What is the synthesis in a case interview, and how do you give one?

In this article

1. Before you start
2. Why interviewers ask this
3. The synthesis from first principles
4. Worked example
5. Trade-offs and edge cases
6. How to answer it in an interview
7. Common mistakes
8. Follow-up questions
9. Sources

Before you start

This page is one part of the case interview guide, and it covers only the closing recommendation of a case. The McKinsey Personal Experience Interview (PEI) is a separate part of that firm's process, covered in McKinsey PEI questions.

Why interviewers ask this

A case that ends without a decision leaves the interviewer unsure whether you could stand in front of a client and say what to do. Bain's case interview page says the firm is looking for how you think, structure problems and build on ideas as the conversation evolves, not for a single right answer. The closing recommendation is where that structure becomes audible. That link is an inference, because the Bain page does not use the word synthesis.

None of the firm pages used for this article defines synthesis or says how it is scored. Coaches commonly describe it as its own signal, so treat the shape below as coaching convention, not a published firm rule.

A weak close recaps the work: it lists what you looked at and stops. A strong close opens with a decision, gives the two or three facts that force it, names one risk and proposes one test. The listener can act on it even if the case ends right there.

The synthesis from first principles

Synthesis

Synthesis is a decision built from analysis. In a case interview the synthesis is the whole closing answer: it opens with the recommendation and then gives the two or three reasons, one risk and one next step, spoken as a clear position rather than a list of the work you did.

Interim synthesis

Interim synthesis is a one- or two-sentence statement made after a single finding or exhibit during a case that says what you learned, what it means and what you will check next. It keeps a long case pointed at the final answer and makes the closing synthesis quick to assemble.

Recommendation

A recommendation is a spoken statement of the action the client should take, given as a clear position. It is the opening sentence of the synthesis and the part that says what to do, not only what was found, and the reasons that follow it support it.

In plain words

Think of a doctor with test results. Reading the numbers aloud is a summary. Saying what is wrong is the insight. Saying what the patient should do is the recommendation. The analogy breaks in one place: a doctor usually has the full results, while in a case you decide with partial data, so you must also say what you are unsure of.

A summary is a recap of what happened. An insight says what the findings mean. A recommendation says what to do. The synthesis is the whole closing answer: it opens with the recommendation and carries the insight as its reasons.

The structure comes from Barbara Minto. The Pyramid Principle is a method of presenting ideas as a pyramid under a single point, so the reader meets the conclusion first. Minto's site describes the ideas as organised "under a single point". Her Situation, Complication, Question (SCQ) framework is a tool for identifying the question in the reader's mind. At the end of a case, the interviewer's question is "what should the client do?"

Text description of the diagram

• A summary is a recap of what happened.
• An insight says what the findings mean.
• A recommendation says what to do.

Step 1: Start with the answer, not the analysis

1 · Answer2 · Facts3 · Risk4 · Next stepReplay

Step-by-step description

1. Open with the decision.
2. Two or three facts support it.
3. One risk tied to the answer.
4. One test that checks the risk.

• The answer comes first.
• Two or three facts support the answer.
• One risk is tied to the recommendation.
• One next step tests the risk.

The first sentence of the synthesis is the decision. Input: everything you have worked out. Output: one sentence that names what the client should do. The CaseLab handbook gives the order "Use: objective, answer, evidence, risk, next step." A simple wording of our own is "Given the goal of restoring profit, I recommend X."

The answer comes first because the Pyramid Principle puts one point at the top and organises everything beneath it. A listener who hears the point first can judge every later sentence against it.

Compare two openings. "We looked at revenue, costs and competition" is a recap. "I recommend winning back Region B's lost traffic" is a decision. Both are our own examples built from the same analysis, and only the second can be acted on. You now know the opening move. The next step is choosing what supports it.

Step 2: Give two or three facts that make the answer feel inevitable

Two or three facts support the answer. Input: your findings. Output: the few facts that make the decision feel obvious. Pick from your analysis; do not repeat all of it. More reasons are not better, because the aim is a clear answer, not a complete one.

Each fact should answer "why?" for the answer above it. If a fact does not help the listener believe the answer, leave it out.

Use numbers where you have them, and say them plainly. A single percentage change from the case beats three vague adjectives. The case math drills cover the arithmetic behind such figures.

Step 3: Name one risk tied to the recommendation

One risk is tied to the recommendation. Input: your answer and its weakest assumption. Output: one specific sentence about what could make the answer wrong or costly. A risk is not a general worry such as "the market may change". A fitting example is deep discounting that hurts margin: it belongs to the particular action recommended.

Naming the risk is not hedging. The answer stays firm and the risk is stated separately. "I think they could maybe focus on Region B" is the weak version, because it gives no position. "I recommend focusing on Region B; the main risk is X" gives a position and shows you can see how it could fail.

You now have the answer, its support and its weak point. The last step turns the weak point into an action.

Step 4: Propose one next step that tests the risk

One next step tests the risk. Input: the risk you just named. Output: one concrete action that would confirm or kill the recommendation. A good next step is small and fast, such as a pilot in a few locations, a data pull or a customer call.

This is how you stay decisive without full data. Refusing to recommend because data is missing gives the client nothing. A pattern of our own: "Based on what we know, I would recommend X. The main uncertainty is Y, so I would test Z."

On length, Hacking the Case Interview says "Your recommendation should take 60 to 90 seconds. Two minutes is the absolute maximum." Road to Offer says "A strong synthesis is 60-90 seconds". Those are coaching conventions, not rules from any firm.

Worked example

The CaseLab handbook uses a coffee chain whose profit fell 15% last quarter. Region B revenue fell from 32 million dollars to 24 million, a decline of 8 million dollars, which is 25 percent of 32 million.¹ Traffic fell after a competitor opened. Costs rose only slightly, the other regions are stable or growing, and customer ratings are solid.

The handbook does not give the chain's total profit, so assume an illustrative prior-quarter profit of 40 million dollars. A 15 percent fall is then 6 million dollars. If an illustrative half of each lost revenue dollar would have been profit, because costs barely moved, Region B's loss costs 4 million dollars of profit, about two thirds of the drop.¹

The model answer below has 102 words. At an illustrative speaking rate of 130 words a minute it takes about 47 seconds, and a 60 to 90 second convention is a budget of roughly 130 to 195 words at the same rate.¹

The three moves on this case, in wording of our own, are these. The summary is "We looked at revenue, costs and competition." The insight is "The problem is concentrated in Region B and is mostly lost revenue." The recommendation is "I recommend focusing the turnaround on Region B traffic recovery." The handbook's own strong answer reads "I recommend focusing first on Region B and winning back lost customer traffic there". Only the recommendation tells the client what to do, and the model answer below is a full synthesis that opens with it.

Text description of the diagram

• Region B revenue fell from 32 to 24 million, a drop of 25 percent.
• The other regions are stable or growing.

For a practice case with the same shape, see the profitability case walkthrough.

Trade-offs and edge cases

Interim synthesis: do it all through the case, not only at the end

The synthesis is not only a closing move. Coaches commonly advise an interim synthesis after each finding: what you learned, what it means and what to check next. A candidate who states meaning after each exhibit has a final recommendation almost written before the interviewer asks.

When the data is incomplete

Recommend anyway. State the answer, state the uncertainty separately and propose the test that would resolve it. Bain's page says the firm is not looking for a single right answer, so a clear, reasoned position with a named uncertainty fits what the firm describes. The same applies when a market-sizing estimate ends the case: finish with what the number means for the decision, not with the number alone.

When the interviewer cuts you off

Because the answer sits at the top, you can stop after any part and the listener still has your position. That is a practical advantage of the pyramid order, not a published rule. If you are cut off after the first sentence, the decision is already on the table.

Fewer reasons, not more

A longer list does not make the answer feel more complete. If you have five findings, choose the two or three that would convince a sceptic, and mention that you checked the others only if asked.

How to answer it in an interview

The spoken structure: say the decision, give the facts that force it, name the one risk, then end on the test. If you need it, ask the interviewer for a moment to organise your thoughts before you close; coaches commonly advise this. Practise it out loud and time it; the 20 original practice cases give you cases to close.

Common mistakes

• "We looked at revenue, costs and competition." (wrong) This is a recap and contains no decision. (right) "I recommend focusing the turnaround on Region B traffic recovery."
• Saving all synthesis for the final minute. (wrong) By then the case has drifted and nothing has been tied together. (right) After each exhibit, say what it means and what you will check next.
• "I think they could maybe look at Region B." (wrong) It takes no position. (right) "Based on what we know, I recommend Region B traffic recovery. The main uncertainty is whether the lost customers return, so I would test a local offer."
• Listing every finding to look thorough. (wrong) It buries the answer. (right) Pick the two or three facts that make the answer feel inevitable.
• "I cannot make a recommendation because we need more data." (wrong) It gives the client nothing to act on. (right) Recommend on what you know, name the uncertainty and propose a test.

Follow-up questions

What is the difference between a summary and a synthesis in a case interview?

A summary recaps what you did; a synthesis is the closing answer that says what it means and what to do. Hacking the Case Interview gives "revenue is down and costs are up" as a summary. The insight explains what the facts mean, and the recommendation, which opens the synthesis, says what the client should do.

How long should a case interview synthesis be?

Coaching convention is 60 to 90 seconds. None of the firm pages used here publishes a limit, so aim for one clear minute and let the interviewer ask for more.

Does synthesis only happen at the end of the case?

No. Interim synthesis after each finding keeps the case on track and makes the final recommendation quick to assemble.

What if you do not have enough data to make a recommendation?

Recommend anyway, name the uncertainty and propose the test that would settle it. A clear position with a stated risk is stronger than a refusal.

How do you practise a synthesis?

Close every practice case with a timed spoken answer in the four-part shape, then compare it with a model solution. The practice cases with model solutions work for this.

Does the synthesis mean anything else?

Outside interviews the word also names the philosophical sequence of thesis, antithesis and synthesis, and the making of chemical compounds. This page covers only the closing recommendation of a case interview.

Sources

1. The Minto Pyramid Principle (official site) — Minto Books International (Barbara Minto), 26 Oct 2026
2. Case interview preparation — Bain & Company, 26 Oct 2026
3. Chapter 5: Synthesis and recommendation, CaseLab Handbook — CaseLab, 26 Oct 2026
4. Case Interview Recommendation: How to Finish Strong — Hacking the Case Interview, 26 Oct 2026
5. Case Interview Synthesis: 60-Second Recommendation — Road to Offer, 26 Oct 2026