The hardest discipline in product communication is not saying things well. It is deciding what not to say. A PM lives in a state of permanent informational abundance — a dozen workstreams, a hundred open questions, a thousand details, all of which feel important because all of them are yours. The instinct, entirely natural and entirely fatal, is to convey it all. But communication is not a transfer of information from your head into someone else’s. It is a contest for a scarce and shrinking resource — the audience’s attention and memory — and in that contest, more is almost always less. This piece is about the single most important act of message discipline: ruthless prioritization, and its most useful heuristic, the rule of three.
The rule of three, and its relationship to the Rule of Seven
I have written at length that a single manager can hold high-fidelity alignment for roughly seven direct reports — the Rule of Seven, which I took from George Miller’s 1956 work on the magical number seven, plus or minus two. So it might look like a contradiction to now insist that your message be held to three. It is not a contradiction. It is the same physics measured at two different points in the network, and the gap between the two numbers is the entire lesson.
Seven is the capacity of a single node under near-ideal conditions: one manager, focused, receiving directly, one hop from the source. Three is what survives transmission across a lossy human network under real conditions. You budget to three not because three is magic but because three is what is still legible after the packet loss. Seven is what one node can receive. Three is what the network can carry.
For a PM the operational meaning is precise and demanding: manage your ongoing narrative down to the three highest-level, most impactful messages, and subordinate everything else to them. Stop a random stakeholder in the hallway, ask what your team is focused on, and the answer you would want them to give — those are your three.
Why less is more: the argument from how memory actually works
The case for severe prioritization rests on the mechanics of memory. Working memory is the mental bench where an audience holds your message while deciding what to do with it, and it is small. Present more than a few messages and your audience is more likely to retain nothing than to retain everything.
This is deeply counterintuitive, which is exactly why smart people get it wrong. Messages do not stack. They compete. Each additional point dilutes the salience of the others and raises the processing cost of the whole, until the audience, unable to tell what matters most, defaults to retaining whatever was easiest to remember — which is almost never the thing you most needed them to keep. When everything is emphasized, nothing is. Prioritization is not the sacrifice of part of your message. It is the precondition for any of it surviving at all.
The two moves: collapse and eliminate
Collapse: raise the altitude
The first move is to collapse — reduce granularity by moving up a level of abstraction, folding many specific items under one higher-level message. Five related feature efforts become one message about a strategic capability. A dozen reliability fixes become one message about earning enterprise trust. Done well, it lets you honestly represent an enormous amount of work with a small number of messages, because you have found the right altitude — the level at which many things resolve into one thing.
Eliminate: stop talking about it
The second move is harder. Some things cannot be collapsed into your three. They are simply not among the three most important things, and the only honest move is to eliminate them from your narrative. This does not necessarily mean stopping the work. It means you stop talking about it at the level of your headline messages, so it stops competing with what matters most. A message that competes with your priorities is not neutral. It actively costs you, by consuming the exact attention and memory your priorities need to survive. Strategy is choice, and a narrative that refuses to choose has no strategy in it, only inventory.
How to actually find your three
Write down everything you are tempted to say, without editing — the full sprawl, twenty or thirty items. Then, against each, ask one question borrowed from strategy rather than communication: if this item disappeared from the narrative, what would the audience get wrong? Most items, honestly answered, change nothing — the audience would form the same correct picture without them, which means they are detail, not message. A handful, if removed, would leave the audience with a materially wrong model. Those are your candidates.
There is a predictable force working against you the whole time: priority inflation. Every stakeholder wants their thing in your top three, because being in the narrative is how work becomes visible and funded. The discipline is to treat every addition as a subtraction — because it is one. When you frame it that way — “if this goes in, which of these two comes out” — the conversation changes, because now they are trading against a peer’s priority rather than against your stubbornness.
The paradox: over-communicate the few
Here is where this becomes genuinely counterintuitive. Having narrowed to three, the correct next move is not to state each once and move on. It is to over-communicate them — to repeat them far past the point where you personally are sick of the sound of them. People will forget. A message said once in a meeting is gone from most heads within days. People will ask you the very questions you answered explicitly last week, and the correct internal response is not exasperation but the recognition that this is simply how organizational memory works — it is lossy, and you are the error correction.
Say them the same way each time. Varying the wording to keep yourself entertained fragments the message across versions; the identical phrasing, repeated until it is boring to you, is what finally lets it lodge in someone else.
Why this is hard, and why the difficulty is the point
Prioritizing your message means deliberately leaving the visibility of some of your work on the table for the sake of the impact of the rest. Every time, in the moment, that trade feels like a loss. But the trade is exactly where the leverage is. Anyone can list everything they are doing; that requires no judgment and produces no clarity, and in the agentic era a machine will happily generate the exhaustive list for you in seconds. What is scarce, and therefore what you are actually paid for, is the judgment to decide what matters most and the discipline to say only that, repeatedly, until an organization has genuinely absorbed it. In a world of infinite things to communicate, the constraint is the craft.