← Back to Guides · SEO
The Title-Tag-First Rule: A Second Search Console Audit Pass (2026)
Short answer: A title tag or meta description rewrite is almost always cheaper than a content rewrite, and it is the first thing to try when Search Console shows impressions with no clicks or a near-miss position in the teens-to-thirties. This is a follow-up to our first Search Console audit: the same method, run again a few weeks later, with one new page in the mix (/commands) and a sharper rule for when a snippet rewrite is enough versus when the page itself needs a structural fix.
If you already read the first audit post, the short version of what's new here: /commands showed the same "0% CTR at a double-digit position" shape as /skills did the first time, which means the pattern generalizes — it isn't a one-off. We also tightened the rule for when to stop at the title tag versus when to also touch the first paragraph, because that distinction wasn't obvious the first time through.
The four rows this pass
| Page | Query | What Search Console showed | What we changed |
|---|---|---|---|
| /skills | cursor skills | 47 impressions, 0 clicks, average position 66.1 | Title tag rewritten to lead with "Cursor Skills" as an exact-match phrase plus the year, meta description tightened to state a concrete count and outcome ("47 ready-to-install Cursor skills..."), a definition of "cursor skills" added to the first 100 words, and an internal link added to the Claude Code skills listicle to separate informational intent (what is a skill) from catalog intent (browse skills) |
| /compare/cursor-rules-vs-skills | cursor rules skills docs | 13 impressions, 0 clicks, average position 29.1 | Added a short "official docs summary" section near the top that cites Cursor's actual docs structure directly, and retitled the meta description to say "official docs explained" — the word "docs" in the query is a strong signal the searcher wants a reference answer, not commentary |
| /commands | cursor commands | 5 impressions, 0 clicks, average position 32.2 | Added a short definition/benefit blurb above the fold that answers "what are cursor commands" directly, and tightened the title tag to put "Cursor Commands" verbatim near the front |
| /docs | agentickit | 4 impressions, 0 clicks, average position 3.8 | Rewrote the title to state explicitly what the page documents instead of a generic label, since a near-top branded position with 0% CTR means the SERP snippet isn't giving a navigational searcher a reason to click through rather than typing the URL from memory |
Three of these four rows repeat a shape we already described in the first audit: non-trivial impressions, 0% CTR, position in the 30s-to-60s range. The fourth (/docs) is the branded-query case — near the top of the results, still 0% CTR, which is a pure snippet problem rather than a content problem.
A third pass later found /skills, /commands, and /docs still stuck at roughly these same numbers, and traced it to a gap this round's fixes didn't address: the pages weren't naming Cursor explicitly enough for Cursor-specific queries.
The rule we sharpened this time: title tag first, content second
The first audit conflated two different fixes under "answer the query in the first 100 words." Running the method a second time made the actual sequencing clearer, and it's worth stating directly:
- If the page is a near-miss (position 15-40) and the query contains a reference word like "docs," "list," or a product-ish noun, the SERP snippet is doing most of the work. Start with the title tag and meta description — put the literal query phrase near the front, and make the description state a concrete outcome (a count, a time, a result) rather than a vague benefit. Don't touch the page body yet.
- If the page is stuck deep (position 50+) with real impressions, the snippet alone won't move it — Google is still deciding whether the page's content matches the query at all. That's when the first-100-words definition and the internal link matter: they're relevance signals for the crawler, not just click bait for the SERP.
- Check both, but in that order. A rewritten title tag on a position-66 page (like
/skillshere) still needs the content fix too — the snippet change buys a better CTR if it ranks, the content change is what moves the ranking in the first place.
That ordering is why /skills got both treatments (title/meta plus an on-page definition and a new internal link) while /compare/cursor-rules-vs-skills only needed the "docs summary" section added near the top — it was already ranking on page 3, so the job was credibility and snippet framing, not ranking from zero.
Why /commands needed the same fix as /skills, almost verbatim
/commands at position 32.2 with 5 impressions and 0% CTR is structurally identical to what /skills looked like before its first fix: a catalog page that Google already associates with the right query, but that doesn't answer "what is X" fast enough for either a searcher or a crawler to be confident about relevance. The fix was the same move — a short, plain-language blurb above the fold that defines "a Cursor command" before the catalog grid starts, plus a title tag that puts the query phrase verbatim near the front instead of relying on the brand name to carry it.
This is the actual value of re-running an audit on a schedule rather than once: it turns a one-off diagnosis into a repeatable checklist you can run against every new page as it accumulates impressions. The first time, "definition in the first 100 words" was a fix for one page. The second time, it's a rule you apply on sight.
The listicle-vs-catalog split, and why it matters for internal linking
One detail from the /skills fix is worth calling out on its own: the internal link we added points from the catalog page (/skills, which lists every skill) to a listicle-style post (the rules-vs-skills-vs-commands-vs-hooks comparison). That's deliberate query-intent separation. Someone searching "cursor skills" could want either "show me the actual skills" (catalog intent) or "explain what a skill is and how it's different from a rule" (informational intent). Letting one page try to serve both usually means it serves neither well. Splitting them and linking between them lets each page rank for its matching intent while still keeping the user one click from the other.
This matches a pattern that shows up across how the wider ecosystem documents the Claude Code / Cursor agent toolchain: comparison and "which one do I use" content tends to live separately from install-it-now catalog content. A recent practical comparison of Claude Code skills, subagents, hooks, and workflows makes the same split explicit — it's framed entirely around "which mechanism for which job," not as a catalog of specific skills to install, and it links out to install-focused resources rather than trying to be one itself. Our own rules vs. skills vs. commands vs. hooks post plays the same role for /skills and /commands: the decision-table page for intent, the catalog pages for action.
Checklist: running this audit a second time
- Re-pull Search Console → Pages, same filters as last time (impressions descending, 90-day window).
- Diff against the last audit's rows. A page that showed up before and still shows the same shape (impressions, 0% CTR, same position band) means the first fix either hasn't propagated yet or wasn't applied with enough weight — give it longer before re-touching it.
- Any new row with the same shape (our
/commandscase) gets the same treatment as last time's matching case, not a bespoke new approach. Consistency across similar pages is itself a relevance signal. - For branded queries, re-check CTR in isolation from non-branded rows every time — a branded query's CTR tells you about your SERP snippet, not your ranking, and the two should never be fixed with the same lever.
- Log the date of each change next to the row it was applied to, so the next audit pass can tell which rows are "waiting on a fix to propagate" versus "still unaddressed."
FAQ
How is this different from the first Search Console audit post? The method is the same, but this pass has one new page (/commands) with the same 0%-CTR-at-a-near-miss-position shape, which let us generalize the fix instead of treating it as one-off, and it adds a clearer rule for sequencing title-tag fixes before content fixes.
Should every page get both a title-tag rewrite and a content rewrite? No — near-miss pages (position 15-40) with a reference-style query usually only need the snippet (title/meta) fixed. Deep-position pages (50+) need the content/relevance fix too, since the snippet alone won't move a ranking that low.
Why add an internal link instead of just adding more content to one page? Because the query had two distinct intents (catalog vs. informational) bundled into one page. Splitting them across two pages that link to each other serves both intents better than one page trying to do both.
How often should I re-run this audit? Every 2-4 weeks is enough to see whether a fix propagated, since Search Console data itself lags and position is noisy at low impression counts — re-running weekly just adds noise, not signal.
Sources
- Google Search Console Performance data for agentickit.co (first-party, not published externally)
- Claude Code Skills vs Subagents vs Hooks vs Workflows: Which to Use in 2026 — referenced for the catalog-vs-decision-content framing, not quoted
Written by Rakshit Yadav. Published 2026-10-29, last updated 2026-10-29.
Related Tools & Agents
Run this checklist automatically
The on-page SEO skill and /seo-audit command apply this exact method to any page you point at.
See the on-page SEO skill →