PublishedVerifiedEvery 6 weeksSources5 namedAuthored bySquareRank Team
Gemini · § 1.4.4 · Reference
Squarespace Gemini Checklist
Twelve items in five groups — Crawlers, Schema, Content, Freshness, Multimodal — each copy-paste codable from a Squarespace 7.1 site. Ship one group per session and the whole list takes about ninety focused minutes. The checklist is the operational version of the four-layer install the cluster hub walks through in narrative form.
This is the reference page. Each item is one action with a verification step. The leaves linked from each group cover the why and the how in depth; this page is the do-it list you keep open while you work.
§01How to use
How to use this checklist
Work through the groups in order. Crawlers first because Gemini citation eligibility starts with Googlebot reachability. Schema second because it is the foundation everything compounds on. Content third because that is where the most labour lives. Freshness fourth because Gemini leans on recency. Multimodal fifth because it is the Gemini-specific accent that distinguishes this cluster from the AI Overviews install. Each item has a verification step. Tick an item only when the verification passes.
The checklist assumes you have already worked through the AI Overviews checklist or the equivalent classical-SEO foundation: clean heading hierarchy, Person and Organization schema, 134-167 word lead at the top of every page. This list ships the Gemini-specific accents on top of that work. Business plan or above is required for the Code Injection items.
§02Crawlers
Group 1 — Crawlers (2 items)
The two crawler-layer items resolve the question most owners stumble over first: do I block Google-Extended, and does that affect Gemini citations? Total time on a Squarespace site is about five minutes including verification. The honest answer lives in the Google-Extended leaf; the actions live here.
01. Search-engine checkbox is on. Settings → Crawlers → "Allow search engines to crawl my site" should be on. Gemini citation eligibility starts with Googlebot reachability2. Verify: open robots.txt; expect no Disallow rule for Googlebot.
02. Google-Extended posture is intentional. Decide whether the AI exclusion checkbox is on or off based on your training opt-out preference, not your Gemini visibility hopes. Google's own documentation states the token does not affect Search inclusion or ranking1. The decision tree lives in the Google-Extended leaf. Verify: note the posture; revisit annually.
§03Schema
Group 2 — Schema (3 items)
The schema-layer items require Code Injection on Business plan or above. Total time on a fresh site is about twenty minutes including JSON-LD authoring and Rich Results Test validation. Article schema goes on every editorial page with an honest dateModified value. ImageObject schema goes on hero images. Both must validate without errors.
03. Article JSON-LD on every editorial page. Inject Article with headline, datePublished, dateModified (ISO 8601), author (Person, URL pointing to /founder/), publisher (Organization), and image (matching the ImageObject URL). The pattern is in the freshness leaf. Verify: Rich Results Test passes with zero errors.
04. ImageObject JSON-LD on the hero image of top pages. Inject ImageObject with contentUrl, caption, creator (Person URL), representativeOfPage true, width, height3. Pattern lives in the multimodal leaf. Verify: Rich Results Test passes; the contentUrl matches the visible hero image.
05. Article image field references the ImageObject URL. The two schema blocks should point at the same image URL, creating the entity handshake between text and visual. Verify: grep both JSON-LD blocks for the image URL; the values should match exactly.
§04Content
Group 3 — Content (2 items)
The content-layer items overlap with the AI Overviews install — section-extractable passages and named-source citations work across both surfaces. The Gemini accent is that chunk-level extraction is slightly more aggressive in Gemini's grounding pipeline, so the passage discipline carries a bit more weight here than it does on AIO.
06. Each H2 opens with a 134-167 word self-contained passage. Bolded lead at the top of every section, then expansion underneath. Gemini's grounding emits citations chunk by chunk; each section should hold up alone as a citation candidate. Verify: read the opening sentence of every H2; each should answer the section's sub-question without context from elsewhere on the page.
07. 2+ named 2026 sources per editorial page. Replace vague attribution with publisher name and date inline. Gemini's grounding leans on dated, attributable claims. Verify: count distinct named publishers per page; should be 2 or more, with at least one 2026 date.
§05Freshness
Group 4 — Freshness (2 items)
The freshness-layer items are where Gemini diverges most clearly from ChatGPT and Perplexity. Gemini's grounding favours newer content via the underlying Google Search ranking, and Squarespace's auto-emitted dateModified is unreliable enough to warrant the supplemental Code Injection block. The discipline is to be honest about when claims actually changed.
08. dateModified in JSON-LD matches the visible date strip. Both should reflect the most recent substantive update. The cosmetic-date pattern Google's December 2025 core update targeted4 does not pay; only real updates pay. Verify: view source on the page; the JSON-LD dateModified should equal the visible Published or Modified value.
09. Quarterly refresh calendar set for every editorial page. Calendar reminder 90 days after the page's modified date triggers a review: are claims accurate, do sources resolve, have any vendor names or stats moved? Verify: a calendar or spreadsheet shows the next-review date for at least your top ten pages.
§06Multimodal
Group 5 — Multimodal (3 items)
The multimodal items are the Gemini-specific accent that distinguishes this cluster. Gemini reads images natively via Embedding 2 rather than OCR, so alt text, file names, captions, and ImageObject schema each carry real weight. The work pays on hero images and key illustrations, not on decorative background imagery. Ship the layer on your top five pages, then expand as time allows.
10. Alt text on every informational image. Descriptive, 6-12 words, what the image actually shows. Not keyword-stuffed. Verify: view rendered HTML; the alt attribute is populated and reads naturally if you covered the image.
11. Descriptive file names on hero images. Hyphen-separated, lowercase, 5-8 words describing the content. Set the name on your local machine before upload; Squarespace's CDN preserves descriptive names more reliably when files are uploaded through the Files panel rather than the in-editor image picker. Verify: the served image URL contains a recognisable slug rather than a hash string.
12. Visible caption beneath the hero image. Matches the ImageObject caption property in language, anchors the image to surrounding text. Squarespace exposes the caption field directly on the image block. Verify: caption is visible to users and matches the schema caption value.
What done looks like
0
Rich Results Test errors across Article, Person, Organization, and ImageObject blocks site-wide.
