How to analyze a business case before making a decision

Start with a decision, not a company biography

Write the choice you need to make in one sentence. “Should we introduce a cheaper plan for occasional users?” is useful. “What can we learn from Netflix?” leaves the research without a boundary. Specify the decision owner, the time available, the resources at risk and the outcome you want to change. This gives you a reason to include a case and a reason to leave one out.

Harvard Business School describes case learning around a decision with incomplete information. Our worksheet applies that principle to practical research; it is not an HBS teaching case or a substitute for its courses. When reading a retrospective account, temporarily set the outcome aside and reconstruct what could have been known at the decision date.

Build three columns: evidence, interpretation, uncertainty

In the evidence column, record what the source actually establishes: a price change, a product withdrawal, a reported result or a dated announcement. In interpretation, record the explanation offered by the case or by you. In uncertainty, list what you cannot establish. A revenue increase after a redesign does not by itself establish that the redesign caused the increase.

Attach a source and date to each important claim. Keep the denominator with every number: a percentage of customers is different from a percentage of transactions. Preserve whether the figure is a company statement, an estimate or an independently documented outcome. If a source does not support the strength of your wording, weaken the wording or remove the claim.

Compare the mechanism and the constraints

Choose a small set with different outcomes. For a pricing decision, useful comparators might include a price cut, a simplified bundle and a decision to maintain price. Ask what changed customer behavior, what the company needed to execute it, and what made the result possible. A recognizable brand is not enough to make a case comparable.

A simple comparison table has six fields: customer, original problem, action, required capability, observed outcome and mismatch with your situation. The mismatch field is essential. A company with proprietary distribution or years of accumulated trust may have options your team does not. Write that difference before copying its action.

Finish with a test and a reversal condition

Imagine a small software team considering a lower-priced plan. Its case review might suggest that onboarding friction, rather than price, is the first obstacle. That is an illustrative hypothesis, not a measured result. The next step could be a limited onboarding test with a predefined activation measure, while keeping support load and cancellations as guardrails.

Your final memo should say what you will do, what evidence could change your mind and when you will review the decision. Case research is complete when it narrows the uncertainty that matters. A larger pile of examples is not automatically a better decision.

Use this worksheet for your decision

Copy this into your notes. Leave unknowns visible instead of filling gaps with guesses.

Decision and deadline:
Options, including doing nothing:
Evidence and source dates:
Possible mechanism:
Why these cases are comparable:
Important differences:
Smallest useful test:
Success measure and guardrails:
What would reverse the decision:

Find evidence in the archives

Choose a supporting case and a counterexample with a different outcome. Read their sources and check whether their conditions match yours.

• mistake.club
• change.archi

Cases to practice with (read and check the underlying sources):

• In 2020 Airbnb's travel business plunged 72%; cost cuts led to a $100B IPO
• NRG's consumer-green pivot cost CEO Crane his job and $1.5bn in goodwill writedowns