Share of search is a brand's share of total category search volume over a defined period. If a category generates one million brand-related searches in a month and a brand accounts for 120,000 of them, its share of search is 12%.
The metric is useful for three reasons: it is measurable consistently across competitors without their cooperation, it is difficult to inflate, and it frequently moves before sales do.
Share of search versus share of voice#
The two are routinely confused and behave very differently.
| Share of search | Share of Voice | |
|---|---|---|
| Who generates the signal | Consumers | Brands, media, and consumers |
| Can it be bought directly | Not really | Yes |
| What it indicates | Demand | Attention |
| Typical relationship to sales | Leading | Coincident or lagging |
Share of voice rises when a brand publishes more, buys more amplification, or activates more creators. Share of search rises only when more people independently decide to look for the brand. That makes the second much harder to manufacture, and it makes the gap between them the diagnostic worth watching: a brand whose share of voice climbs while its share of search stays flat has bought attention without converting it into demand.
Why it leads sales#
For any category with a consideration step, searching happens during evaluation and buying happens at the end of it. The lag between the two signals is approximately the length of the category's consideration cycle — short for replenishment and impulse purchases, substantially longer for high-value or infrequent ones.
This is what makes share of search a leading indicator rather than merely another way of describing the present. The important qualification is that the lead time is category-specific and must be established empirically. Measuring the historical relationship between the two series in your own category is the only way to know how far ahead the signal runs, and assuming a generic lag is a good way to act early on a movement that never arrives.
Measuring it so the series stays comparable#
Four things have to be fixed and held constant:
The category definition. The denominator is the set of brand terms that count as "the category". Add or remove a brand mid-series and each brand's share moves for reasons unrelated to demand.
The term set per brand. Include common misspellings, abbreviations, and both English and local-language forms. A brand with a widely used local-language name whose term set omits it will be understated systematically — not randomly, which means the error never averages out.
The platform. Marketplace search, general web search and in-app social search capture different intents. Marketplace search sits closest to purchase; social search sits closest to discovery. A blended figure across all three describes no single behaviour, so pick the platform that matches the question and stay on it.
The window. Long enough to average out promotional spikes. Weekly readings in a category with an active promotional calendar are mostly noise.
Pitfalls#
- Generic brand names. A brand whose name is an ordinary word absorbs unrelated search volume and is overstated unless the query set is filtered. This is the single largest source of bad share-of-search figures.
- Campaign spikes. A burst of activity produces a short-lived movement that does not indicate durable demand. Read direction over a sustained window, not a level at a point.
- Direct-navigation habit. An established brand that consumers go to directly, without searching, can read lower than a challenger whose buyers are still evaluating. The metric measures active consideration, which is not the same as strength.
- Seasonality. Compare like periods against like periods, or use a rolling window.
Using it well#
Share of search earns its place as one input among several, not as a standalone verdict.
- Read it against sales in the same category. Divergence is the signal — search rising while sales are flat means either a lag or a conversion problem, and which one it is can be resolved by looking at whether the gap closes.
- Read it against Share of Voice. Attention without demand is a distinct and correctable problem.
- Read it at category level, not brand level alone. A brand's share can fall while its absolute search volume rises, if the category is expanding faster. Those are different situations and only one is a problem.
- Treat sustained movement as signal and single periods as noise.
For the social-platform side of the same question, where discovery search behaves differently again, see Xiaohongshu Marketing. For the forward-looking use of demand signals, see Trend Forecasting.
Where to look next#
For the attention-side counterpart, see Share of Voice. For discovery-led search behaviour, see Xiaohongshu Marketing. For where demand signals come from, see Market Intelligence Data Sources.
Common questions#
What is the difference between share of search and share of voice?#
Share of search measures what consumers actively look for; share of voice measures how much is being said, including by brands themselves. That difference in who generates the signal is what makes them behave differently. Share of voice can be bought — more publishing, more paid amplification, more influencer activity all raise it directly. Share of search is far harder to inflate, because it requires someone to independently decide to look for you. A brand whose share of voice rises while share of search stays flat has usually bought attention without creating demand, and the gap between the two is one of the more informative diagnostics available.
Why does share of search often move before sales?#
Because for considered purchases, searching happens during evaluation and buying happens at the end of it. The lag between the two is roughly the length of the category's consideration cycle: short for impulse and replenishment categories, considerably longer for high-value or infrequent purchases. That gap is what makes the metric useful as an early indicator — a sustained movement in share of search is a reasonable leading signal for sales in the same category. The caveat is that the lead time is category-specific and has to be established empirically rather than assumed.
How should share of search be measured to stay comparable?#
Fix four things and hold them: the category definition, meaning the full set of brand terms that count as the denominator; the term set per brand, including common misspellings and both English and local-language forms, since omitting a major variant understates that brand systematically; the platform, because search behaviour on a marketplace differs from search on a general engine or a social platform; and the time window, long enough to average out promotional spikes. Change any one of them mid-series and the trend measures the change in method rather than the market.
What are the main pitfalls?#
Three recur. Brand names that are ordinary words attract unrelated search volume and are overstated unless the query set is filtered. Promotional and campaign spikes produce short-lived movements that do not indicate durable demand, so single-period readings are unreliable. And a brand with a strong direct-navigation habit — consumers going straight to it rather than searching — can be understated relative to a challenger whose buyers are still in an evaluation mode. None of these invalidate the metric, but each argues for reading direction over a sustained window rather than a level at a point in time.