Niche down is the most repeated advice in agent marketing and one of the least followed. That gap is usually blamed on stubbornness, which is unfair. The resistance is rational: in a slow market, deliberately narrowing the set of people who might hire you feels like negligence.
The advice and the objection are both right, because they are answering two different questions that keep getting merged.
Two decisions that get confused
What you are known for is a positioning decision. It determines which strangers find you, what people say when they describe you, and which questions you get associated with.
What you accept is a business decision, made one call at a time.
These are independent. Nothing about a page describing your work with townhouse sellers prevents you from taking a detached listing from a referral. Nobody reads an agent's site as a list of restrictions, and the people most likely to call you already know you and are not reading it at all.
Almost all the fear about niching lives in the second decision, while all the benefit lives in the first. Separate them and the advice becomes much easier to act on.
What narrowing actually buys
The traditional argument is trust: a specialist seems more credible for the thing they specialise in. True, and vague enough to ignore.
The mechanical argument is stronger, and it has become more important recently. Both of the systems that now decide who gets found reward specific claims over broad ones, for the same underlying reason.
Google's helpful content guidance explicitly rewards content demonstrating first-hand expertise, including the kind that comes from having actually visited a place. Depth like that is straightforward to produce about a narrow subject and nearly impossible to fake across a broad one. The generalist site drifts toward summary because summary is all you can sustain about everything.
AI assistants apply the same preference from a different angle. Asked to recommend an agent for a specific situation in a specific place, an assistant selects a source it can defend quoting. A site claiming every property type in every neighbourhood matches that question weakly, because it matches all questions weakly. A site that says one thing clearly matches strongly, and strong match wins the citation. We described that selection process in how leads arrive from AI search.
So breadth is a disadvantage precisely where modern visibility gets decided. That is the part of this argument that is genuinely newer than the advice itself.
Choosing one without guessing
The usual failure is picking a niche that sounds appealing rather than one you have evidence for. Luxury, because the commissions are larger. Investors, because it sounds sophisticated. Both are reasonable ambitions and terrible starting positions if you have never done the work.
Do this instead. List your last twenty transactions with three columns: property type, area, and client situation. Look for the cluster. It is usually obvious and it is frequently not the one you would have chosen, because it reflects where your referrals actually come from and where your knowledge is genuinely uncommon.
That cluster is your niche, and its main advantage is that you can start demonstrating it immediately rather than building credibility from nothing.
Narrow enough to be specific, broad enough to eat
Both failure modes are real.
Too broad is the default: serving all of a metropolitan area, all property types, all client types. That claim is unfalsifiable and therefore unusable by anyone, human or machine.
Too narrow is rarer and more painful: a segment that turns over a handful of times a year cannot support a business regardless of how well known you are within it. Before committing, check actual turnover in your chosen segment from board data rather than intuition, the same discipline we recommended for choosing a farm area.
The usable middle is generally geography plus a situation. Townhouse sellers in two named districts. First-time condo buyers in one area. Specific enough to be defensible, broad enough to pay.
What sellers are actually weighing
Worth checking positioning against what clients say they want. NAR's 2025 Profile of Home Buyers and Sellers reports sellers' priorities when choosing an agent include help marketing the home, pricing it competitively, and selling within a specific timeframe.
None of those is a demand for a specialist. All of them are easier to prove if you are one, because your evidence comes from the same property type and the same streets as their situation. Specialisation is not a credential clients ask for. It is what makes the claims they do care about checkable.
Showing it without closing doors
On the site, the niche lives in two places and nowhere else. It goes in the identity layer, the plain sentence in your bio and homepage stating what you handle and where. And it gets its own genuinely deep pages, the ones that could only be written by someone who does this work.
Everything else stays open: your contact page, your service pages, your general availability. The niche does the attracting. The rest of the site serves whoever arrives, including the person with the detached house who was told you are good.
Testing a niche before committing to it
Positioning feels permanent, which is why agents postpone the decision for years. It is not permanent, and you can test one cheaply.
Write the two or three pages the niche would require, publish them, and change the sentence in your bio. That is a weekend, and it is reversible. Then leave it alone for two quarters and watch three things.
Whether enquiries arrive pre-qualified for that specific thing, which is the earliest genuine signal.
Whether people describe you using the niche when they refer you. When someone says you should talk to her, she does townhouses in the east side, the positioning has taken hold. This usually happens before any measurable change in transactions.
Whether you enjoy the work. This sounds soft and it decides whether the strategy survives, because a niche you find tedious will quietly get abandoned regardless of how well it performs.
If the answers are negative after two quarters, change the sentence back. The cost of testing is a weekend of writing and a bio edit. The cost of postponing the decision indefinitely is years of being findable for nothing in particular, which is what most agents actually choose by default, usually without noticing they have chosen anything at all.
Two quarters is also roughly the shortest window that means anything, since positioning works by repetition and a claim nobody has encountered twice has not been tested yet.
The takeaway
The advice to specialise is right, and the fear behind refusing it is answering a different question. Decide what you are known for, and leave what you accept alone. Choose the niche from your own transaction history rather than from ambition, check that the segment turns over enough to live on, and then state it plainly in the one or two places that shape how strangers and machines describe you. The agents struggling most with visibility are almost never the narrow ones; they are the ones whose site could belong to anybody.
