Most agencies pick blog topics from a keyword tool, a client's wish list, or whatever the account manager thought of in the shower. All three ignore the best data source you already have: the queries Google is already showing the site for, just not well enough to earn clicks.
The fastest way to find blog topics in Google Search Console is to filter queries to average position 11-20 — the top of page two — with at least 100 impressions over the last 90 days. Those queries are pre-validated: Google has decided the site is relevant, real people are searching, and you're one decent page away from traffic. Here's the exact workflow.
Why page 2 beats a keyword tool
A keyword tool tells you what the internet searches for. Search Console tells you what Google thinks this specific domain deserves to rank for. That's a different and much more useful signal.
The numbers make the case. A query sitting at position 14 with 2,400 impressions in 90 days is earning roughly 0.3% CTR — maybe 7 clicks. Move it to position 6 and you're looking at 5-7% CTR, or about 130 clicks for the same impressions. Move it to position 3 and it's closer to 250. You didn't create demand; you just stopped leaving it on the table.
Position 11-20 also means the hard part is done. The site has topical relevance, some link equity, and indexing. You're not starting a new topic from zero — you're finishing one.
The exact filters to apply
Open Search Console, go to Performance > Search results, and set it up like this:
- Date: Last 3 months. Shorter windows produce noisy positions; 16 months buries recent opportunities under old data.
- Search type: Web only. Image and video queries distort position averages badly.
- Metrics: Turn on all four — Clicks, Impressions, CTR, Average position.
- Tab: Queries.
- Filter 1: Position > 10.
- Filter 2: Position < 21.
- Filter 3: Impressions > 100.
Sort by Impressions descending. On a typical local service business you'll get 40-150 rows. On a mature content site, several hundred. Export to Sheets — the in-browser table caps at 1,000 rows and is painful to work in.
Two filters that sharpen the list fast
Add a query filter: doesn't contain [brand name]. Branded queries sit on page two for all sorts of boring reasons and they're not blog topics.
Then run a second pass with a custom regex filter to surface informational intent, which is where blog posts actually win:
^(who|what|why|when|where|how|which|can|does|is|are|should|do|best|vs|versus)\b
Set the query filter to "Custom (regex)" and paste that in. What comes back is a list of questions and comparisons your audience is already asking Google, on a domain Google already half-trusts to answer them. That is the single highest-conversion topic list you will find anywhere.
Decide: new post or update?
This is where most people get it wrong. Before writing anything, find out which URL is currently ranking for the query.
In Search Console, click the query row, then switch to the Pages tab. You'll see the URL (or URLs) earning those impressions. Three outcomes:
- A relevant page is ranking at 11-20. Update it. Expand the section that addresses the query, add it to an H2, tighten the title tag. This is the cheapest win in SEO — often 2 hours of work for a position 14 to 6 move.
- An unrelated page is ranking. A service page is picking up an informational query by accident. Write a dedicated post and internally link it from that page. You'll usually see the new post take over within 3-6 weeks.
- Multiple pages are splitting the query. That's cannibalization, not a topic gap. Consolidate first. Writing a fourth page on the same theme makes it worse.
Cluster before you write
Do not write one post per query. In your exported sheet, sort alphabetically and scan for groups. A plumbing client's page-two list might contain "how much does a water heater cost to replace", "water heater replacement cost 2024", "cost to install tankless water heater", and "is it cheaper to repair or replace water heater" — that's one 1,400-word post with four H2s, not four thin posts.
A good rule: if two queries would be satisfied by the same page without the reader feeling misled, they belong together. Aim for 3-8 queries per post. Sum their impressions — that becomes your opportunity estimate.
A simple scoring formula
Add a column in your sheet: Opportunity = Impressions × 0.05 − Clicks. That's a rough estimate of the clicks you'd gain at position 5-6 CTR. Sort descending and you have your publishing queue, ordered by actual business value rather than gut feel.
Example from a real audit: a 12-location dental group had 61 clustered page-two opportunities. The top five clusters totalled 38,000 impressions and 410 clicks. At a 5% CTR they represented about 1,500 additional monthly clicks — which reframed the content retainer conversation entirely. That's a far better client pitch than "we'll write four posts a month."
Sanity-check the SERP before committing
Search Console tells you demand exists. It doesn't tell you whether you can win. Open each shortlisted query in an incognito tab and look at the top 10:
- If the results are all YouTube videos or Reddit threads, a blog post won't break in. Skip it.
- If the results are all product pages or category pages, the intent is transactional — that's a service page brief, not a blog brief.
- If the top results are 2019 posts with thin content and no author, that's a green light.
- If an AI Overview answers the query completely in two sentences, discount the traffic estimate by half. Still worth doing for topical authority, but don't promise the clicks.
Make it a monthly ritual, not a one-off audit
The page 2 list regenerates itself. Every post you publish creates new impressions on new queries, half of which land at position 11-20 and become next quarter's topics. Agencies that run this on the first Monday of every month across every client never run out of briefs again.
The practical bottleneck isn't the data — it's turning 15 clustered opportunities into 15 briefs, drafts, approvals, and publishes across eight clients. That's a production problem, and it's where most content retainers quietly stall. Tooling helps: pulling Search Console data automatically, generating drafts in each client's voice, and scheduling publishing is exactly the kind of work an AI blog writer built for agencies should handle so your strategists stay on strategy. If you're deciding how much volume to commit to in the first place, this breakdown of blog posts per month for local SEO gives realistic numbers by vertical.

Nick Quirk

