Make the claim testable
A hypothesis is a provisional explanation or a conditional view about a decision. "The business should grow" is too broad to help you choose an analysis. "A fall in repeat orders may explain most of the profit decline" gives you something to measure. "Launching in a second city could meet the growth target if acquisition and delivery economics remain attractive" makes the conditions of a recommendation explicit.
You do not need to announce an answer before you have evidence. If the prompt supplies little information, start with competing explanations and a comparison that would distinguish them. For a profit decline, asking how revenue and costs changed may be more useful than choosing a favourite explanation at random. Keep the initial explanation provisional until you have tested it.
Choose a test that could change your mind
For each hypothesis, identify the observation you would expect if it were true. If repeat orders are the problem, order frequency among comparable returning customers should have fallen enough to explain a meaningful share of the decline. A lower company-wide average alone may be misleading if the company acquired many new customers who have had little time to reorder.
Request a comparison with an appropriate denominator and period. You might ask for orders per returning customer in the same quarter this year and last year, alongside the number of such customers and contribution per order. Explain that this comparison lets you estimate the profit impact of lower frequency while keeping other drivers visible.
Evidence that changes the next step
A fictional repair service has stable job volume but falling operating profit. You initially suspect higher materials costs. The exhibit shows materials cost per job unchanged and average net revenue per job down 8%.
You could respond with "Materials cost does not explain the decline in this period. I would now split net revenue into the mix of jobs and the price realised within each job type." It is a stronger analytical move than finding a reason to keep discussing materials.
Separate an accounting explanation from a cause
A profit bridge can show that lower price explains the decline. It cannot, by itself, tell you why prices fell. Possibilities include a promotion, more lower-priced products in the sales mix, discounting by the sales team or a shift in customer contracts. Those are different mechanisms and can lead to different actions.
Write down the observed result and its financial driver. Then identify a possible cause and a test for it. If realised price declined, compare like-for-like products before concluding that individual products were discounted. If discounts increased, ask whether customer retention or conversion improved enough to justify them. Keep the distinction between the measured financial driver and its possible cause clear as you investigate.
Prioritise impact, uncertainty and effort
You rarely have time to explore every branch equally. Start with an issue that could materially affect the decision, remains uncertain and can be tested with the available information. A very large but already established cost may not be the best next question. A smaller variable that could reverse the recommendation may deserve attention first.
Imagine two options are separated by $40,000 in annual contribution. Their office rent differs by $2,000, while one option depends on retaining a customer contract worth $150,000 in contribution. The contract's contribution is large enough to reverse the choice between the options. Explain that prioritisation out loud; the interviewer should not have to guess why you want the next dataset.
Do not turn this into a rigid scoring formula. The point is to compare what the analyses might change. Record a condition for returning to a secondary issue. For example, "If demand is sufficient, I will check whether capacity requires additional fixed investment." You can then continue with the demand analysis and return to capacity if needed.
Keep a visible record of what changed
After each major exhibit or calculation, note how it changed your view and which questions remain open. "The price decline accounts for most of the shortfall; we still need to distinguish discounts from product mix." This protects you from repeating a question or carrying a disproven assumption into the recommendation.
In solo practice, pause before revealing the next checkpoint. Predict which evidence would support your view and which would weaken it. Then compare your prediction with the information you receive. If you only read the explanation after the answer is visible, you may feel you understand the reasoning without having practised choosing the test yourself.
At the end, review one moment when you changed direction. Check whether the new evidence tested your explanation and justified that change.
Practice question
You suspect falling customer numbers explain a revenue decline. Customer numbers are stable, but orders per customer fell. What is the best next move?
The business scenarios and figures are fictional practice examples written for CaseMentor.