Your niche: the onliness test

Marty Neumeier's book Zag (2006) is reported to argue that a brand should be able to say it is the only one that does a particular thing, and to offer a minimal template for saying it. As reported, the template runs: Our [offering] is the only [category] that [benefit]. The exact wording in the book was not verified for this page, so treat that as a reported summary of the idea, not his sentence.

For one person, the template needs three swaps. The offering is you. The category is the field you work in, said the way the people who hire you would say it. The benefit is what changes for a specific group of people, which is where most drafts go vague. A working version for an individual:

I help [audience] [change] by [method]. Proof: [artefact].

Every slot is a noun or a number somebody could check. "Businesses" is not an audience; "grocery planners" is. "Drive growth" is not a change; "cut stockouts" is. A method is something you actually do. A proof is something a stranger can open. If a slot can only be filled with an adjective, it is not filled yet.

Take your line, cross out your name, and write in the name of someone who does roughly what you do: a colleague, a peer at another company, the person who last got the role you wanted. Read it again. If it is still true for them, it does not yet distinguish you; a stranger could not use it to choose you over them.

This is a practitioner test, widely used in positioning work, and it is not attributed to Neumeier here. It is useful because it is fast and because it is uncomfortable in the right way: it shows how much of a line is borrowed vocabulary. It has one blind spot. A line can pass the swap and still be wrong for you, because it is false, or because it names a niche you do not want. Passing proves the line is yours; it does not prove the line is good.

Company positioning, naming and visual identity follow the same instinct at a larger scale, and they have their own home: learn-ai-branding-with-phoebe.

Kevin Kelly's page on The Technium is dated March 4, 2008, and it carries two versions of the argument. At the top sits a shorter rewrite (it appeared in Tim Ferriss's Tools of Titans; the page does not give the year). Below it is the 2008 original. They do not use the same number in the same way, and quoting one as the other is the most common mistake made with this essay.

• The rewrite. A true fan is someone who "will buy anything you produce". Each makes you $100 of profit a year, so 1,000 fans make $100,000 a year.
• The 2008 original. Each true fan spends $100 a year on your work, framed as one day's wages. That is spending, not profit.

Kelly then says plainly that "The number 1,000 is not absolute": at $50 a fan you need 2,000, at $200 you need 500. And the arithmetic only works under one condition, a direct relationship: "they must pay you directly". He also writes, about his own model, "My formula may be off by an order of magnitude". Read together, the essay is a sketch of how small a direct audience can be, offered with its uncertainty on the label.

On April 27, 2008, Kelly posted "The Case Against 1000 True Fans" on the same site. In it Jaron Lanier challenges the idea from experience: he says he has not found a single musician who meets the definition. Kelly published the challenge next to his own argument rather than answering it away.

That is the honest state of the evidence. The 1,000 is a model, its author flagged how far off it could be, and a serious critic disputed whether anyone fits it, on the author's own page. Nobody has settled it, and nothing here settles it either.

The transfer, as an interpretation. Most people building a professional reputation are salaried, not selling to fans, so the money arithmetic does not carry over. What carries over is the smallness. For a salaried professional the niche is the small set of people who hire you, refer you and cite your work: often dozens, not thousands. A line written for them can afford to be narrow, because they are the only readers who change anything. That reading is ours, not Kelly's. Segments, funnels and channels as a business discipline belong to learn-marketing-with-phoebe.

Press each draft button, then type your own line. The checker prints four counts and a percentage, each marked heuristic.

• Audience words: who the line is for, from a list of roles and sectors (planners, clinics, warehouses, founders).
• Outcome words: what changes for them (cut, fewer, catch, stockouts, delays).
• Proof: any number, or a word for something a stranger can open (teardown, repo, talk, case study).
• Vague words: the shared vocabulary (passionate, results-driven, insights, thought leader).

Ines's three constructed drafts score as follows. Draft 1: 11 words, audience 0, outcome 0, proof 0, vague 7, specificity 0 percent. Draft 2, "Data analyst who builds forecasting dashboards for retail teams.": 9 words, audience 1, outcome 0, proof 0, vague 0, specificity 100 percent. Draft 3, "I help grocery planners cut stockouts with forecasts they can check, one public teardown at a time.": 17 words, audience 2, outcome 4, proof 1, vague 0, specificity 100 percent. Two test lines are worth trying too: "Seasoned, strategic thought leader and lifelong learner." scores 7 words and vague 4, and "Helping 3 warehouse teams find late orders before 9am." scores 9 words, audience 1, outcome 2, proof 2.

A good line is not one with a high percentage. It is one with no zero in the first three counts and nothing in the fourth.

Draft 2 is the warning. It scores 100 percent specificity, the same as Draft 3, with exactly one audience word. That word is "analyst", and it names Ines, not the people she helps. The checker cannot tell the difference between the writer and the reader; it only knows the word is on a list. Read the counts, not the percentage: outcome 0 and proof 0 say Draft 2 is still missing two of the three things a stranger needs.

The lists are finite. A word that is not on a list scores nothing. "Retail" and "teams" are not on the audience list, so Draft 2 gets no credit for them, and a line in a field the lists do not know, a sharp line about veterinary rostering, say, can score low while being exactly right. All four lists are printed under the widget so you can see what it is looking for.

It cannot check truth. In the test line, "3 warehouse teams" counts as proof because it contains a number, and "teardown" in Draft 3 counts because it is a listed word. Whether the three teams or the teardown exist is for a stranger to check, which is the reason to link them. Ines's first teardown does not exist yet; session 3 builds it, which is why her line promises a habit rather than claiming a count. A count of zero is a useful prompt; a high count is not a verdict.