thresholds you set in advance that would make you stop or change direction: a number, a timebox and a decision. “If fewer than 5 of 40 qualified people pre-pay within two weeks, I stop and rework the offer.” Their value comes from being written down before you are emotionally committed, when you can still be honest about what would change your mind.

How the next tests are selected

Tests are chosen by the ratio of information gained to money spent. The highest-severity untested assumption is targeted first, with the cheapest test that could genuinely falsify it. A test that cannot produce a “no” is not a test — so preference goes to methods that require some commitment from the other person: pre-orders, deposits, paid pilots, booked calls, waitlists with payment details.

Limitations of model-generated analysis

• The analysis reflects general patterns of business failure, not your specific market. It can be confidently wrong.
• It does not use proprietary market research, private datasets or live market data unless you explicitly supply that information in your description.
• It has no access to your customers, your competitors' internals or your local conditions.
• Output quality tracks input quality: vague descriptions produce generic risks.
• It cannot guarantee whether a business will succeed or fail. No tool can.
• Outputs are decision support — not financial, legal, tax or other professional advice. Where the stakes are high, verify with a qualified professional.

See it applied

The clearest way to judge the method is to read a finished report. You can see a full failure forecast end to end, or follow the complete business idea validation framework to run the same reasoning by hand.