Search volume is an estimate of how many times people search for a keyword in a given period, usually shown as a monthly average. It is the most familiar keyword metric and the most misread. The figures come from Google Ads data, clickstream panels and modeling, so they are approximations, they are often rounded or grouped, and different tools show different numbers for the same keyword.
Where the numbers come from
Google Ads' Keyword Planner reports search volume, but accounts without meaningful ad spend usually see broad ranges rather than specific figures, and Google groups close variants together. Third-party tools combine Google data, clickstream data from panels and their own models to produce a single estimate.
None of these sources is a precise count. Treat volume as an order-of-magnitude signal: this keyword is searched a lot more, or a lot less, than that one.
Why volume misleads SaaS teams
In B2B software, the searches that lead to revenue are often small. A query like “SOC 2 compliance software for startups” may show low volume, or none, while being made almost entirely by people with budget. Tools tend to under-count these long, specific searches because their samples are thin at the low end.
Volume also says nothing about intent. A large keyword can be dominated by students, job seekers or people looking for a free alternative.
Volume versus traffic
Ranking first for a keyword does not mean receiving its full volume as visits. Ads, AI Overviews, featured snippets and other SERP features take attention, and some searches are answered on the results page. A page also usually ranks for many related variants, so its real traffic can be higher or lower than the volume of its main keyword.
Search Console's impression counts are a better measure of real demand for searches you already appear for, because they come from Google's own logs.
How to use it well
Use volume to compare keywords within the same tool, to spot a topic with no demand at all, and to see seasonal patterns. Do not use it to forecast traffic or to rank your content plan on its own. Pair it with an explicit judgment of business value.
Example
The small keyword that sells
A made-up HR software company sees two candidates. “Employee onboarding” shows a high volume estimate. “Onboarding checklist software for remote engineering teams” shows almost none.
Their sales team says the second phrase describes their best customers word for word. They write that page first. Search Console later shows impressions for dozens of related variants that no tool had listed, which is a common pattern for specific B2B searches.
Common mistakes
Treating estimates as counts
Volume figures are modeled and rounded. Planning traffic targets from them sets up expectations no page can meet reliably.
Discarding zero-volume keywords
A tool showing zero often means its sample was too small to measure. If customers use the phrase, the search is real.
Ignoring close variants
One page can rank for many phrasings of the same need. Judging a topic by a single keyword's volume undercounts its demand.
Questions
Each tool uses different data sources, sampling and models, and some group variants differently. The differences are normal; pick one tool for comparisons and stay consistent.
Search Console shows impressions, which count how often your site appeared for a query. It is real Google data, but only for searches where you already show up, and Google filters out some rare queries for privacy.
Generally no. Keyword tools estimate searches on search engines. Questions asked in standalone AI assistants are not measured by standard volume figures, and good public data on them is limited.
