One assumption at a time is not the whole downside

Changing one assumption at a time shows what that assumption does to a model. It does not show what happens when two changes arrive together. A business can remain profitable in each separate test and lose money in the combined case.

The separate tests are useful. The mistake is treating them as the limits of what could happen.

Keep one starting point

Consider an invented maintenance service planning one month. It expects 100 identical jobs at $100 each. Variable operating expenses are $60 per completed job, and fixed operating expenses are $3,000 for the month. Those figures include every operating expense in this illustration, including management pay and any depreciation. Revenue and its related expenses are recognized in the same month.

The existing team and equipment can deliver between 80 and 100 jobs at the same service standard without another fixed expense. The comparison holds job mix and variable expense per job unchanged within that range. Its operating result is before financing and income tax; no equipment spending or payment dates have been supplied.

The starting calculation is 100 × ($100 − $60) − $3,000 = $1,000.

Two tests that each stay positive

First, reduce the price to $95 while keeping 100 jobs. Contribution after variable expenses falls to $35 per job. The month leaves $500 after fixed operating expenses.

Then restore the original $100 price before reducing volume to 80 jobs. At $40 contribution per job, the month leaves $200. Each test changes one input from the same starting case.

The Open University's sensitivity-analysis lesson demonstrates that reset when testing individual inputs in an investment model. Here, the calculation concerns one month's operating result rather than a discounted investment value. In both settings, keeping the starting point visible prevents one test from quietly carrying changes from another.

Now put the changes together

Supply a different case: 80 jobs at $95 each. Revenue is $7,600, variable expenses are $4,800 and fixed operating expenses remain $3,000. The operating result is negative $200.

Original monthly model: separate changes and their combination

Scroll across to see every column. Use the arrow keys when the table is focused.

Adding the two separate declines would subtract $500 and $800 from the original $1,000, producing negative $300. That is not the combined result. The $5 price reduction applies to 80 jobs in that case, not 100. The multiplication changes as well.

Put debt dates inside the downside

To examine financed risk, add actual payment dates to a separate case. A positive operating result cannot establish that the money for principal will be available when due.

Consider a different fictional company planning January 2027. It starts with $8,000 of usable business cash; household reserves and money assigned elsewhere are excluded. It pays $7,000 of complete operating costs on January 10, including every worker and all owner work. There are no noncash operating charges. Its base case completes and earns $10,000 of accepted work and collects the full amount January 15. The team can deliver either the base work or the lower-volume case below within those same paid costs.

A completed loan already has its proceeds reflected in opening cash. Its January 20 payment is $3,000: $2,500 principal and $500 interest. There are no additional fees, tax payments, capital purchases, other debt payments or other cash events through January 31. No new borrowing, household contribution or refinancing has been committed.

The base schedule leaves $1,000 after January 10 costs, $11,000 after January 15 collection and $8,000 after the debt payment. Its $3,000 operating result differs from the amount available after financing payments; principal is not another operating expense.

Now supply two changes together: only $8,000 of work is completed and earned, and its entire collection moves to January 25. Costs and debt dates stay unchanged. The operating result is still positive $1,000 before financing, or $500 after the stated interest. But on January 20 the company has only the $1,000 left after operating payments. The $3,000 payment is $2,000 short.

Reducing receipts alone while retaining January 15 collection would leave $6,000 after debt service. Moving collection to January 25 creates the earlier funding failure; the lower amount also leaves less after it eventually arrives. Do not treat that later receipt as money already available at the deadline, or describe this schedule as executable without another funded arrangement.

The two changes have supplied no probability of default or universal limit for borrowing. They expose a specific question for the decision: what money or actually agreed timing change would meet January 20 without taking a reserve assigned to somebody else? If that remains unanswered, a favorable earnings case does not complete the financing plan. Keep the owner's separate capacity beside this company record.

A scenario does not supply its own probability

Nothing here establishes that a lower price causes fewer customers, or that this combination is likely. They are supplied conditions. A useful scenario needs a reason to examine its combination: perhaps a competitor's offer and weaker demand could occur in the same period. That possibility would need evidence in a real business.

Nor should every unfavorable input automatically be combined. Some conditions may be incompatible. Write down the relationship you are assuming before treating the combined case as a realistic downside.

At $95 per job, this model needs $3,000 ÷ $35, or about 85.71 jobs, to cover its operating expenses: 86 whole jobs. The proposed 80-job case falls short. That calculation has not found six more customers. At 80 jobs, a price of $97.50 would cover the same costs; it has not established that customers will accept it.

Keep the separate tests to see what each input changes. Then examine combinations that matter to the decision. Use expected value only when outcomes and defensible probabilities are supplied. Use the cash forecast for the dates money is needed. This month's operating result cannot answer either question on its own.