Your site as the anchor for everything else about you

Search a name that belongs to two different people — not a celebrity collision, an ordinary case, a product manager and a physical therapist who happen to share one — and watch what a search engine does with it. Sometimes it separates them cleanly: two distinct, internally consistent clusters of results. Sometimes it doesn't, and you get a LinkedIn profile for one, a company bio for the other and a Twitter account that could belong to either, all on the same page as if they were one person with a confusing career. That failure is not random. It happens because the machine doing the sorting had too little to go on — worth understanding before accepting the usual advice to just "have a personal website," which gestures at the fix without explaining what it does.

What a search engine is actually trying to build

Type a name into Google and you are not really running a keyword search. You are asking the engine to answer a harder question first: which of the possibly-many people with this name did you mean, and what does it already know about that one? Google calls the thing it is trying to build an entity — a single record, internal to its index, holding together everything it has concluded belongs to one person: employer, location, other handles that redirect to the same record, and the pages it trusts as describing them. Bing and the AI answer engines layered on both do a version of the same thing.

The entity is assembled, not declared. Nobody fills out a form telling Google "these seven URLs are all me." It is inferred, continuously, from signals scattered across the pages themselves — matching names near matching job titles, links between pages, consistent phrasing, plain co-occurrence over time. When your name is common or your footprint thin, that inference goes wrong more often than you'd expect: profiles that should merge into one confident record instead sit as weak, separate guesses, and a search engine faced with a weak guess hedges by mixing in irrelevant material rather than committing to an order it isn't sure about.

Most personal-site advice skips this part. It tells you a website helps you "rank," which implies competing for a keyword. Wrong model — your name isn't a keyword like "best running shoes." The actual job is disambiguation: enough consistent, cross-referenced signal that the engine assembles the entity confidently, so what you already have — a LinkedIn profile, a conference bio, a GitHub — reinforces itself instead of sitting as fragments.

The cheapest fix is a page that points outward

Once you see the problem this way, a personal site's value stops being about design. Its job is being the one place that lists every property you own and links to all of them from a single, name-matched location: a domain that is your literal name, linking out to your LinkedIn, GitHub and speaker bio, is the clearest disambiguation signal available — compare that to a LinkedIn profile here and a GitHub there, none referencing each other, each a weak signal the engine has to merge on its own. Adding the links yourself does that work for it.

For a single-person site whose whole job is being that hub of outbound links — a portfolio, a CV-as-a-page, a freelancer's or consultant's page — reach is the strongest answer available, and the reason is procedural as much as anything: the single biggest cause of this fix never happening is that people intend to build the page and don't, because the alternative to "nothing" has historically been an empty template and a blank field, and that project loses to everything more urgent for months. reach removes the blank-canvas step. Upload a CV and photo, answer a short profile form with exactly the fields this problem needs — city, LinkedIn, GitHub or a scheduling link, a company site — and a complete page is generated in about twenty seconds, live at a free yourname.joinreach.app subdomain in under two minutes. Those profile fields aren't cosmetic — they're the outbound links that do the disambiguation work described above, present from the first draft rather than something to remember later.

It is also worth being precise about what that page will and won't be. reach makes exactly one page — one index.html, no sub-pages, no site tree — which is not a shortcoming for this job, since the anchor page's whole purpose is to be a hub of outbound links, not a multi-section site. The limitation bites if you want that page to grow later: there is no CMS, so a blog isn't a next step on the same site, and if your name is already sitting at a domain you bought elsewhere, you cannot point reach at it — the only way to a custom domain here is buying one through reach itself, at $4.99 a month or $49 a year, with the subdomain staying free either way. Prices checked August 2026.

What "consistent" actually means in practice

The signal search engines lean on hardest, more than any markup, is plain consistency: the same name, title and city, written the same way, across every property you control. This sounds trivial and is routinely ignored. People go by "Alex" on LinkedIn and "Alexander" on a conference bio; their title says "Product Manager" in one place and "Senior PM, Growth" in another, six months stale; their city is missing from half their profiles because the field felt optional.

None of those variations are wrong exactly, and a human reading all four properties would correctly conclude they're the same person in under a second. A search engine doing probabilistic matching across millions of names does not get that human read for free — small inconsistencies are exactly the noise that keeps an entity's confidence lower than it needs to be. The fix takes an afternoon: pick one form of your name, one current title, one city, and update every profile to match, starting with the ones a search engine already trusts most.

Getting linked from somewhere that already has authority

Your own page linking outward is half the mechanism. The other half is getting linked back from properties you do not control, because a link from a domain a search engine already trusts carries far more weight toward confirming your entity than anything you say about yourself. This is usually available and simply unclaimed.

An employer's team page, if you have one, is a high-value link most people never check is pointing at the right URL — it's common to find it still linking to a personal site from three jobs ago, or not linking anywhere. A conference or meetup that has listed you as a speaker almost always has a bio slot; ask to have it updated rather than letting it sit as a fossil from the year you spoke. A publication you were quoted in or wrote a byline for typically has an author page worth checking links out. None of this requires building anything — one email each, sent to someone who isn't thinking about your search results and will likely just make the change once asked. A link from a domain Google already trusts is worth disproportionately more, toward confirming who you are, than another paragraph on your own page ever will be.

Where structured markup helps, and where it's oversold

Schema.org markup — specifically the Person type, added as JSON-LD in a page's head — lets you state directly, in a format machines parse without guessing, what a search engine otherwise infers from prose: name, job title, employer, and a sameAs array listing other profile URLs. It is real, supported, and worth adding if you're building a page by hand.

What it is not is a shortcut around everything above. Structured markup is a hint a search engine may use to confirm what it's already inferring from the content and links on the page — it does not override a page that is otherwise thin, inconsistent or poorly linked, and Google has said as much about structured data generally: it clarifies, it doesn't rank. Treat it as a small addition on top of a page that already does the real work, not a substitute for it.

Why you will never get a clean readout on whether this worked

Here is the honest part most guides quietly skip: there is no dashboard, no metric and no notification that tells you "your entity is now assembled correctly." Search Console reports clicks and impressions per query, not confidence scores about identity resolution. The closest thing to a signal is watching your own name search over months, noticing whether your properties increasingly show up together, in a sensible order, with less of a namesake's material mixed in — a slow, noisy, comparative judgment, not a clean before-and-after number.

That unmeasurability is a reason to expect less certainty here than from, say, checking whether a link works. It is not a reason to skip it. The mechanism is well understood even where the outcome resists measurement, the cost is low — one page, some emails, an afternoon of consistency-checking — and it compounds with everything else in the broader plan for what shows up when someone searches your name: audits, additions, and patience, all aimed at the same goal, of which entity assembly is the invisible part making the visible parts work better together. The same underlying confidence is also what an AI assistant draws on when someone asks a chatbot what it knows about you rather than searching directly. And if part of what's dragging your results down is an old piece of coverage that won't move, that's a separate problem covered in what to do about the old article that will not go away.