Four pillars.
Four ways a machine reads a store.
SEO, AEO, GEO and AIO — how search engines, answer engines, generative engines and systems that act on a record each read a Shopify catalogue, and how each is scored. Every public scan scores all four.
- Order
- SEO → AEO → GEO → AIO
- Pillar layer
- v1.0 · 29 checks — SEO 9 · AEO 5 · GEO 5 · AIO 10
- Weight totals
- SEO 100 · AEO 86 · GEO 82 · AIO 108 — relative within each pillar
- Composite
- Rubric v1.0 · 9 catalogue checks, weights totalling 95
- AI agents checked
- 10 — GEO pillar and composite
- Search crawlers
- Googlebot, Bingbot — SEO pillar only
- Fix classes
- 4 automatic · 20 assisted · 5 guided — the design
- Fixes
- Planned fixes · Planned verification
SEO → AEO → GEO → AIO
Conventional search, then answer engines, generative engines and the systems that act on a record. Each pillar assumes the one before it has been cleared; each is scored on its own inputs, so each can fail on its own.
One discipline each.
Always in this order.
Each definition below is the one the scanner, the report and every Rank Sniper page use — rendered, never paraphrased.
SEO
Search engine optimisation. Whether conventional search engines can crawl, parse, and rank the catalogue.
9 checks · weight sum 100
Engage →AEO
Answer engine optimisation. Whether the catalogue can be lifted as a direct answer to a shopper's question.
5 checks · weight sum 86
Engage →GEO
Generative engine optimisation. Whether generative engines will cite this store when composing a recommendation.
5 checks · weight sum 82
Engage →AIO
AI optimisation. Whether an AI system can correctly parse and act on the record without guessing. Includes transactability as a named sub-score.
10 checks · weight sum 108
Engage →Why four
and not one.
One number averages four different failures into something nobody can act on.
A single “AI readiness” score hides which system is turning a store away. Search engines, answer engines, generative engines and agents that transact read different parts of a Shopify store, through different crawlers, for different purposes — and they fail for different reasons. A catalogue can be perfectly indexed and still unbuyable by an agent, or richly structured and still unread by a generative engine because of one robots.txt group.
Four pillars keep those failures apart. Each is scored separately, from its own checks, with its own weights and — where one applies — its own crawler rule. A store gets four answers to four questions, and the public result flags the weakest pillar as the first place to work.
Why these four? They are the distinct jobs a machine does with a catalogue that a storefront scan can observe: find and index it, quote it, draw on it for a composed answer, act on it. The acronyms have no consensus meaning across the industry, so Rank Sniper publishes its own definitions, renders them verbatim wherever they appear, and lists every term in the lexicon.
From a crawler
to an agent.
Conventional search, answer engines, generative engines, then systems that act on a record — each one further from a list of links.
- TGT-019 checks
SEO — Search engine optimisation
A crawler fetches the storefront, an index keeps one URL per product and files it into the result sets it can compete in. Everything the other three pillars measure starts from that fetch.
- TGT-025 checks
AEO — Answer engine optimisation
An engine answers the question in place and quotes where it can. The page has to hold a sentence, a list or an answer that survives being lifted out of it.
- TGT-035 checks
GEO — Generative engine optimisation
An engine composes an answer from what its crawlers were allowed to read. The store has to be readable, and its claims traceable to a named source.
- TGT-0410 checks
AIO — AI optimisation
A system reads the record to act on it: quote a price, pick a variant, start a checkout. Every fact has to parse, be stated once and agree with itself.
Four questions,
four different readers.
What each pillar asks, who asks it, where on a Shopify store the answer lives, and how the scorer weighs it today.
| SEOSearch engine optimisation | AEOAnswer engine optimisation | GEOGenerative engine optimisation | AIOAI optimisation | |
|---|---|---|---|---|
| The question it answers | Whether conventional search engines can crawl, parse, and rank the catalogue. | Whether the catalogue can be lifted as a direct answer to a shopper's question. | Whether generative engines will cite this store when composing a recommendation. | Whether an AI system can correctly parse and act on the record without guessing. Includes transactability as a named sub-score. |
| Who reads it | Search crawlers and indexes — Googlebot, Bingbot | Answer engines and the retrieval crawlers that feed them | Generative engines — their training and retrieval crawlers, and control tokens | Parsers, product feeds and agents that act on a record |
| Shopify surfaces | Theme head and H1, sitemap.xml, robots.txt.liquid, product_type, tags, handle, image alt | body_html, headings in the product template, FAQPage / HowTo / QAPage JSON-LD | robots.txt agent groups, /llms.txt, vendor, updated_at, units and entities in descriptions | Product JSON-LD offers, variant SKU and weight, option names, images, vendor |
| Main failure mode | Deduplicated, uncategorised or uncrawled — not in the running | Summarised instead of quoted — the attribution goes elsewhere | Unread or unattributable — missing from the composed answer | Contradictory or incomplete — described, but not actionable |
| Checks in the pillar layer | 9 · weight sum 100 | 5 · weight sum 86 | 5 · weight sum 82 | 10 · weight sum 108 |
| Heaviest check | Title length and uniqueness · sitemap.xml validity (14 each) | Description depth · FAQPage / HowTo schema (20 each) | llms.txt presence and quality (20) | Structured attributes (16) |
| Crawler penalty | 6 points per fully disallowed search crawler (Googlebot, Bingbot), capped at 30. Search-crawler blocks never touch the composite. | No crawler penalty applies to this pillar. | 6 points per fully disallowed AI agent, capped at 30, counted over all 10 agents the scan checks. The same penalty is subtracted from the composite. | No crawler penalty applies to this pillar. |
| Lexicon | SEO, defined | AEO, defined | GEO, defined | AIO, defined |
| Pillar page | /pillars/seo → | /pillars/aeo → | /pillars/geo → | /pillars/aio → |
Ordered, not nested.
Each fails on its own.
Each pillar assumes the one before it — but each is scored on its own inputs.
An answer engine can only quote a page it can reach; an agent can only act on a record it can read. But the pillars do not gate each other in the score, so any one can fail while the others pass:
- SEO passes, GEO fails. Googlebot is welcome; GPTBot and ClaudeBot are disallowed. The SEO pillar is untouched; GEO and the composite each lose 12 points.
- GEO passes, AIO fails. Every AI agent may read the store, and the product markup states two prices. Readable is not actionable.
- AIO passes, AEO fails. Offers, currency, availability and SKUs are complete; the descriptions are two lines of brand voice. An agent could buy the product; an answer engine has nothing to quote about it.
- AEO passes, SEO fails. The descriptions are model answers; every product type is blank and the theme omits the canonical link.
Some inputs are read by more than one pillar, for different questions. The description feeds AEO’s depth and opening checks and GEO’s units and entities; the vendor field feeds GEO’s attribution and AIO’s consistency.
Four pillar scores.
One composite. Not a sum.
The composite and the pillar layer are computed from the same audit and reported side by side. Neither is derived from the other.
- Composite · rubric v1.0
- 9 catalogue checks with weights totalling 95 — published as they are, not rescaled — blended 80% with a structured-data sub-score at 20%, then 6 points off per disallowed AI agent, capped at 30. Rounded; never below 1.
- Pillar layer · v1.0
- 29 checks in four pillars, each normalised to 100 over the checks that fed it, with its own weights (totals: SEO 100 · AEO 86 · GEO 82 · AIO 108). Pillars do not sum to the composite, and this layer cannot change it. An input the scan could not read is unscored — excluded, not failed.
- Shared inputs
- 9 of the composite’s 9 catalogue checks also feed a pillar check, often at a different weight. The rubric shows each pair.
- Canonical Rank Score
- In development One versioned scoring contract for all four pillars. Until it ships, both layers above are reported, each labelled as what it is.
What AIO
means here.
The one acronym of the four that the industry uses in three incompatible ways.
Parse and act.
AIO here means machine readability — whether the record parses and can be acted on. It is not the umbrella definition that treats AIO as all four disciplines at once, and it is not the off-site trust definition.
What no pillar
measures.
Stated so none of the four scores reads as more than it is.
- No engine is queried. The scan measures what the storefront serves to crawlers — structure, completeness, access — not what any engine answers on a given day.
- No ranking, citation or traffic is promised. Where an engine places or quotes a store is that engine’s decision.
- Nothing is written to a store today. Fixes are Planned; the scan reads and scores.
Five field guides.
Mapped to the pillars.
Cornerstone guides for Shopify, each tied to the pillars it serves. The lexicon, the crawler index, the rubric and the tutorial are linked below.
Structured data for Shopify
Product, Offer and ProductGroup markup on a Shopify storefront, and how theme and app markup collide.
Read →Answer-engine readiness
Descriptions, headings and question-and-answer content an engine can lift.
Read →llms.txt for Shopify
What the proposal is, what it is not, and how a store would publish one.
Read →Product data for AI agents
SKUs, identifiers, variants, prices and availability as an agent reads them.
Read →Agentic commerce for Shopify
What the platforms say about agents buying from Shopify stores, dated and sourced.
Read →Terms, defined
66 search and AI-commerce terms, the four pillars among them.
Open →Who reads a store, by category, is in the crawler index. How each score is built is on the rubric. The merchant learning path is the tutorial; every guide is listed in the field guides.
Four scores.
One free scan.
The public scan reads robots.txt, the product listing, up to three product pages, sitemap.xml and llms.txt, and scores all four pillars.
