Google Search Console keyword research: 202 AI impressions in 48 hours
One page Distribb published on August 2 got 202 impressions in Google's generative AI features within 48 hours.
We found the topic by pairing every Search Console query with the page currently earning its impressions. Google already associated the site with the query, but the ranking page covered a broader topic. That Google Search Console keyword research gave us a specific page to write.
The same process found two more pages. "Best Automated SEO Monitoring Tools" started ranking three days after publication. "Best AI Tools for Backlink Outreach" reached position 3 after nine days and collected 106 Search Console impressions.
If you want Distribb to find these query gaps, write the pages, publish them and track the results, start the three-day trial.
The six-step Search Console playbook
1. Connect Google Search Console to Claude Code
Give Claude or your agent read-only access to Search Console. Pull the queries, impressions, clicks and ranking page for each result.
Use the last three months. A longer range keeps one unusual day from controlling the list.
2. Run one prompt
Florian Darroman used this prompt:
Find keywords where we rank, but we don't have any specific content targeting this keyword.
Ask the agent to return every query beside the URL currently earning its impressions. That pairing tells you when Google associates your site with a topic but has no page written for the exact search.
3. Find the query gap
Open the page currently getting the impressions and compare its topic with the query.
Distribb had a page about automated SEO software. Search Console showed 273 impressions for "automated SEO monitoring" going to that broader page.
Someone searching for monitoring tools wants products that track rankings, site changes or SEO performance. The query needed a focused page instead of another section inside the broader software article.
4. Check demand and cannibalization
Validate the query before Claude writes anything.
- Check monthly search volume in Ahrefs or DataForSEO.
- Read the Google results to confirm what the searcher expects.
- Search your site for a page that already answers the query.
- Skip the query when a new article would compete with an existing page.
"Automated SEO monitoring" had 800 monthly searches and a keyword difficulty of 4 in Ahrefs.
5. Publish one page for the exact query
Use the query in the page title, URL and opening answer. Keep the article focused on that search.
Distribb published "Best Automated SEO Monitoring Tools" on August 8. The new page answered the monitoring query while the older article stayed focused on automated SEO software.
6. Check the new URL and repeat
Search the exact query after Google indexes the page, then filter Search Console by the new URL. Record the page's position, impressions and clicks.
Give those results to Claude before the next content batch. The automated SEO monitoring page was ranking after three days. The backlink outreach page reached position 3 after nine days.
The backlink outreach page had collected 106 Search Console impressions when we checked it.
This loop is one of the workflows behind 44,000 impressions in Google's AI features over three months. Run the check before each content batch so the queries Google already associates with your site enter the queue first.
Frequently asked questions
How do I find keyword gaps in Google Search Console?
Pair every query with the page currently earning its impressions. Open that page and check whether it answers the exact search or covers a broader topic. A query with impressions and no matching page is a candidate for a focused article.
Which Google Search Console date range should I use for keyword research?
Use the last three months for this workflow. That range is long enough to reduce the effect of one unusual day while staying close to the topics Google currently associates with your site. Export the queries, impressions, clicks and ranking pages from the same period.
How do I know when a keyword needs its own page?
A keyword needs its own page when the searcher expects a different answer from the page currently earning impressions. In the source example, a broad automated SEO software page earned impressions for "automated SEO monitoring." Monitoring tools have a narrower purpose, so Distribb published a separate article for that query.
How do I avoid keyword cannibalization when creating a new page?
Search your site before publishing and skip any query that an existing page already answers directly. Then compare the proposed page with the current ranking page. The two pages should serve different searches instead of competing for the same one.
What data should I check before writing a page for a Search Console query?
Check monthly search volume, keyword difficulty, the current Google results and your existing pages. The automated SEO monitoring query had 800 monthly searches and a keyword difficulty of 4 in Ahrefs. Those numbers supported the decision, but the separate search intent was the reason it needed its own page.
How long does a focused SEO page take to rank?
The two source examples started ranking within three and nine days, but the workflow does not promise a fixed timeline. "Best Automated SEO Monitoring Tools" ranked after three days. "Best AI Tools for Backlink Outreach" reached position 3 after nine days and had 106 Search Console impressions when checked.
Which Search Console query is already sending impressions to the wrong page on your site?
Try Distribb if you want the query research, writing, publishing and tracking handled in one workflow.
I like this approach because it starts with data you already have instead of guessing which keywords you should write about.
I've been working on building out blog content, and one of the difficult parts is deciding which topics deserve their own page and which should stay as part of a broader article.
Looking at queries where Google is already giving impressions but sending them to a less specific page seems like a useful signal. I also like the cannibalization check before creating anything new. It's easy to fall into the trap of publishing more and more pages when improving an existing one might actually make more sense.
The query-to-page pairing is brilliant as a measurement system: Google's impressions data is the most honest signal you have about what your site is actually good for in the market's eyes, not what you think it should be good for. Most teams never read that signal. They see "we got impressions on X" and treat it as noise—one more metric. But you're treating it as a decision framework: where is Google saying we've already won the trust, but the content doesn't match the question? That's measuring competitive signal, not just traffic. The hardest part is probably the discipline to skip queries that are just noise. The cannibalization check exists for a reason, but I'm curious: when a low-volume query keeps appearing in the 200-impression clusters, do you eventually archive or merge it, or leave it alone?
The strongest part of this workflow is the query-to-page pairing, because it turns Search Console from a reporting tool into a source of product decisions. I’d keep the three-month window, but add one guardrail: record the query’s current intent and the page’s current promise before publishing. That makes it easier to tell whether the new page won because it matched intent or simply because it was newer. For the 202 impressions, I’d also watch the ratio of impressions to meaningful clicks and returning visitors—AI visibility is useful, but the follow-through is the real signal. One question: after a focused page ranks, do you revisit the broader page and add a deliberate handoff, or keep the two pages completely separate to avoid cannibalization?
The interesting part here is the loop between what Google is already associating with the site and what you decide to actually build content around. Curious how you decide when a query deserves its own page versus being noise you leave alone?