49 questions Shopify merchants, agencies and affiliates ask before they trust a score, each answered in the tense that is true today and linked to the page that owns the detail. Where something is not built, the answer says so.
What Rank Sniper is, which part of it you can use now, and what has not happened yet.
A Shopify-only tool that measures how well a store’s catalogue can be read by search engines, answer engines, generative engines and AI agents, reported as four pillar scores (SEO, AEO, GEO, AIO) beside one composite score. It has two parts: a free public scan that runs today, and an embedded Shopify app that is in development. Fixes, verification and monitoring are designed to live in the app; none of them runs yet. How the workflow fits together.
The public scan (Available): a composite score and four pillar scores for any Shopify storefront, with no install and no account. Its robots.txt read of named AI crawlers is Available. The full report by emailed sign-in link is in Beta, and support tickets with Rank Sniper staff are Available. Everything that connects to a store, writes a fix or watches for drift ships with the Shopify app and does not run today. The changelog records what has actually landed.
No. No fix has been applied to any store. The app that would write fixes is not on the Shopify App Store and is installed on no store, and every fix type is still on the roadmap: automatic markup (Planned), automatic metadata (Planned) and assisted batch rewrites (Planned). What exists today is measurement: the public scan reads a storefront and scores it. If any page implies Rank Sniper has repaired a store, it is wrong, and we want to hear about it.
Not yet. Installing from the Shopify App Store is In development, so there is nothing to install and no store has granted the app access. Tier prices are ruled and listed on the pricing page, but nothing can be bought until the app lists, because subscriptions will be billed through Shopify (Planned). Until then, the free scan is the part of Rank Sniper you can use.
What the scan reads from a Shopify store, and how the app is designed to touch one.
No. The scan reads what Shopify storefronts serve (/products.json for the catalogue, product pages, /sitemap.xml and /robots.txt) and the app is being built against the Shopify Admin API. A store on another platform, or a headless storefront that does not serve /products.json, cannot be scored, and the scan says so rather than guessing. How the scanner is built.
From /products.json it reads each product’s description (body_html), product_type, vendor, tags, images, options and variants, including each variant’s SKU and weight in grams, plus updated_at for freshness. The composite’s 9 catalogue checks test those fields across the products it samples. Shopify’s developer documentation read on 2026-09-11 does not document /products.json as a supported API, so treat it as storefront behaviour, not a contract. products.json, defined.
Shopify generates a default robots.txt. To customise it, add a robots.txt.liquid template to your theme; it renders /robots.txt and exposes Liquid objects for its groups, rules, user-agents and sitemap. Shopify strongly recommends building on those objects rather than replacing the file, because its default rules are updated regularly. Add your own groups for the agents you want to name. The crawler index sets out which token means what, and the lexicon defines the template.
It is not designed to. Structured-data fixes are designed to ship through a theme app block rather than edits to your theme files (Planned). Changes outside the app’s write scope (theme templates, sitemap configuration and root-domain files such as llms.txt) are designed as guided fixes: exact steps you apply, which the next scan checks. Theme app blocks, defined, and the designed write contract.
None has been requested from anyone, because the app is not installable, and the final access-scope list is not published yet. As designed, writes pass through an allowlist: Rank Sniper’s own metafield namespace for automatic output, and the specific native fields an assisted fix names once you approve it (Planned). A mutation that is not on the list is refused before it is sent. Access scopes, defined.
The four disciplines as Rank Sniper defines them. The definitions below are published verbatim; the answers explain how they are scored.
Whether conventional search engines can crawl, parse, and rank the catalogue.
Read the pillar →Whether the catalogue can be lifted as a direct answer to a shopper's question.
Read the pillar →Whether generative engines will cite this store when composing a recommendation.
Read the pillar →Whether an AI system can correctly parse and act on the record without guessing. Includes transactability as a named sub-score.
Read the pillar →Because a store can pass one discipline and fail the next. A catalogue can be well indexed by Google (SEO), give an answer engine nothing it can quote (AEO), block the crawlers generative engines send (GEO), or carry prices an agent cannot parse (AIO). Rank Sniper keeps them apart, always in the order SEO → AEO → GEO → AIO, from conventional search to agents that act. The definitions above are Rank Sniper’s own; each pillar has its own page.
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. In the scan, AIO covers structured attributes, SKU coverage, variant-level offers, price with priceCurrency, availability and conflicting markup, and it carries a named transactability sub-score. The AIO pillar.
No. The composite is rubric v1.0: 9 catalogue checks whose weights total 95, blended 80% with a structured-data sub-score at 20%, minus the crawler penalty, rounded, with a floor of 1. The pillar scores are a separate layer (v1.0) with their own weights, each normalised to 100 over the checks that could be read. Neither layer changes the other. One canonical four-pillar Rank Score is In development. The rubric.
No. Rank Sniper does not score backlinks, domain authority, or any link graph. That data comes from third-party indexes Rank Sniper would have to buy and resell, and it says nothing about whether an agent can read a product record, which is what every check measures. Rank Sniper also never promises rankings, citations, traffic or revenue: a score measures what your storefront serves, and where an engine places you is that engine’s decision.
Exactly what the public scan fetches, what it obeys, and how robots.txt enters the score.
In this order: /robots.txt first, which it obeys; then /products.json; up to three product pages as raw HTML; /sitemap.xml; and /llms.txt. That is the whole list. It never logs in, submits a form, opens checkout or account pages, writes anything or runs JavaScript, so markup a script injects later is not seen. It identifies itself as RankSniperBot; its identity page shows the exact User-Agent.
The scan does not run. RankSniperBot reads robots.txt before anything else; if the catalogue endpoint is disallowed, nothing is fetched and the person who asked is told the store declined. Narrower rules are honoured path by path: disallowing /products/ skips the product-page sample and leaves those checks unscored. Two lines refuse it outright: User-agent: RankSniperBot and Disallow: /. How it matches rules.
10, by name: GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, anthropic-ai, PerplexityBot, Google-Extended, Applebot-Extended, Bytespider, CCBot. Each is checked against the robots.txt group that applies to it. Googlebot and Bingbot are read too, for the SEO pillar only. Agents outside the list are not checked, including Claude-SearchBot, Claude-User, Perplexity-User; adding one changes scores, so the list moves only with a rubric version. The crawler index covers 33 agents, with who runs each and what a block changes.
Each listed AI agent your robots.txt fully disallows subtracts 6 points from the composite and from the GEO pillar, capped at 30. It is a capped penalty, not an exclusion from the score: catalogue quality still counts. Only a root Disallow: / without an overriding Allow: / counts, and a wildcard block applies to every listed agent the file does not name, so User-agent: * with Disallow: / reaches the cap by itself. Control tokens such as Google-Extended count the same as crawlers today.
Because an input could not be read. If a product page times out, is refused, or is disallowed to RankSniperBot, the checks that depend on it are excluded from that pillar rather than scored as failures, and the pillar normalises over the checks that remain. A missing file is different: a 404 on sitemap.xml or llms.txt is a finding, because absence is what those checks measure. An unreadable robots.txt is treated as no restriction, the standard’s default. Unscored checks, defined.
How the Shopify app is designed to act on a scan. Every fix below is on the roadmap, not in a store.
As designed, each pillar check has one of three fix classes. Of 29 pillar checks, 4 are automatic: structured data published in Rank Sniper’s own layer, with no native field changed. 20 are assisted: native Shopify fields such as title, description, tags or SKU, drafted and written only when you approve them. 5 are guided: theme, sitemap or root-domain changes you make from exact steps. All of it is Planned; nothing is written to any store today. Fix classes, defined.
No. A SKU, a GTIN, a shipping weight, a material or a brand is something only the merchant holds, so the design asks you for it rather than guessing. An agent quotes whatever a store publishes, and a plausible invented value is worse than a blank one: it is wrong in a way no check can catch. Only fixes built from facts your catalogue already holds, such as structured data, are designed to be automatic. GTIN, defined.
Start with crawler access, then the weakest pillar. Each blocked listed agent costs 6 composite points, up to 30, and removing an accidental block is a robots.txt edit. If the block was deliberate, keep it and read the score with that in mind. The public result flags the weakest pillar; the full report by emailed link (Beta) lists that pillar’s failing checks, heaviest-weighted first. The workflow.
What the app is designed to change, what it will not, and what protects a store when it does. None of it has run on a store.
Only if you approve it. The app is installed on no store, so nothing has been written anywhere. As designed, automatic fixes publish through Rank Sniper’s own metafield namespace and structured data and never touch your copy. Assisted fixes do change native fields (titles, descriptions, tags, alt text) but only as drafts staged for you; nothing is written until you apply it. Assisted batch rewrites are Planned. The designed write contract.
That is the design, and it is not built. Versioned rollback on every write: Planned. The plan stores the prior value of every field before it is written, with per-product restore, and rollback is designed to ship before the first automatic fix does. Until it exists and has been exercised on real products, Rank Sniper does not describe any change as reversible in practice. Rollback, defined.
Both are planned. Post-fix verification — what agents actually receive: Planned, designed to re-fetch the delivered HTML after a fix, without JavaScript and as each declared AI user agent. Recurring re-scans, drift detection and alerts: Planned. Today nothing runs on a schedule: a re-scan happens when you ask for one, so run the free scan again after a theme update or a new app install. Verification, defined.
Nobody has installed it, so today there is nothing to remove. As designed, automatic output lives in Rank Sniper’s own metafield namespace and theme app block, separate from your native fields, while assisted changes you approved are edits to your own fields and remain yours. That is a design statement, not a tested one: uninstall behaviour can be verified only once the app exists. The changelog records what has landed.
Who charges for Rank Sniper, what can be bought today, and where the Founders programme stands.
Shopify is the only billing authority for Shopify merchants. Rank Sniper subscriptions and add-ons will be billed through Shopify when the app is available; nothing is charged on ranksniperhq.com. Shopify billing is Planned, and there is no checkout on this site. The pricing page lists the ruled tier prices, banded by catalogue size, with all four pillars in every tier.
No. Nothing is sold on ranksniperhq.com. What you can use today is the free scan, which needs no account and no payment. Subscriptions and add-ons will be charged through Shopify inside the app, and the app is not on the Shopify App Store yet (In development). A page or message asking you to pay Rank Sniper directly is not a Rank Sniper channel; tell us about it.
A founding-member programme for stores that back Rank Sniper before it lists, capped at 50 stores: the size of the programme, not a countdown. Founder terms are being reconciled. No Founder purchase is available, and a request is recorded, never charged. Its commercial terms stay frozen until they are reconciled, the entitlement is defined and a Shopify-compliant billing path exists. Where it stands.
No. Rank Sniper offers no refund guarantee today, and it does not promise rankings, citations, traffic or revenue, because none of those is Rank Sniper’s to give. What it does commit to is narrower: every score is computed from a versioned rubric you can read, and the availability statements on these pages read from one release table. Audit the rubric yourself.
For agencies, freelancers and in-house teams operating Shopify stores they do not own.
Yes, through the free scan (Available): point it at any Shopify storefront, whether a client’s, a prospect’s or a competitor’s, and get a composite and four pillar scores without an install or a client login. The multi-store portfolio view is In development; team permissions (Planned), client reporting (Planned) and white-label reports (Planned) are not built. The agency lane.
No. Agency status does not create referral commissions. An agency that also wants to promote Rank Sniper may separately apply to the affiliate programme; applications are reviewed, and approval is not automatic. Affiliate attribution and commissions are governed independently from agency client access, and neither relationship grants the other.
The merchant. As designed, an agency works through an explicit client relationship the merchant has agreed to, and a store access grant the merchant gives per store and can revoke; revoke the grant and the agency’s access ends. The store, its data and its subscription stay with the merchant. These access objects are Planned. The agency access model.
For creators, publishers and advertisers who refer Shopify merchants. A separate programme from the agency lane.
Rank Sniper enrols affiliates directly; there is no self-serve sign-up. An affiliate earns 20% of revenue Rank Sniper has actually collected from a store they referred, for 12 months from that store’s first payment, with a 60 days last-click cookie. Referral links are Private Beta; attribution through to an install is Planned and payouts are Planned. The full terms.
No. Referring a merchant grants no access to that merchant’s store, data or dashboard. The affiliate relationship is “I referred this merchant”; operating a client’s store is the agency lane, a separate programme with a separate enrolment. An affiliate enrolment does not include agency functionality, and agency status does not create referral commission. What an affiliate does not get.
Not in practice. Commission accrues only on revenue Rank Sniper has collected, and nothing can be bought until Shopify billing is live (Planned). Payouts are Planned: the payout rail is not built, and the Affiliate Command portal is Planned. Referral links already record visits. Crediting runs from 2026-08-05. The programme page.
What the scan reads, what is stored, what is public, and what is not yet documented. Where a policy does not exist yet, the answer says so.
Only public URLs your storefront already serves to any visitor: /robots.txt, /products.json, up to three product pages, /sitemap.xml and /llms.txt. It reads product records and page markup. It does not log in, submit forms, open checkout or account pages, or collect personal data, and it does not follow links to discover more of your site. The crawler’s own page.
Yes, it is stored. A completed scan is saved so its result page at /scan/your-store loads without re-scanning, and it is re-used for 24 hours; anyone who enters the same storefront address in that window sees the stored result. Result pages carry a noindex directive, so they can be shared by link but are not offered to search engines for indexing. Published statistics are aggregate: a SKU count rounded to a band and a category, never the hostname.
It is stored with the storefront you scanned and used to send a sign-in link that opens the full report. The emailed report is in Beta: end-to-end link redemption is still being verified. How long the address is kept, and which providers process it, are not yet set out in a final privacy policy. That gap is stated rather than papered over; the next answer explains why.
Not a final one. The privacy page is a clearly labelled placeholder: it states what is true today and says it has not been reviewed by a lawyer. Retention periods, sub-processors, legal bases and a data-subject request route are not yet documented, because they depend on hosting, mail and Shopify app decisions that are still open. Until that policy exists, ask about anything specific through the contact form.
No. The free scan reads only public pages and never authenticates. Reading a catalogue through the Shopify Admin API needs the app installed and your permission, a connected scan, which is Planned. No store has installed the app, so Rank Sniper holds Admin API access to no store. The Admin API, defined.
What Rank Sniper can and cannot tell you about AI answers, and what the vendors themselves say.
No. Rank Sniper does not query AI engines and cannot see whether any of them cites, recommends or ranks your store. It checks structural readiness: whether the crawlers those engines send are allowed in, whether your product records are complete, and whether they parse. Those are conditions an engine needs; whether it then uses your store is the engine’s decision, and nobody outside it can promise the result. Citation, defined.
No score can promise that. Google says there are no additional requirements to appear in AI Overviews or AI Mode, and no special schema.org structured data to add for them; Googlebot’s robots.txt rules are the control for how a site is crawled for them. What Rank Sniper measures (access, complete records, valid markup) are fundamentals, not a shortcut. Google on AI features.
Unproven. llms.txt is a community proposal, first published by Jeremy Howard in 2024, not a standard. Google states that Google Search ignores it, so it neither helps nor harms visibility there, and none of the major AI crawler pages read for the crawler index mentions consuming it. Rank Sniper’s GEO pillar reads whether you publish one, as a statement of your terms, not as a claim that any engine uses it. The llms.txt guide for Shopify.
No. Google limited FAQ rich results to well-known, authoritative government and health sites in 2023, then stopped showing them: its Search Central updates log records that the feature no longer appears in Google Search from May 7, 2026. FAQPage remains valid schema.org markup that other machine readers can parse, and Rank Sniper’s AEO pillar reads it as one question-and-answer signal among several. Nobody can promise a rich result from it. Google’s updates log.
Decide per class, not per brand. Blocking a training crawler such as GPTBot or ClaudeBot is an opt-out from model training; by OpenAI’s and Anthropic’s own accounts, their search agents are separate. Blocking an AI search crawler such as OAI-SearchBot can remove you from that product’s answers. Several vendors say their user-initiated fetchers may ignore robots.txt. Rank Sniper’s score penalises every listed agent the same way today, so a deliberate choice can cost points. Decide on the crawler index.
The terms the other answers lean on. Each links its full lexicon entry.
A training crawler collects content that may be used to train models. An AI search crawler builds the index an AI product retrieves from when it answers. A user-initiated fetcher reads one page because a person asked. A control token, such as Google-Extended or Applebot-Extended, is not a crawler: it is a name in robots.txt that decides how content another crawler fetched may be used. Each is a different decision.
No. RFC 9309, the robots.txt standard, says the protocol “is not a substitute for valid content security measures,” and Google says robots.txt is not a mechanism for keeping a page out of Google; it points to noindex or password protection instead. robots.txt asks crawlers not to fetch. It stops nobody, and some user-initiated fetchers say they may not follow it. robots.txt, defined.
Not by Google. Google says structured data is not required for generative AI search and no special schema.org markup is needed. It still decides rich-result eligibility: for merchant listings Google requires a Product with name, image and offers, and an Offer with price and priceCurrency. Rank Sniper’s AIO pillar treats conflicting markup as worse than none, because a parser cannot choose between two prices. Structured data for Shopify.
GEO, generative engine optimisation, was named in a 2023 research paper, “GEO: Generative Engine Optimization” by Pranjal Aggarwal and co-authors, accepted to KDD 2024. AEO, answer engine optimisation, has no single canonical origin traceable to a primary source; industry accounts conflict. Rank Sniper publishes its own definitions of both, verbatim, because the field has no consensus meaning. The GEO paper.
The fastest way to make any of these answers specific is to scan your own store. If a question is not here, a person reads every message.
