A feasibility study is usually written by people who want the project to go ahead. That does not make it dishonest. It does mean the document is an argument for a decision, and arguments should be tested before money moves.
Stress-testing is the discipline of trying to break the case on paper, so that reality does not do it for you later. This article sets out a practical sequence for founders, executives and investors reviewing a feasibility study, whether it is yours or someone else's.
Why feasibility studies fail quietly
Most weak studies are not wrong in obvious ways. The arithmetic adds up and the formatting is tidy. The problems sit underneath:
- Assumptions that are stated once and never questioned.
- A single base case presented as the forecast.
- Demand figures taken from market reports rather than from buyers.
- Costs built from quotes that exclude delays, rework or ramp-up.
- Funding plans that assume everything arrives on time.
The goal of a stress test is not to produce a pessimistic version of the study. It is to find out which parts of the case carry the weight, and whether those parts can bear it.
Start with the decision, not the model
Before opening the spreadsheet, write down the decision the study is meant to support. For example: commit a fixed amount of capital to build a facility, enter a new market, or acquire a business.
Then agree what 'feasible' means in your context. It might be a minimum return, a payback period, a ceiling on peak funding, or a limit on how much you can lose without threatening the wider business. Without this, a stress test has nothing to measure against, and every result can be explained away.
Step 1: Audit the assumptions
List every assumption that drives revenue, cost and timing. Most models have far more than the summary suggests. For each one, record three things:
- Source. Is it a contract, a quote, a pilot result, a benchmark or a judgement call?
- Owner. Who is accountable for it being right?
- Impact. If it were wrong, how much would the outcome change?
You will usually find that a handful of assumptions decide the result. Typical candidates are price, volume, conversion or utilisation, input costs, time to first revenue and capital cost. Spend your effort there, not on the long tail.
Be wary of assumptions that are backed only by the author's confidence. 'We expect to win a fair share of the market' is not evidence. A signed letter of intent, a paid pilot or a repeatable sales conversation is.
Step 2: Test the model's mechanics
A sound case can still sit inside a faulty model. Check the structure before you trust the outputs.
- Do inputs live in one place, clearly separated from calculations?
- Do the balance sheet, income statement and cash flow reconcile?
- Are timing conventions consistent, such as monthly versus annual, or cash versus accrual?
- Do growth rates, working capital and depreciation behave sensibly in later years?
- Is the terminal value, if one is used, doing most of the work?
A useful habit is to change an input to an extreme value and see whether the model responds sensibly. If doubling the price barely moves revenue, or setting volume to zero still produces profit, something is wrong.
Step 3: Run sensitivity and scenario tests
This is the core of the exercise. Use three layers.
One variable at a time
Move each key assumption up and down by a range that is plausible, not just convenient. Note which ones change the outcome most. Rank them. The ranking tells you where to gather better evidence.
Combined scenarios
Real problems rarely arrive alone. A delay in opening often comes with higher costs and slower early sales. Build a downside scenario in which several things go moderately wrong together, and a severe scenario in which they go badly wrong. Ask whether the business or the investor survives each one, and on what terms.
Reverse stress test
Instead of asking 'what is the return if volume falls 20%?', ask 'what would have to be true for this project to lose money, or to breach our funding limit?' Then judge how likely that is. Reverse tests often expose conditions that sound unlikely in theory but are common in practice, such as a single customer leaving or a licence taking longer than planned.
Step 4: Check the evidence behind demand and cost
Demand and cost are where optimism tends to collect.
For demand, look for evidence from actual buyers: orders, deposits, pilots, repeat purchase behaviour, and the reasons people give for not buying. Top-down market sizing may be fine for context, but it is a weak basis for a revenue forecast. Ask how the study gets from a market to a customer, and how long that takes.
For cost, ask what is excluded. Common gaps include contingency, commissioning and ramp-up, hiring lead times, compliance, currency movements, and the management time the project will absorb. If the study covers more than one country, check that local costs, taxes and regulations have been considered by someone who knows them. Tax and legal points should be confirmed with qualified advisers.
It also helps to compare the plan with similar projects that have already happened, including ones that went badly. Overruns in comparable projects are a better guide than a single optimistic quote.
Step 5: Examine funding, timing and cash
Projects that look viable on returns can still fail on cash. Review:
- Peak funding need, and when it occurs.
- How much headroom exists if revenue starts late.
- Debt terms, covenants and what happens if they are breached.
- Whether further capital is assumed but not secured.
- Working capital, especially if customers pay slowly or suppliers want upfront payment.
A useful question is: how many months of delay can the project absorb before it needs new money? If the answer is 'none', the plan has no margin.
Step 6: Define kill criteria and stage the commitment
A stress test should change how you commit, not just whether you do. Where possible, break the investment into stages with clear checkpoints.
- What evidence must exist before the next tranche is released?
- Which results would cause you to pause, change course or stop?
- Who decides, and by when?
Agree these criteria before the money is spent. It is much harder to stop a project once people, contracts and pride are involved. Staging also gives you a cheap way to test the weakest assumptions first.
Questions to ask the author
Whether the study is internal or from an outside party, these questions are worth putting to the author:
- Which three assumptions would change the decision if they were wrong?
- What evidence do we have for each, and how recent is it?
- What is the downside case, and can we survive it?
- What did you leave out, and why?
- What would make you recommend we do not proceed?
A good author will answer these without defensiveness. Vague answers are information in themselves.
What to do next
Take the feasibility study you are relying on and do three things this week. List the five assumptions that matter most and mark the evidence behind each. Ask for a combined downside scenario and a reverse stress test. Then write down the conditions under which you would pause or stop.
If you would like an independent view on a study, a model or an investment case, you can contact GainWells to discuss it. For general investment decisions, speak to a qualified financial, legal or tax adviser as well.