Every brand already turns people away. The question is whether you chose the reason. The usual account of brand positioning as a filter stops one step early — pick a position, accept that it turns some people away, move on — and the step it skips is the one that decides everything. What variable is your filter actually sorting on? Few brands ever chose. An accidental filter sorts on whatever was easiest to signal rather than whatever you meant to select for.
Treating positioning as a filter is conventional advice, and good advice. A clear position “acts as a filter for business decisions,” as designer Paul Syng puts it, so that anything not reinforcing the position is easier to refuse. It works the same way on the demand side: positioning decides who you serve and who you don’t. The gap worth closing is between the filter you think you built and the one that is actually running.
What “positioning as a filter” actually means
Positioning filters in two directions at once. Inward, it governs decisions: a defined position makes “does this fit what we are?” an answerable question, so campaigns, hires, and partnerships that pull against the position get caught before they ship. Outward, it selects an audience. April Dunford describes positioning as context-setting that makes a product’s value obvious to “a well-defined set of customers.” Define the set, and you have defined who the work is not for.
That exclusion is not a flaw in positioning. It is the mechanism. Al Ries and Jack Trout built their entire framework on it: “the essence of positioning is sacrifice,” they wrote in Positioning: The Battle for Your Mind; a product “can’t be all things to all people and have a powerful position.” Seth Godin makes the same point from the audience side with his “smallest viable audience”: the smallest group that can sustain the work, which forces a decision about who the work is not for. Exclude no one and you don’t have a position. You have a description.
So the exclusion is settled. You are going to turn people away. What remains open, and what almost no one decides on purpose, is the rule deciding who.
The honest case for letting your register do the filtering
There is a real argument for letting the register do the sorting, and it deserves a fair hearing before the reframe. Write in an enterprise voice, frame the work for organizations with budget and appetite, and the small prospect reads the room and self-selects out before a sales call that would have ended in a number she could not say yes to. That looks like honest signage. Building a beautiful entrance to a room someone cannot afford is the crueler option.
The argument has support. In signaling terms, a price or register is an honest signal when it costs more to send than a low-fit buyer is willing to bear, which is why a premium tier communicates itself in advance. There is direct evidence that explicit exclusion can work: Wallach, Tanenbaum, and Blair, writing in the Journal of Consumer Research in 2026, found that “dissuasive framing” — telling some consumers a product is not for them — raised engagement among the customers it was for, because the signal made the product read as more specialized.
Taken on its own terms, the case holds. A filter that sorts cleanly by fit is doing exactly what positioning is supposed to do, and doing the prospect a favor in the process.
Where a register filter goes wrong: it sorts on posture, not fit
The case holds only if the register is sorting on fit. Often it is sorting on something else: posture. A register-based filter selects for whether the reader recognizes herself in the framing, and recognition is a proxy for fit only when the reader already knows her own situation well enough for the two to line up. The researchers behind that Journal of Consumer Research study flagged the same seam themselves: their work focused on “consumers who already knew their preferences,” and they called for future tests where choices “reflect self-expression rather than product fit.” Self-expression is posture. When the filter sorts on posture, it stops tracking fit.
A filter sorting on the wrong variable fails in both directions at once. A capable, well-funded prospect reads a frame pitched at a register she does not see herself in, decides the work is “not for people like me,” and leaves, a good-fit client lost to a stylistic signal. Meanwhile a poor-fit prospect recognizes the posture, feels the pull of a frame built for someone like who he wants to be, and converts. Engagement from the wrong audience reads like traction, but it only signals that the framing is landing with people who aren’t in a position to buy. Two failures, one cause, and both invisible unless you go looking.
The difference between the variable your filter should sort on and the one it often does:
| What the filter should sort on | What a register-based filter sorts on |
|---|---|
| Fit: can we do this work well for them? | Posture: do they recognize themselves in the framing? |
| Readiness: are they at the stage where this lands? | Familiarity: does the room look like rooms they already belong to? |
| Price: can the economics work for both sides? | Aspiration: do they want to be the kind of buyer this is for? |
Posture correlates with fit often enough to feel like it is working. The cost shows up in who you lose and who you keep, not in the signal itself.
The hidden cost: mistaking your client mix for market truth
This is where the wrong variable does its real damage. A brand reads its client mix as market feedback: these are the people who showed up, so this must be the market. But a self-selected mix doesn’t read the market. It reflects the filter that produced it. Sort by posture, and you assemble a client base of people who recognized the posture, then mistake their presence for proof of demand.
This is selection bias with a brand attached. When the sample is skewed, the conclusions drawn from it misread the broader market. The textbook case is self-selection: when people choose whether to respond to a survey, the ones who opt in distort the result. A client roster assembled by an accidental filter is exactly that kind of skewed sample. The well-fitted prospects who bounced off the register never enter the data. The brand optimizes harder for the audience it already has, sharpens the posture that selected them, and the loop tightens. Every turn reads as the market talking, when it is the frame talking back to itself.
How to tell what your filter is really sorting on
You can audit the variable. The check runs in two directions, matching the two ways the filter fails.
On the front end, watch the gap between attention and conversion. As one positioning audit frames it, “high traffic paired with weak conversion rates often signals a positioning disconnect,” and when a sales team “spends most discovery calls correcting assumptions,” the position is pulling in people on a misread. Those are posture signals: recognition without fit.
On the back end, code your losses by reason. For prospects who did not buy, sort the why into fit, readiness, price, or “didn’t understand what this was.” A roster of capable prospects lost to the last category is a filter sorting on posture, not fit. Run the same coding on who converted. If the people sailing through are recognizing a posture rather than matching on fit and budget, the filter is selecting for the wrong thing, however principled it looks from the inside.
Coherence is whether who you admit matches who you serve
This is a coherence problem before it is a marketing one. Coherence is the degree to which a brand’s signals agree with each other, and the register is a signal. When the register admits one kind of client and the strategy is built to serve another, two signals are contradicting each other, and the contradiction is making a strategic decision nobody approved. Brand strategy is supposed to be the filter running through every significant decision. A register left on default quietly overrides it.
The reframe is not “stop excluding.” You can’t; exclusion is what positioning is. The reframe is to decide the variable on purpose and then check that your signals enforce it. That check belongs to the brand, not to the register that happened to write the copy. A brand doesn’t fully own its meaning — the audience constructs that — but it does own which signals it sends, and a signal sorting on the wrong variable is a coherence break you can fix. The register is a surface; the structure beneath it is where the decision should live. A filter that looks principled while sorting on the wrong variable is the difference between a simple position and a shallow one: both read as exclusion, but only one is doing the job on purpose.
Deliberate beats accidental in every direction. Decide what you are sorting on, or accept that you are not really sorting at all.
What matters most
Every brand positions, and every position filters. The work is to know what yours is sorting on. Choose the variable deliberately (fit, readiness, price), then hold your signals to it, because a register left on default will pick a variable for you and call the result the market. The pitfall to avoid: reading the clients your frame attracted as evidence of who wants the work, when they may only be evidence of who recognized the frame.
