Google Search Console holds 16 months of query data about your site, it is free, and it is the only place Google tells you the truth about what it thinks of your pages. Most founders open it twice a year, panic at a graph, and close it again.
That habit is the whole problem. The 50.1K clicks and 5.54M impressions we pulled on promptbuilder.cc did not come from a clever trick. They came from opening five reports every week and letting them decide what got shipped next.
Here are those five reports, in the order you should read them, and the exact decision each one is supposed to produce.
Treat Search Console as a weekly decision engine
Most SEO advice treats GSC as a reporting tool: something you check to see how you did. That framing is why it gets ignored. Reporting is backward-looking, and a founder with 40 things to do does not need another rear-view mirror.
Flip it. GSC is an input to next week's build queue. Every report below should end with you writing something into a list: a page to refresh, a URL to fix, a keyword to target, a link to chase. If a weekly review produces no queue items, you read it wrong.
Twenty minutes, once a week, same day every week. That is the whole commitment.
Set the date range to "Last 3 months" and turn on the compare toggle against the previous period before you read anything. Absolute numbers on a young site are noise. Direction is the signal.
Report 1: Performance, where next week's post comes from
The Performance report shows clicks, impressions, average CTR, and average position for every query and page. Turn on all four metric tiles, then work the Queries tab.
You are hunting three specific patterns.
Page-two queries. Filter to positions between 6 and 20. These are terms Google already considers you semi-relevant for, and moving one from position 12 to position 6 is a fraction of the work of ranking a brand new page. This is the single highest-ROI list in the entire tool.
High impressions, near-zero clicks. Google is showing you and nobody is clicking. That is almost always a title and description problem, not a ranking problem. Rewrite the title tag to match what the searcher actually wants and re-check in three weeks.
Queries you never targeted. Sort by impressions and read the phrasings you did not plan for. This is free keyword research from your own audience, and it is more reliable than any tool's suggestions because it is measured demand on your actual domain. If a query you never wrote about is pulling impressions, that is your next post, already validated.
Then switch to the Pages tab and sort by clicks. The pages that get outsized traffic tell you which format and topic to repeat. The pages that get none, on purpose or not, tell you what to stop making.
Report 2: Page Indexing, because unindexed pages cannot rank
Page Indexing splits every URL Google knows about into indexed and not indexed, with a reason for every exclusion. Read the reasons, not the totals.
The ones that matter:
- Discovered, currently not indexed. Google found the URL and chose not to crawl it. On a new domain this usually means low authority. On a big site it means crawl budget. Either way, more thin pages will make it worse.
- Crawled, currently not indexed. Google fetched the page and decided it was not worth storing. This is the thin-content filter talking. The fix is information gain, not resubmission.
- Duplicate, Google chose a different canonical. Two of your pages are competing. Merge them or differentiate them.
- Excluded by noindex / blocked by robots.txt. Almost always an accident on a small site. Fix immediately.
This report is non-negotiable if you run programmatic pages. An indexation rate that drops as you publish more is the clearest possible signal that your template is too thin, and it is a reason to stop generating and start adding unique value per page.
The number to track week over week is one ratio: indexed pages divided by total submitted. If it is falling, nothing else in this list matters yet.
Report 3: URL Inspection, the single-page diagnostic
Performance and Page Indexing tell you about the site. URL Inspection tells you about one page, and it is the tool you reach for when something specific is wrong.
Paste any URL into the search bar at the top of GSC and you get the indexing status, the canonical Google selected (which is often not the one you declared), the crawl date, the discovering sitemap, mobile usability, and any detected structured data.
Two habits worth building:
- Inspect every new post the day you publish it, then hit "Request Indexing" once. It is not a ranking boost, it is a queue jump for discovery. Do it once per URL, not repeatedly.
- Use "Test Live URL" when a page behaves oddly. It shows the rendered HTML Google sees right now, which is how you catch JavaScript that never hydrated, schema that silently failed to parse, or a canonical pointing somewhere unexpected.
If your JSON-LD does not appear in the live test's structured data section, it does not exist as far as Google is concerned, no matter how good it looks in your editor.
Report 4: Links, the reality check on your authority
The Links report gives you external links (who points at you, and to which pages) and internal links (how your own site distributes authority).
On the external side, look at "Top linking sites" and ask one question: is this list growing? A flat referring-domain count over three months means your content is not earning links on its own, which is the normal state of a young startup. That is exactly the gap directory submissions are built to close, and it is why the fastest DR lift for a new domain comes from being listed rather than from writing more.
The internal side is the part nobody reads, and it is the free win. Sort by "Top linked pages" and find your money pages near the bottom. Any page with zero or one internal link is effectively orphaned: Google struggles to find it and passes it almost no authority. Every post you publish should link to at least two older posts and get linked from at least one.
Internal linking is the only authority lever you fully control. Use it.
Report 5: Core Web Vitals, the one you check least
Core Web Vitals reports field data from real Chrome users: Largest Contentful Paint, Interaction to Next Paint, and Cumulative Layout Shift, grouped into Good, Needs Improvement, and Poor buckets by URL group.
Be honest about the priority here. Page experience is a tiebreaker, not a ranking engine. A fast page with weak content still loses to a slow page that answers the question better. So this report gets five minutes a month, not twenty minutes a week, with two exceptions worth acting on immediately:
- A URL group that flips from Good to Poor after a deploy. That is a regression you shipped, and it is cheap to catch now and expensive to catch in three months.
- Layout shift on mobile. It wrecks conversion long before it wrecks rankings, so fix it for the revenue reason and take the SEO benefit as a bonus.
New sites often see "not enough data" here. That is fine. It means too little traffic for field measurement, and it will populate as you grow.
The 20-minute loop that turns a launch into compounding growth
Same day every week, in this order:
- Performance, last 3 months, compared. Pull the positions 6 to 20 list. Pick one page to refresh. (5 min)
- Performance, Queries tab. Note any query with impressions you never targeted. Add it to the content queue. (3 min)
- Page Indexing. Check the indexed ratio against last week. Investigate only if it dropped. (3 min)
- URL Inspection. Inspect last week's new post, confirm it is indexed and the schema parsed. (2 min)
- Links. Check referring domains, then find one orphaned page and link to it from a post that already gets traffic. (5 min)
- Write the queue. Two items: one page to refresh, one page to write. (2 min)
That loop is the difference between a launch spike and a compounding curve. A launch gets you a week of traffic. Ship, measure, refine, repeated weekly, is what turned a zero-traffic domain into 50.1K monthly clicks, and it is the same review I run on nicklaunches.com, the launch directory where those products get listed in the first place.
None of it requires a paid tool. It requires the same 20 minutes, every week, on a cadence you actually keep.
Want the full step-by-step version of this playbook, plus a new breakdown every week? Grab the free guide, and follow along on X.