Mine Search Console for the queries you already half-rank for
Google has already told you where it trusts the domain. The cheapest page to rank is the one filling a gap it named.
Inside this
- Pull the query-to-page table for the last 90 days20m
Query, landing page, impressions, clicks, position. Three months so one freak day cannot set the agenda, and leave off the newest days: Google marks them preliminary, with a dotted line on the chart.
- List the queries whose landing page is about something else30m
Impressions against a page that only half answers the query. That mismatch is the whole signal.
- Kill the gaps that would cannibalise a page you already have30m
Check volume and difficulty, read the live SERP for intent, and drop any query an existing page already answers. Splitting one intent across two pages ranks neither.
- Publish one page per surviving query, then re-check its position4h
Exact query in the title, URL and opening answer. Log position and impressions before the next batch — this loop is only as good as its feedback.
- Push the queries sitting on page two before writing anything new2h
Position 11–20 with impressions is a page Google already half-trusts. Exact query in the title, H1 and first hundred words, then three internal links to it. An afternoon, and cheaper than any new page.
- Rewrite the titles that get seen and not clicked1h
High impressions and few clicks is a title problem, not a ranking problem. Promise what the query asked for in the title and description, then read the click-through rate again in a month.
- Take your own name out of every Search Console view15m
On a young site most clicks are people typing the name, which says nothing about whether the content ranks. Google’s branded filter needs a top-level property with enough impressions, so a new site usually filters the query instead: “not containing” the name and its misspellings. Read every step in this group on that view.
- When a page-two query will not move, count the links behind the pages above it1h
Once the title and internal-link push has had four weeks, compare the referring domains to your page with those of each of the top three, page against page, not site against site. A gap you can close in a quarter is a link job for that URL, never your homepage. A gap you cannot close is a reason to pick a smaller query.