You can hide a half-formed idea behind a slide. You cannot hide it inside a paragraph. A bullet point lets you gesture at a conclusion and move on before anyone asks how you got there; a sentence has to commit, subject, verb, claim, and the moment you try to write the connective tissue between two bullets, you discover whether you actually had an argument or just a list. That is the whole case for executive writing: the document is where the thinking is forced to be real.

The quick version

  • Executive writing means producing the short, decision-grade documents a leader lives on, memos, proposals, board notes, update emails, where the job is to move a busy, skeptical reader to understand and act, fast.
  • The discipline starts with one habit: answer first. Lead with your conclusion and recommendation, then support it, don't make the reader assemble it.
  • Writing it out is the point, not the chore. A clear memo exposes the gaps in your logic that a slide deck would have let you skate past, so writing improves the decision, not just the communication of it.
  • The trap is fluency without substance: prose that reads smoothly and says nothing checkable. Plain words and a stated "so what" are the cure.

Write to think, not just to inform

The most quoted experiment in this field is also the most instructive, because a company bet its meeting culture on it. Amazon banned slide presentations internally in favour of narrative memos that everyone reads in silence at the start of the meeting. In his 2017 letter to shareholders, Jeff Bezos set out the reasoning: "We don't do PowerPoint (or any other slide-oriented) presentations at Amazon," he wrote, because the discipline of full sentences forces a completeness that bullets allow you to dodge. A list of phrases can imply a logical chain that was never actually built; a paragraph has to build it.

So the move is to draft your next significant decision as prose before you make a single slide. Not a deck you will narrate, a document that stands alone and survives being read by someone who can't ask you a follow-up question in the room. If you cannot write the connecting sentence between two of your points, you have just found the weak joint in your own reasoning, which is far cheaper to find at your desk than in front of your board.

"We don't do PowerPoint … presentations at Amazon.", Jeff Bezos, 2017 letter to shareholders

An honest limitation. The memo is not a universal tool. Bezos himself noted in the same letter that "the quality of these memos varies widely", the format guarantees nothing on its own, and a bad six pages is worse than a good three slides. Long-form writing is also slower to produce and slower to consume; for a status ping or a calendar nudge it is overkill. The narrative memo earns its cost on consequential, contested decisions, where the thinking genuinely needs to be tested, not on every message that crosses your desk.

Answer first: the structure that respects the reader

Once you have decided to write, the single highest-leverage structural choice is to put the answer at the top. This is the core of the Pyramid Principle, developed by Barbara Minto, McKinsey's first female MBA hire, and still taught to new consultants across McKinsey, Bain and BCG (McKinsey Alumni). Minto's rule is deceptively simple: lead with the governing thought, your recommendation, your "so what", then group the supporting arguments beneath it, then the evidence beneath those. As she put it, you think from the bottom up, but you present from the top down.

flowchart TD
  A(["The answer
your recommendation / 'so what'"]) --> B(["Supporting argument 1"])
  A --> C(["Supporting argument 2"])
  A --> D(["Supporting argument 3"])
  B --> B1(["Evidence, data, detail"])
  C --> C1(["Evidence, data, detail"])
  D --> D1(["Evidence, data, detail"])
Minto's pyramid, the conclusion governs the page; support hangs beneath it, not in front of it. Leaders Loop

The military reached the same conclusion independently and gave it a blunter name: BLUF, Bottom Line Up Front. It is written into U.S. Army correspondence doctrine (Army Regulation 25-50), which instructs writers to lead with the bottom line because, as the guidance frames it, the greatest weakness in ineffective writing is that it doesn't quickly transmit a focused message (background on BLUF). Two very different institutions, a consultancy and an army, converged on the same instinct because both serve readers whose time is scarce and whose decisions are costly.

So the move is to write your conclusion as the first line of the email or memo, before any context. "I recommend we delay the launch to Q3; here's why" beats three paragraphs of background that bury the ask. The reader who agrees can stop reading; the reader who doesn't knows exactly what they are arguing with. Answer-first structure is, at bottom, a courtesy, it is how you respect the fact that the most senior reader has the least time. It pairs naturally with the broader habit of structured communication, which is Minto's method applied across every channel a leader uses.

Beware fluency: smooth prose that says nothing

The opposite failure of the muddled memo is the polished one, writing that reads beautifully and commits to nothing. This is the harder problem now that anyone can generate fluent text on demand. Larry McEnerney, long-time director of the University of Chicago's Writing Program, makes the point sharply in his widely watched lecture: readers don't reward writing for being clear or well-organised; they reward it for being valuable to them, for changing what they think or do ("The Craft of Writing Effectively"). Clarity that delivers nothing the reader can use is just well-lit emptiness.

William Zinsser, in On Writing Well, named the everyday version of this disease: clutter. "Clutter is the disease of American writing," he wrote, the unnecessary words, the circular constructions, the jargon that lets a sentence sound official while saying less than it appears to. His remedy is to prune ruthlessly and to put the missing "I" back into institutional prose, so a human is visibly making a claim a reader can hold them to.

So the move is to run a two-pass edit on anything that matters. First pass: delete every sentence that, if challenged, you could not defend with a fact, a number, or a clear logical step. Second pass: cut the qualifiers, "somewhat," "in order to," "it should be noted", that pad the word count without adding meaning. What survives is the part that was actually doing work. The test is brutal and useful: read each paragraph and ask, "what does the reader now know, or do, that they didn't before?" If the answer is nothing, the paragraph is decoration.

A worked example

Take a product lead, call her Priya, who needs sign-off to pause a feature and reallocate two engineers. (Illustrative throughout; not a real memo.) Her first draft opens the way most do: a paragraph on market context, a paragraph on the team's history with the feature, a paragraph of caveats, and the actual request buried near the bottom of page two. It reads fine. It also fails, because the director skims the first screen, sees no decision, and moves it to "read later", where it dies.

Rewritten answer-first, the same memo opens: "Recommendation: pause the notifications feature for one quarter and move its two engineers to onboarding, where churn data shows the larger problem. Below: the evidence, the cost of pausing, and the one risk I'm watching." Now the director knows the ask in eight seconds. The context that follows is no longer throat-clearing, it is support for a claim already on the table, so every paragraph has a job.

flowchart LR
  A(["Context paragraph"]) --> B(["History paragraph"])
  B --> C(["Caveats"])
  C --> D(["The actual ask
(buried, page 2)"])
  D -.->|"reader gave up"| X(["'Read later' → dies"])
  E(["The ask, line 1
+ why, in one sentence"]) --> F(["Evidence that supports it"])
  F --> G(["Cost + the one risk"])
  G --> H(["Decision made"])
Same facts, two orders. The buried ask dies in the inbox; the answer-first version gets decided. Leaders Loop

Then Priya does the unglamorous part Bezos insists on: she sets the draft aside, comes back a day later, and cuts a third of it. The press of writing it out has already done its quiet work, in drafting the "cost of pausing" section she realises she has no number for it, which is itself the most important finding. Better communication was the smaller prize. The real win was that writing the memo exposed the hole in the decision while it was still cheap to fix.

Frequently asked questions

Isn't a good slide deck just as effective as a memo?

For some jobs, yes, slides are excellent for data a reader should see rather than read, and for live, exploratory discussion. The narrative memo wins on consequential decisions specifically because it removes the presenter. A deck is performed; gaps can be smoothed over by a confident speaker. A memo is read in silence, so the logic has to hold up on its own. Use slides to show; use prose to argue.

Won't "answer first" feel abrupt or presumptuous to senior readers?

It is the opposite. Burying the ask reads as either timidity or a failure to know your own point. Stating the recommendation up front signals that you have done the thinking and respect the reader's time, and it lets them engage with the actual question instead of reverse-engineering it from your evidence. You can soften the tone without burying the point: "I recommend X, and I want to flag the trade-off honestly" is direct and humble at once.

I don't have time to write a six-page memo for everything. How do I choose?

You shouldn't, and Amazon doesn't either. Match the format to the stakes. A one-line BLUF email handles most traffic. Reserve the long, written-and-rewritten memo for decisions that are contested, expensive, or hard to reverse, exactly the cases where the thinking most needs testing. The format is an investment; spend it where a wrong decision would cost the most.

Does writing skill even matter now that AI can draft documents?

It matters more, not less. Tools generate fluent prose easily, which floods inboxes with text that reads well and decides nothing, the fluency trap at scale. The scarce skill is no longer producing sentences; it is judging whether a document is valuable to its reader and carries a defensible argument. That judgment, knowing what to cut and what claim is load-bearing, is the part you still have to own.

How do I get better at this without a writing course?

Edit, don't just write. Take a memo you sent last month and cut it by a third without losing meaning; the gap between the two versions is your lesson. Ask a trusted colleague to read a draft and tell you, in one sentence, what you're recommending, if they can't, your structure is the problem, not their attention. This is reflective practice applied to prose, the same loop covered in self-awareness & reflective practice.

Related in the Toolkit

Where to go next