toolkit

Scenario · Content and search

What should we be targeting?

“We want to rank for more procedures — where do we start?”

"We want to rank for more procedures" is a request for a keyword list, and the useful answer is a short one with a reason attached to every entry — because chasing raw search volume alone picks terms the practice has no realistic chance of winning. This starts with what the site is already ranking for, since a term already showing up on page two is nearly always cheaper to move than a bigger term the site has never appeared for, and builds the shortlist from there rather than from volume alone.

What to ask for

Keyword research & planningKeyword research & planning on the SEMrush Analytics API.SEMrush researchOptional integration — SEO-research client for the SEMrush Analytics API.Search Console seriesOptional integration — reads Search Console's Search Analytics API and writes the data series nothing else has.Content briefContent brief generator — turns a keyword into a brief a writer can execute without further research: search intent and audience, a recommended title / meta description / H1, an H2–H3 outline, entities and questions to cover, internal-link anchors, and a word-count target.

See it work

A real run of Keyword research & planning:

pulling semrush keywords for "dermatologist bend oregon" …

INFO — 7 keyword(s) · dermatologist bend oregon
  easy wins (low KD + volume + intent):
  · botox bend oregon (440/mo, KD 18, transactional)
  · mole removal bend oregon (260/mo, KD 15, transactional)
  · restylane bend oregon (170/mo, KD 20, transactional)

  report: ./out/keyword-research-and-planning-keyword-research-and-planning.html
  json:   ./out/keyword-research-and-planning.json

The captured report, exactly as a run hands it to a client —open the full report ↗

Seen in the wild

Two of this domain's core service-name keywords are each split across multiple competing pages on its own site.

This is the toolkit's keyword-research skill, run against the SEMrush Analytics API. It requires a SEMrush API key (SEMRUSH_API_KEY, set in the toolkit's .env) — without one the skill is inert and exits with a one-line notice rather than a report. Given a domain, it pulls every keyword that domain ranks for — position, monthly search volume, keyword difficulty, intent — in one call. That's a different view than a single-page audit: it can't see two pages competing with each other, because it only ever looks at one page at a time.

The three keywords this run flagged, by real position

"ccip" — 1,900 searches/month3 pages rank for it: #27, #65, #77
"ocip" — 2,900 searches/month2 pages rank for it: #43, #81
"ccip meaning" — 390 searches/month3 pages rank for it: #41, #75, #88
Keyword difficulty for all three17–22 of 100 — genuinely easy

None of these are competitive terms. A term this easy, with content this specific already published, should be a single page-one result — not three pages tied on page three.

What the full ranking-keyword pull showed

  • "ccip" ranks from 3 different pages on the same site: positions 27, 65 and 77Search
  • "ocip" ranks from 2 different pages: positions 43 and 81Search
  • "ccip meaning" ranks from the same 3 pages again: positions 41, 75 and 88Search
  • 19 of the domain's 100 tracked ranking keywords show this same multi-page splitSearch
3 separate pagesWhat a manual, page-by-page read of the site sees
19 keyword collisionsWhat one full domain keyword-ranking pull found

The request is for keywords and the useful answer is a shortlist with a reason per row. Volume alone picks the terms the practice has no chance at; what is already ranking picks the terms it is closest to winning.

  1. Start with what the site already ranks for. Search Console shows the terms bringing impressions and where they sit. A term on page two with impressions is worth more than a higher-volume term the site has never appeared for, and it is cheaper to move.
  2. Research the terms properly. Keyword research gives volume, difficulty and the related and question terms around each. Where the competitive picture matters, pull the research on what competing practices rank for — the gap between their terms and yours is the shortlist nobody has to guess at.
  3. Group by intent, not by volume. Someone searching a procedure name is at a different stage from someone searching “how much does X cost” or “X near me”. A plan that mixes them produces pages that serve nobody.
  4. [manual] Agree the shortlist with the practice. Which procedures they actually want more of, and which they cannot staff. This is the filter no data provides, and applying it late wastes the content budget.
  5. Brief each agreed term. The content brief turns one term into something a writer can execute — intent, title, outline, entities, questions, word-count target, internal-link anchors. A keyword handed over without a brief comes back as a page about the wrong thing.
  6. Record the baseline. Current position and impressions per target term, so the work can be shown to have moved something later. Without this, “did the SEO work” has no answer six months on.

Page two is the cheapest ground

The instinct is to pick high-volume terms. The wins are almost always terms the site already appears for and ranks badly — the content exists, the authority exists, and the gap is usually a title, a heading structure, or four hundred words. Step 1 is first for that reason.

What this does not cover

Whether the practice can compete for what the data recommends, and whether it wants the patients a term brings. Search volume says people are looking; it does not say the practice has the authority to outrank a hospital system, or that the procedure is one they want more of. Both of those are business calls, and a keyword plan that ignores them produces content nobody funds.

← All scenarios