Keyword research: finding what people actually search for

Content written around a phrase nobody actually searches for can be excellent and still get zero organic traffic — not because it ranks poorly, but because it was never targeting a real, existing demand. Keyword research is the step that closes that gap, before a single sentence gets written.

The phrase you'd guess vs. the phrase people type

A developer writing about Python's requests library might instinctively title it around "Python HTTP client library" — a technically accurate description that's rarely how anyone actually searches. Real search behavior tends to be more specific and more problem-shaped: "python requests timeout," "python requests post json," "requests vs urllib." Keyword research tools (Google's own free Keyword Planner, or third-party tools like Ahrefs and Semrush) surface the actual phrases and their approximate monthly search volume — replacing a guess with real data before committing to a content angle.

Search volume against competition

A keyword with high search volume is only a good target if ranking for it is actually achievable — a generic, high-volume term ("python tutorial") is dominated by massive, long-established sites with enormous existing authority (part 15 covers domain authority directly). A newer or smaller site competing head-on for that exact phrase is fighting a battle it's very unlikely to win any time soon.

Long-tail keywords: lower volume, higher intent, real reach

High-volume, high-competition:  "python tutorial"
Long-tail, lower-competition:   "python requests library timeout not working"

A long-tail phrase — longer, more specific, lower individual search volume — is both easier to rank for and often more valuable per visitor: someone searching the exact phrase above has a specific, immediate problem, and content that solves it precisely converts that visit into genuine value far more reliably than a vague visitor arriving from a broad, generic term. A large share of all search volume, in aggregate, comes from the long tail rather than a small number of head terms — many individually small searches add up to real traffic overall.

Matching a keyword to the content that should rank for it

"python requests library"          → a tutorial or reference page
"python requests vs urllib"        → a comparison article
"python requests 401 error"        → a specific troubleshooting post

The same base topic splits into meaningfully different content types depending on the exact phrase — a broad reference tutorial doesn't serve a searcher looking for one specific error message nearly as well as a page written directly around that error would. Keyword research isn't just "find words with volume," it's understanding what kind of page a specific search actually expects to find — which is really the beginning of the search-intent question part 8 covers in full.

Where real keyword ideas come from

Beyond a dedicated tool, a page's own "People also ask" and "Related searches" sections in live search results are genuine keyword research, sourced directly from real search behavior at zero cost. So is a site's own internal search log, if one exists — the exact phrases visitors type into a site's search bar are keyword research a competitor has no access to at all.

Next: writing content that actually satisfies what a searcher was looking for — the difference between matching a keyword and matching the intent behind it.