Advantage and positioning choice

A team runs a positioning workshop. The output is a sentence: "The AI-powered workspace for modern teams that helps everyone do their best work." It is put on the site, in the deck, and at the top of the roadmap document.

Two weeks later the product decisions have not changed at all. An enterprise lead asks for meeting summaries and the team says yes, because enterprise leads are modern teams. An individual user asks for a faster personal editor and the team says yes, because individuals do their best work too. The sentence has excluded nobody, so it cannot decline anything.

The failure is that the team treated positioning as a description of the product rather than as a choice about who the product is for. A description can be accurate and still carry no information, because information comes from what is ruled out. And the failure is comfortable: the sentence tested well internally, nobody had to tell a customer no, and no revenue was refused. The cost only appears later, as a product that is slightly worse than a focused competitor for every population it serves.

The cost arrives through a chain of individually sensible yeses:

flowchart LR
  S["'The AI-powered workspace<br/>for modern teams'"] -->|"excludes nobody"| X["Every request<br/>qualifies as in-audience"]
  X --> Y1["Enterprise lead asks for<br/>meeting summaries → yes"]
  X --> Y2["Individual asks for a<br/>faster personal editor → yes"]
  Y1 --> W["Slightly worse than a focused<br/>competitor for every population"]
  Y2 --> W

No step in that chain is where the mistake happened. The mistake was the first box, which was approved because it offended nobody.

Ask this: which real customer request does this positioning tell us to decline? If the honest answer is none, you have written marketing copy.

A positioning choice is strategic when it makes four commitments at once. Miss one and it stops being able to decide anything.

Commitment	The question	Failure when vague
Audience	Who specifically, defined by behaviour	Every request qualifies as in-audience
Alternative	What they would use instead, named	You compare against nothing and win by default
Valued difference	What is better, in terms that audience already cares about	You claim a difference the audience does not pay attention to
Sustaining capability	What in your company makes that difference hold	The difference is copied within a quarter and you have nothing left

Difference is not advantage

A difference becomes an advantage only when three things hold at once:

1. Valued. The audience already cares about this dimension. A difference on a dimension nobody weighs is a cost you carry alone.
2. Hard to copy. Not impossible — hard, or slow, or unattractive for a competitor to copy because copying it would damage something they rely on.
3. Sustained by a real capability. Something the company actually has today: accumulated data, a distribution channel, an operating model, deep domain knowledge, or a shipped technical asset.

Test the third one strictly. A capability that is in design or planned is a statement of intent. Building a position on it means the position is currently undefended, which can be a legitimate bet — but it must be labelled as one.

Run a candidate difference through all three gates in order. Only one exit is an advantage; the other three are worth knowing about before you build on them:

flowchart TD
  D["A difference you have"] --> V{"Does the audience<br/>already weigh it?"}
  V -->|"no"| C1["A cost you carry alone"]
  V -->|"yes"| H{"Hard, slow, or unattractive<br/>for a rival to copy?"}
  H -->|"no"| C2["A lead measured in weeks"]
  H -->|"yes"| L{"Sustained by a capability<br/>that is live today?"}
  L -->|"in design or planned"| C3["An undefended bet.<br/>Label it as one."]
  L -->|"live"| ADV["Advantage"]

Most positioning arguments never reach the third gate, because the first two are answered by assertion.

The giving-up test

Positioning that gives up nothing is not positioning. Make the sacrifice concrete with three questions:

• Which population will now get a product that is worse for them than a focused competitor's?
• Which request type will be declined even when it arrives with revenue?
• Which capability will you deliberately not build, and for how long?

If you cannot answer all three, you have added an emphasis, not made a choice.

Both of these are positions for Noted. One of them can be used by an account manager holding a request:

The strong version is disagreeable, which is the property that makes it usable. It also states a sacrifice a colleague can carry into a conversation you are not in.

Giving up an audience is not the same as refusing their money

This distinction prevents most of the panic in this lesson. Giving up an audience means: they no longer shape the roadmap, they are no longer the population you optimize the core experience for, and their requests are handled against the standard you set rather than the standard they set. Existing contracts can continue. Support continues. What ends is their claim on your direction.

Boundary

Positioning is a commitment with a duration, not a permanent identity.

Two situations legitimately soften it. First, a genuinely early product may not yet know which audience it serves. Holding two candidate audiences while you test is honest, as long as the resource split is stated and time-boxed — "both, until the end of the quarter, then one." What is dishonest is holding two indefinitely and calling it focus.

Second, an audience can be a bridge. A population you do not intend to serve long-term can still be the one that funds or teaches the product now. That is a real strategy, and it has a specific requirement: name the condition under which you leave. Without that condition, "bridge" becomes the story you tell yourself about drifting.

The model also stops working when the four commitments are made at different levels of abstraction. An audience defined tightly and an alternative defined loosely produce a position that seems sharp and still cannot decline anything.

Read the Noted case if you have not already.

Take Signal 3: P95 response time rose 34% for large workspaces.

The details matter here. "Large" means more than 500 documents, which is about 4% of workspaces. Those workspaces hold a disproportionate share of paid seats. No support tickets mention speed; the signal came from monitoring. Separately, the case tells you that small teams are where most paid conversion happens today, that individual knowledge workers are the largest population and mostly free, and that the team collaboration environment is still in design.

Before you name a sustaining capability, check which ones Noted actually has. The case separates them, and the separation decides whether your position is defended or is a bet:

Capability	Status in the case	A position built on it is
Personal document workspace with AI assistance	Live	Defended today
File storage and retrieval	Live	Defended today
Public publishing path	Live	Defended today
Team collaboration environment	In design	Undefended — label it a bet
Private draft-review workflow	Planned	An intention, not a position
AI-first onboarding	Planned	An intention, not a position

Any position you write that leans on a row from the bottom half must say so in the sentence, not in a footnote.

Your task. Make a positioning choice for Noted and commit to what you give up.

1. Choose one audience from the three populations in the case. Define it by behaviour, not by label.
2. Name the specific alternative that audience uses today for this job. Include the option of doing nothing.
3. State the valued difference, in terms that audience already weighs. Say how you know they weigh it, or mark it <unknown>.
4. Name the capability that sustains the difference. Mark it live, in design, or planned, using the status table in the case. Do not blur the three.
5. State how long the difference would take a competitor to copy, or mark it <unknown>. Either way, say what would make copying unattractive to them — that part you can reason about without data.
6. Run the giving-up test. Name the population that now gets a worse product, the request type you will decline even with revenue attached, and the capability you will not build.
7. Decide what the large-workspace slowdown means under your position. It is a different kind of fact if your audience is the 4% holding paid seats than if it is the free majority.
8. Write the position as one sentence that could be disagreed with.

Expect to be pushed on: whether your valued difference is one any competent competitor could copy in a quarter; whether your sustaining capability is actually live rather than in design; and whether your sacrifice is real — you will be asked what you would say to the account manager holding the enterprise relationships.

Committed to a position? Check it against what a strong answer holds

• Defines one audience by observable behaviour and names the alternative that audience uses today, including doing nothing.
• States a valued difference and says how you know the audience weighs it, or marks that <unknown> rather than asserting it.
• Labels the sustaining capability live, in design, or planned using the case's status table, and calls any position that leans on the bottom half a bet in the sentence itself.
• Passes the giving-up test with three concrete answers: who gets a worse product, which request is declined even with revenue attached, and which capability stays unbuilt.
• Reads the large-workspace slowdown through the chosen position rather than as a separate problem, and ends in one sentence a colleague could disagree with.
• The most common weak move is naming a difference any competent competitor could copy in a quarter and calling it an advantage. A difference is only an advantage when it is valued, hard to copy, and sustained by something live today.

Produce artifacts/ST-03-positioning-choice.md using the template below.

Write it for the people who will field requests without you: an account manager, a support lead, a designer choosing a default. They should be able to read it and decline something on their own, and know what to say when they do.

In your Product Decision Case, this is the artifact that converts the ST-02 chain into a commitment. The audience row of your hypothesis map and the audience here must be the same population. If they are not, one of the two documents is out of date.

# Positioning choice · <product>

- **Date:** <YYYY-MM-DD>
- **Holds until:** <date or condition — positions have durations>
- **Owner:** <who owns this choice>
- **Diagnosis this serves:** <one line from ST-01>

## The position

<One sentence a reasonable colleague could disagree with. It must contain the
audience and the difference.>

## The four commitments

| Commitment | Statement | Status |
|---|---|---|
| Audience | <defined by observable behaviour, not by label> | <evidenced / assumed> |
| Alternative | <what they use today, including doing it manually or not at all> | <evidenced / assumed> |
| Valued difference | <better on a dimension this audience already weighs> | <evidenced / assumed> |
| Sustaining capability | <what makes the difference hold> | <live / in design / planned> |

## Why the difference holds

- **Valued because:** <evidence that the audience weighs this dimension>
- **Hard to copy because:** <what makes copying slow, costly, or unattractive>
- **Time to copy, honestly:** <weeks / quarters / years, and who could do it>
- **If the capability is not live:** <state plainly that the position is
  currently a bet, and what would make it real>

## What is given up

| Sacrifice | Specifics |
|---|---|
| Population that now gets a worse product | <who, and worse in what way> |
| Request type declined even with revenue attached | <the type, and the sentence you will say> |
| Capability deliberately not built | <what, and for how long> |

## How others should use this

<Two or three examples of requests, and how a support lead, account manager, or
designer resolves each one using this document alone.>

## Evidence and limits

| Claim | Source | Population | Window | What it cannot establish |
|---|---|---|---|---|
| <claim> | <source> | <who> | <when> | <limit> |

## Uncertainty

- **Confidence:** <low / moderate / high> — <why>
- **Strongest case for the audience you gave up:** <argue it properly>
- **What it would cost to reverse this choice:** <effort, time, relationships>

## Revision trigger

<The observation and threshold that would reopen this choice, and who would see
it first. Include the bridge condition if this audience is a bridge: what
specifically would make you leave it.>

## Gaps

<Everything marked `<unknown>`, as open questions with who could close them.>