A personal brand isn't more people knowing you
It's a few people thinking of you first, for one specific thing

The mechanism is how memory is retrieved, not how much attention people pay you. Brains don't index by name, they index by cue: the more the cue present now overlaps with the cue that was there at encoding, the easier the recall. And the more things a concept carries, the more thinly activation spreads across those links — every extra label you claim costs a little strength on all the others. So "narrow" isn't modesty. It's concentrating limited activation on a single line.

• Hard evidence · retrieval runs on cue overlap: Tulving & Thomson (1973, Psychological Review), the encoding specificity principle — whether something comes back depends on whether the retrieval cue matches the cue present at encoding; a cue that wasn't stored alongside the memory fails to reach it even when it's semantically related. A classic paradigm, well replicated, hard. Neurally, cue-driven retrieval of this kind depends on hippocampal and medial temporal lobe pattern completion (a partial cue restoring the whole trace), which is consensus in memory neuroscience. Honest note: extending it to personal branding is an inference — there's no imaging work on branding itself.
• Hard evidence · more attached, slower to retrieve: Anderson (1974, Cognitive Psychology), the fan effect — the more facts associated with a concept, the longer the response time and the higher the error rate for verifying any single one. Reliably replicated, and the foundation of the ACT-R memory model. Hard. This is close to a direct costing of "I do a bit of everything".
• Related evidence · what comes to mind gets chosen: Tversky & Kahneman (1973), the availability heuristic — people judge probability and importance by how easily examples come to mind. When someone is looking for a person, whoever surfaces first often just takes the opportunity, because most people never run a full search. Strength: the phenomenon is solid, but the step from "easily recalled" to "gets the work" is an inference.

Boundaries: narrow positioning is leverage for someone who isn't being thought of yet. Once you're firmly recalled in a slot, narrow turns from leverage into a ceiling, and the move is to extend into the adjacent slot rather than keep hammering the same nail. Also, what's narrow is the label you show outward, not your actual capability. Do whatever you like privately — but every extra thing you announce blunts the hook.

Two layers. First, evaluability: some attributes simply can't be judged good or bad in isolation, so the moment you're placed side by side, the hard-to-evaluate strengths (taste, judgement, experience) go inert and only the most comparable column drives the decision. Second, category membership: people classify and remember by "what's most typical of this kind", so what you want isn't a higher rank inside an old category — it's being the only obvious occupant of a new, narrow one. Together those two explain how the same person can be worth twice as much.

• Hard evidence · judged alone and side by side, the answer flips: Hsee (1996, OBHDP), the dictionary study — one dictionary with fewer entries but in perfect condition, another with twice the entries but a torn cover. Evaluated separately, people pay more for the intact one; evaluated side by side, the order reverses. Because "10,000 versus 20,000 entries" means nothing to someone with no reference point. A repeatedly replicated effect, hard. It says directly: the comparison you're placed in determines price more than your actual merit does.
• Hard evidence · people classify by prototype: Rosch (1975, JEP: General), prototype effects — within a category, the more typical a member is, the faster and more accurately it's judged and recalled. Solid cognitive psychology, hard. That being the prototype of a narrow category beats being seventh in a broad one is my inference, not the original finding.
• Practitioner lore · the skill stack: "you don't need to be top at any one thing, only rare at the intersection of two or three" (Scott Adams' talent stack). Pure anecdote, no controlled work, heavy survivorship bias. Fine as a way of thinking, not as a law.

Boundaries: differentiation is not inventing a category nobody wants. There's exactly one test — is anyone already spending money on the thing you just named? If not, that isn't positioning, it's hiding. The other common failure is going so narrow that the listener has no existing slot to file you under, so they neither remember nor repeat you — narrow, but adjacent to a concept they already hold. As for the price itself, the unbundling-and-comparison work from the "that's too expensive" issue applies directly here.

Three layers stacked. One, mere exposure: seeing the same thing repeatedly raises liking on its own, no reason required. Two, processing fluency: familiar material is cheaper to process, and the brain misreads that ease as familiarity and as "this is probably true". Three, repetition suppression: repeated stimuli genuinely produce weaker cortical responses and lower processing cost — the physical floor under the first two. Net effect: appear in the same spot often enough and people conclude you probably know your subject without reading your whole body of work.

• Hard evidence · seeing more means liking more: Zajonc (1968, JPSP), the mere exposure effect; Bornstein (1989, Psychological Bulletin), a meta-analysis of 200-plus studies, confirms the effect is robust and that it is stronger when participants don't consciously register having seen the stimulus — meaning it doesn't depend on remembering what you said. Meta-analytic, hard. The same meta-analysis also shows the effect declines past a certain number of exposures: it's an inverted U, not "more is always better".
• Hard evidence · repetition makes things feel true: Hasher, Goldstein & Toppino (1977) first reported the illusory truth effect; Dechêne et al. (2010, PSPR) confirmed by meta-analysis that a repeated statement is rated as more likely to be true. Direction stable, hard.
• Hard evidence (brain mechanism) · repetition makes processing cheap: Grill-Spector, Henson & Martin (2006, Trends in Cognitive Sciences) review repetition suppression — when a stimulus repeats, neural responses in the relevant cortical regions (e.g. the occipitotemporal visual pathway) drop markedly while behaviour gets faster and more accurate. Consistent across fMRI and single-unit recording, hard. Subjectively, that "this got easier" is precisely the feeling read as familiarity and credibility.

The ethical line has to be stated plainly: repetition can only make a sentence easier to accept — it cannot make a wrong claim right. Apply it to something you haven't verified and you're trading someone else's cognitive shortcut for your own reach, which is exactly the persuasion-versus-manipulation boundary the six principles of influence issue draws. And illusory truth works on you too: a positioning line you've repeated for three years will feel true to you long after it expired, so run an external check once a year — ask real clients, not your own readers. Failure mode: what compounds is the same position being repeated; repeating the same sentence just becomes tiresome.

Two mechanisms. One, fresh information comes from distance: close friends move in your circles and know roughly the openings you already know; genuinely new information comes from people you're not in regular contact with, precisely because they sit in other circles. Two, passing your name on has a social cost: the person speaking for you cares not only whether they help you but whether saying it makes them look unreliable. A vague compliment invites a follow-up they can't answer, so they'd rather stay quiet. Writing the line concretely is really removing their risk — and that's why they open their mouth.

• Hard evidence · weak ties bring opportunity, causally: Granovetter (1973, AJS) first found that among people who found jobs through contacts, most of those contacts were seen "occasionally" or "rarely" (roughly 80%), not close friends — but that's observational, so correlation only. What nails causality is Rajkumar et al. (2022, Science): multi-year randomized experiments on about 20 million LinkedIn users showing moderately weak ties contribute the most causally to job mobility, more than either the strongest or the weakest. Large-scale randomized experiment, hard.
• Hard evidence · what actually gets passed on: Berger & Milkman (2012, JMR) analysed nearly 7,000 New York Times articles alongside controlled experiments: content producing high-arousal emotion (awe, anger, anxiety) is shared more, low-arousal sadness less; practically useful content is also shared significantly more. Large sample plus experiment, hard. The implication: what gets you repeated is "useful" and "surprising", not "impressive".
• Inference · the line should flatter the messenger: no direct controlled work on this one; it's a rule of thumb derived from the two above plus ordinary observation. Weak evidence, practitioner lore — but the cost is near zero and failure is harmless.

Boundaries: the value of weak ties comes from "the relationship still exists", not from "the relationship is thin" — someone you haven't spoken to in five years and have never helped won't speak for you, so the precondition is long-run giving with no immediate agenda (the networking issue covers exactly this). Don't turn the handed-over line into a mass-mailed template either. And the most-skipped step: when someone does pass your name on, go back and tell them what happened — close the loop once, and they'll speak up again.