Answer first, then the reasons

Most of us explain things in the order we worked them out: the background, the data, the analysis, and finally the conclusion. Senior readers want the reverse. They have many decisions waiting and they read from the top until they have enough. If the answer is at the bottom, many of them never reach it.

Barbara Minto, who joined McKinsey in 1963 as the firm's first female MBA professional hire, turned this into a method she called the pyramid principle, published as The Pyramid Principle in 1985. The structure has three levels:

• The answer at the top, in one or two sentences: what you recommend and why it matters.
• Three or so reasons below it, each a full sentence that supports the answer. The reasons in a group should be the same kind of idea, should not overlap, and together should cover the case. Minto calls this MECE, mutually exclusive and collectively exhaustive.
• Evidence below each reason: numbers, trials, quotes, sources.

You still think bottom-up, from the data to the conclusion. You present top-down. Military writers have a short name for the same habit: BLUF, bottom line up front.

A quick test for any deck or memo: read only the headings or slide titles, in order. If they tell the whole story as full sentences, the structure works. If they read "Background", "Analysis", "Options", "Next steps", the reader has to do the work you should have done. Sentence titles are often called action titles.

Situation, complication, resolution

The pyramid tells you where the answer goes. The opening tells the reader why they should care. Minto's introduction has four parts: the situation (something the reader already knows and agrees with), the complication (what changed, or why the situation is no longer fine), the question this raises in the reader's mind, and the answer. Consultants often shorten it to SCR: situation, complication, resolution, with the question left unsaid.

Each part should be one or two sentences:

• Situation: "'Now' delivery is our headline promise. Offline receiving lifted stock accuracy from 82% to 88% and cut the share of orders with a missing item from 9% to 7%."
• Complication: "Progress has stalled. 4,200 orders a day still lose an item, more than half of those items are 300 fast movers, and we count them once a month, so a wrong record stays wrong for two weeks on average."
• Resolution: "Count those 300 items every afternoon and the full range quarterly instead of monthly, starting with a four-week pilot in 20 stores."

With senior audiences, put the resolution first and use the situation and complication as the reason. That gives you an opening of three sentences: the answer, then "here is why".

Three things go wrong with SCR. The situation turns into a history lesson; keep it to what the reader already accepts. The complication is not a complication: "we want to improve accuracy" is a wish, while "progress has stalled and rivals open in three of our cities this year" is a change that demands a response. And the resolution answers a different question from the one the complication raised. If the complication is about fast movers, the resolution must be about fast movers.

Executive updates that bring bad news early

Between big presentations, leaders follow your work through short written updates, weekly for a pilot or a launch, monthly for a steady product. A good update is read in one minute and has the same shape every time, so readers learn where to look:

1. Status in one line, with the reason: on track, at risk or off track.
2. Three to five numbers against their targets, with the trend. Include the guardrails.
3. What changed since the last update.
4. Decisions needed from the reader, each with a date. Write "none this week" when there are none, so the reader knows you checked.
5. Risks and what you are doing about each.
6. The next update date.

Many teams colour the status green, amber or red. The colours only help if they follow rules everyone knows. A workable set: green means on plan; amber means at risk, and the team has a fix that needs no help; red means off plan, and the team needs a decision or help from the reader. Without rules you get what people call watermelon reporting: green outside, red inside. Every update says green until the week the project is suddenly red, and the leader learns that your green means nothing.

The rule underneath is no surprises. Bad news should reach a leader early, from you, in private, with what you are doing about it. The first time the managing director hears that the pilot is slipping should never be in a room full of people. For deeper reviews of a product's numbers, use the metric review format in Metric reviews that change decisions.

Presenting a business case to a managing director

A managing director weighs a few things in every case, whatever the industry: money (margin, cash, payback), risk (what happens if it fails, and whether it can be undone), customers and competitors, people (the staff who must work differently), and fit with the strategy already agreed. Your case should answer each one before she asks. The calculations themselves (ROI, payback, NPV, sensitivity ranges) are covered in Business cases; here the subject is how you present them.

Before the meeting. Talk privately to each person whose view the managing director will ask for. Agree the numbers with the head of finance, so finance does not challenge them in the room. Hear the strongest objection from the person most likely to oppose you, and change the plan where they are right. Toyota calls this groundwork nemawashi; many teams call it pre-wiring. Send the memo (Writing for decisions) as a pre-read a day ahead.

In the room. For a 20-minute slot, plan about 5 minutes of presenting and 15 of discussion. The discussion is where the decision gets made, so protect it. Five slides are usually enough, each with an action title:

1. The answer and the decision you need today.
2. The problem in numbers: situation and complication.
3. The options, judged on the same criteria, and why you recommend one.
4. The money: cost, value, payback, a low, base and high case, and what you did not count.
5. Risks, gates, and what you are asking for today compared with later.

Everything else goes in an appendix you can jump to when asked. Be clear about the kind of money. Sales kept and margin kept are different numbers, and a leader who discovers the difference herself stops trusting the rest. Give ranges, and name the one assumption that would change your answer.

At the end, say the decision back in one sentence ("So: the pilot is approved, the rollout follows if the three gates pass, and the scanners come back to you separately") and wait for a yes. Within 24 hours, send a short note with the decision, the conditions, the actions with owners and dates, and any open questions from the meeting.

Handling tough questions

Most tough questions are predictable, so prepare for them like an exam. List 15 to 20 questions from the point of view of everyone in the room: the managing director, finance, operations, technology, the sceptic. Rank them by how likely they are and how much damage a weak answer would do. For the top eight, write a one-sentence answer, the one piece of evidence behind it, and where it is in the appendix. Then rehearse with a colleague who plays the toughest person in the room. Some teams call this rehearsal a murder board.

In the room, use the same rhythm for every question:

1. Listen to the end, without planning your answer while she is still talking.
2. Pause. Two seconds of silence gives you time to think, and it looks calm.
3. If it is unclear, restate it: "Are you asking whether the pilot stores are typical, or whether 20 is enough?"
4. Answer in one sentence, the claim first.
5. Add one piece of evidence: a number with its source, or the slide where it lives.
6. Stop, and check: "Does that answer it?"

Some questions need more care:

• You don't know. Say so, give what you do know, and commit to a time: "I don't have the split by city here. I'll send it by 3 pm today." Never invent a number to fill a silence. A wrong number said with confidence will be remembered longer than the whole presentation.
• A number is challenged. Give its source and range. If you were wrong, say so and say what changes in the decision. Often the answer is "nothing, the low case still pays back".
• "Why not just do X?" Treat X as one more option and compare it on your criteria. Sometimes X is better, and saying so builds trust.
• The question behind the question. "What happens during the festival?" usually means "Will this hurt sales in my biggest weeks?" Answer both.
• "Just give me a date." Give a range with how sure you are and what will narrow it: "Most likely the week after the festival freeze. I'll be sure after the pilot readout on 4 December."
• Someone who is not in the room. "What does the head of store operations think?" Represent her view fairly, including where she disagrees. Better still, invite her.