What is share of search

Guide · Glossary & Definitions · 5 min read · last verified 2026-07-27

Share of search is the proportion of category search interest that your brand name accounts for — of all the searches buyers run for brands like yours, the fraction that names you specifically. It is computed from observable search-volume data rather than from surveys or panels, which makes it one of the few brand-strength indicators a small team can track without commissioning research. If buyers in your market search for you and for your alternatives, share of search is your slice of that combined attention, expressed as a proportion and tracked over time.

What the metric actually captures

Typing a brand name into a search box is a small, deliberate act. Buyers rarely arrive at your name by accident; something put it there first — a colleague's recommendation, a podcast mention, an article, a half-remembered ad. Share of search adds those deliberate acts up across an entire category. It works a little like taking attendance in a room you were never invited into: you get to see which names buyers call out, and how often, without asking anyone a single question.

Two properties make this valuable. First, branded search interest reflects memory and intent arriving under their own power — the searcher chose your name before your budget had a say in that particular moment. Second, the metric is relative. Because it is a share, it moves when competitors rise or fade as well as when you do, which forces a market view rather than a mirror view. A brand whose absolute search volume holds flat while the category doubles is losing ground by definition, and share of search is the metric that says so out loud.

How to measure share of search

The mechanics are simple enough to run in a spreadsheet; the discipline is what makes the number trustworthy.

• Define the brand set. List your brand and the alternatives buyers genuinely weigh — the names that come up in your actual deals, not the giants you aspire to beat or the minnows that flatter your ranking. The honesty of the set decides the honesty of the number.
• Gather search-interest data for each name. Public search-interest tools expose relative volumes for brand terms. Disambiguate where necessary: a brand named after an ordinary word needs qualifying terms, or the data may count gardeners and grammarians among your buyers.
• Compute the share. Your brand's search interest divided by the combined interest of the full set, calculated on a consistent interval — monthly is a common practitioner choice.
• Lock the definition. Same brand set, same terms, same data source, every period. When the number moves, you want to know the market moved — not the method. Add or remove a competitor only with a written note, and expect a visible step in your trend when you do.
• Read the trend, not the level. A single reading tells you very little, because the level depends heavily on how you drew the category. Direction over quarters is where the signal lives.

Is share of search a leading indicator?

This is the claim that made the metric famous, so it deserves careful wording. Practitioners who track share of search over long periods often report that sustained movements in it precede movements in market share — that the research phase of a market shifts before the buying phase does. The proposed mechanism is at least plausible: buyers tend to investigate before they purchase, so a change in who gets investigated would naturally appear before a change in who gets paid.

Treat this as a practitioner observation, not a law. The strength of the relationship appears to vary with category, purchase cycle, and how considered the purchase is, and a metric that leads in one market may merely wobble in another. The sober position: share of search is worth tracking as an early signal, worth investigating when it moves, and not worth converting into revenue forecasts with invented confidence. If you need targets attached to it, set them the way How to set growth targets without fake precision describes — as directional commitments, not decimal make-believe.

What share of search does not see

The metric only registers buyers who already know your name. The larger population searching generic category phrases — best tools for this, how to fix that — is invisible to it, and that unbranded territory is usually where your next customers are standing. Share of search reports how you are doing among the already-aware; it says nothing about how the unaware are being introduced, which is a different measurement problem covered in How to measure brand awareness without surveys.

It also misses the questions that have stopped being searches. A visible portion of buyer research has been observed shifting into AI assistants, which answer category questions directly and name brands inside the answer — no results page, no click, no search-volume trace. Which brands current engines name, and on what grounds, is behaviour that keeps changing between model updates, which is exactly why it needs measuring rather than assuming. Share of voice vs share of answers walks through that adjacent metric and how the two complement each other.

Making it operational

Tracked alone, share of search is trivia with a chart. Inside a small system it earns its place: one position metric — this one, share of answers, or both — reviewed on a fixed cadence, with an agreed question attached: what would we do differently if this fell for two consecutive periods? If your team has chosen a single guiding measure — What is a north star metric explains that discipline — share of search usually sits beside it as a leading position indicator rather than replacing it.

The measurement discipline matters more than the metric itself. Locking the brand set, holding the method still, and re-measuring on a schedule is the same discipline Magrios applies to AI answers — a fixed benchmark of buyer questions, re-scanned over time, so that when a number moves you can trust the market moved. Whichever surface you measure, the principle holds: an unstable ruler measures nothing.