You search for a product your store sells. Perplexity recommends three alternatives, including a competitor with a less detailed product page. Your brand is nowhere in the answer.
The tempting response is to add more keywords. Before you do, work out what is actually missing: access to your product information, evidence that the product fits the question, or a reason to recommend your store over another seller.
Those are different problems. They need different fixes.
This guide covers the Perplexity merchant program, how to prepare a useful product feed, and what to check when your products are missing from shopping recommendations. It also explains how to separate brand mentions, citations, and product recommendations when measuring progress.
Start here: Make your product information accessible, specific, and consistent. Confirm the catalog-sharing options available to your business. Then test realistic shopping questions and inspect the answers and sources. Treat submission as a distribution step—not proof that a product will be recommended.
What is the Perplexity merchant program?
Perplexity introduced its merchant program in November 2024 to let retailers share product information for its shopping experience. Its launch announcement described catalog access, checkout integrations, and merchant insights. Participation was presented as free and separate from its advertising products.
The useful distinction is between being available for discovery and being chosen for a particular shopper. A catalog can provide product information; it cannot establish that every item is the right answer to every question.
For planning, separate three jobs: providing accurate product data, earning a recommendation for an appropriate use case, and making the eventual purchase straightforward. Do not use a checkout integration as a substitute for the first two.
How to get started with Perplexity Shopping
1. Confirm the current merchant application or platform route
The official launch article links to a merchant interest form. Check that it is still accepting relevant applications, or confirm the available route with Perplexity or your commerce-platform partner.
An interest form is not the same as an approved integration. Before committing engineering time, ask which markets and categories are supported, how catalog data should be delivered, and whether checkout requires separate onboarding.
Do not plan around a universal five-minute setup or a guaranteed approval date. Get the requirements for your specific account in writing.
2. Choose the catalog that will be your source of truth
Decide whether your storefront, product information system, or feed-management service owns each field. Price, inventory, and descriptive content may have different owners, but changes should not create conflicting versions of the same item.
Ask your integration provider a concrete question: “When this variant changes price or sells out, where does that update go, and how do we know it was accepted?”
A successful file transfer is not evidence that every product record was processed correctly. Build acceptance checks into the workflow.
3. Validate a small, representative product set
Start with products that expose likely problems: an item with several sizes, a sale item, an out-of-stock variant, and a product with a compatibility requirement. Check the landing page, image, identifiers, price, currency, and availability for each one.
Resolve those inconsistencies before expanding. Ten well-understood records make a better test than a full catalog whose errors nobody has reviewed.
4. Check purchasing separately from discovery
Perplexity's November 2025 shopping update described a free shopping experience for US users and checkout through PayPal, with retailers remaining the merchant of record in that arrangement.
That announcement is not a promise that every product or merchant has the same checkout options today. Verify the experience available to your store, including payments, shipping, returns, and customer support. A supported checkout route does not establish a guaranteed ranking advantage.
Product feeds for Perplexity Shopping: what to prepare
A product feed is a structured version of your catalog. Its practical job is to make each item identifiable and its offer understandable.
The checklist below is a preparation guide, not an official Perplexity feed specification. Confirm the accepted field names, formats, delivery method, and update schedule with your integration provider. Do not assume a Google Shopping feed can be submitted unchanged.
| Product information | What to check before sharing it |
|---|---|
| Identity | Use a stable item ID, correct brand and model, and the right variant. |
| Product title | Identify what the item is; add the attributes that distinguish it. |
| Description | Explain use cases, specifications, limitations, and compatibility. |
| Landing page | Link to the correct product or variant, not a generic collection. |
| Images | Show the actual item and variant with accessible, relevant images. |
| Price and currency | Match the offer on the landing page, including active promotions. |
| Availability | Distinguish purchasable, unavailable, and preorder items accurately. |
| Product identifiers | Include valid manufacturer-assigned identifiers where applicable. |
| Variants | Keep size, color, capacity, and other options distinct. |
| Purchase conditions | Make shipping, returns, and relevant restrictions easy to find. |
These are familiar catalog-quality checks. Google's product data specification is a useful reference for maintaining a Google feed, but it is not evidence of Perplexity's submission requirements.
Do all products need a GTIN?
Provide a correct Global Trade Item Number when one has been assigned. Do not invent a barcode to fill a blank. Google's identifier guidance for customized products explicitly recognizes that some custom or one-of-a-kind products lack standard identifiers. Other integrations have their own rules.
Avoid the blanket claim that products without GTINs are invisible to Perplexity. Check the requirements that apply to your catalog and submission route.
Keep feeds and product pages consistent
Check the same facts in the feed, visible product page, and structured data. A shopper should not see one price in a recommendation and another on arrival because an old feed or default variant supplied the wrong offer.
Google documents Product structured data and Merchant Center feeds as complementary ways to provide product information. That is a Google-specific implementation benefit, not proof of a particular Perplexity ranking factor. Our ecommerce schema guide covers the on-page implementation separately.
How to write product information that answers shopping questions
Replace vague praise with details a buyer can use. A useful description should explain who an item suits, what it does, and where its limits are.
Consider this hypothetical example:
Vague: “Premium everyday backpack. Stylish, versatile, and built for adventure.”
Specific: “A 20-liter commuter backpack with a padded sleeve for laptops up to 15 inches. Its water-resistant outer fabric suits light rain, but it is not designed for prolonged exposure. Two external pockets hold a bottle and umbrella.”
The second version gives a reader something to compare: capacity, fit, weather protection, and storage. Use those details only when they are true and supported by the product specifications.
The same principle applies to titles. “Commuter backpack, 20L, 15-inch laptop sleeve” communicates more than “The Everyday, Style 18383.” You do not need to squeeze every attribute into the title; put the rest in clearly labeled specifications.
Use support tickets and pre-purchase questions to find missing details. Answer “Will it fit?”, “What is included?”, and “What is it not suitable for?” on the relevant product page. Do not invent an intended_purpose feed field unless your actual integration supports it.
Why aren't my products showing in Perplexity recommendations?
A missing recommendation is an observation, not a diagnosis. Start by checking which part of the journey is failing.
| What you observe | A useful next check |
|---|---|
| The exact product name does not surface useful information | Inspect page access, indexing signals, catalog delivery, and product identity. |
| The item appears for its name but not a use-case question | Check whether the page actually supports that use case with specific evidence. |
| Your product appears, but another store is the seller | Compare the sellers' availability, offer accuracy, and purchase conditions. |
| The answer uses the wrong price or variant | Trace the cited page or listing and compare update times across sources. |
| A competitor appears repeatedly | Read the cited pages and identify relevant evidence your own content lacks. |
| Your brand is named, but your website is not cited | Check whether the source is a retailer, reviewer, or another publisher. |
These are diagnostic checks, not a disclosed list of Perplexity ranking factors. Several problems can exist at once.
Check crawler access before rewriting everything
Perplexity's crawler documentation identifies PerplexityBot as its search crawler. It recommends allowing the crawler and its published IP ranges when a site wants to appear in search results.
Ask your technical team to check both robots rules and the web application firewall. A page that loads for a person may still present a challenge or block to a crawler. Verify legitimate bot traffic using the provider's published guidance rather than trusting a user-agent name alone.
Compare the same question, not two different shopping situations
“Best backpack” is not the same test as “20-liter work backpack for a 15-inch laptop under $100.” Save the exact prompt, market, interface, date, answer, and cited sources.
Inspect a few repeated runs before drawing conclusions. For a competitor comparison, write down the specific advantage you can observe—such as a clear compatibility table or a current retailer listing—not a guess about the engine's hidden weighting.
Fix the most concrete gap first. Then rerun the same test. This will not prove causation, but it creates a more useful record than changing several pages and assuming the next answer validates everything.
How to earn useful brand mentions and citations
For your reporting, treat a brand mention as the appearance of your brand name in the answer. Count a citation only when the answer includes a source link, and record its destination. Count a product recommendation when a specific item is proposed to the shopper.
These outcomes are related, but they are not interchangeable. A reviewer can be cited while your product is recommended. Your buying guide can be cited without your product being selected.
To make your content worth referencing, publish something that resolves a real decision: a measured fit guide, a compatibility table, a clearly explained test, or a comparison that acknowledges trade-offs. Include the method, date, and limits behind any original result. Do not present an illustrative example as customer research.
For external coverage, start with the sources appearing in relevant answers. Offer reviewers and publishers accurate specifications, useful expert input, or access for genuine testing. Let them make independent judgments. Repeated promotional mentions and paid praise are not substitutes for evidence.
What changes across ChatGPT, Perplexity, and Google AI search?
Reuse accurate product facts across channels, but verify each channel's access and distribution requirements independently.
| Channel | What to verify |
|---|---|
| Perplexity | Search-crawler access, relevant product information, and the catalog or checkout route available to your business. |
| ChatGPT Search | Access for OAI-SearchBot, plus how your content appears in search-enabled answers. |
| ChatGPT shopping results | Product metadata and the applicable catalog or direct-feed route. |
| Google AI Overviews and AI Mode | Indexing, snippet eligibility, accessible content, and established SEO fundamentals. |
OpenAI distinguishes its search crawler from its training crawler: allowing OAI-SearchBot does not require allowing GPTBot for training. Its shopping documentation explains that product selection considers relevance, context, structured metadata, and other content. Direct product feeds have a separate documented specification.
Google's guidance for AI search features says no special AI schema or additional technical requirements are needed. Pages must be indexed and eligible to appear with a snippet. Meeting those conditions does not guarantee selection.
Keep Google Search AI features separate from the standalone Gemini experience in your reporting. A result observed in one should not be recorded as a result in the other.
How to track product recommendations and citation visibility
Build a small set of questions that represent buying decisions, then keep it stable long enough to compare results.
For a backpack brand, a useful starting set might include “Which work backpacks fit a 15-inch laptop?”, “What is a good commuter backpack under $100?”, and “Which compact backpacks are suitable for light rain?” Those are examples, not validated search-volume estimates.
Use separate groups for unbranded discovery, product comparisons, and named-product questions. Branded prompts test whether information can be retrieved; they should not inflate your measure of unprompted discovery.
For each engine, track the share of completed test answers that mention your brand, link to your domain, and recommend your products. Record the raw counts as well as the percentages. Log failed runs separately, and keep the prompts, geography, interface, and denominator consistent.
Alongside those observations, measure qualified website visits and conversions where referral data is available. A Google impression does not demonstrate a ChatGPT citation, and a tracked mention is not an order. Treat monitoring results as a sample of answers—not a census of what every shopper sees.
Where Alhena fits
Alhena AI Visibility helps ecommerce teams examine how their brands and products appear in AI answers. Its documented capabilities include product-level tracking, analysis of how recommendations are displayed, and content-improvement suggestions. The Citations and Prompts views help teams inspect sources and individual tracked questions.
Use those observations to prioritize work: investigate a missing product, correct an inaccurate description, or examine a source that repeatedly appears for competitors. Measurement can guide the next test; it does not reveal or control an engine's private ranking system.
See what AI recommends in your category. Explore Alhena AI Visibility to assess product appearances, cited sources, and the gaps worth investigating.
Frequently asked questions
Is joining the Perplexity merchant program required to be mentioned?
The original announcement described both platform integrations and a merchant program; it did not establish that every recommendation requires a separate application. A mention in a web answer, catalog inclusion, and integrated checkout are different forms of participation. Confirm the route relevant to your goal.
Is the Perplexity merchant program free?
Perplexity described participation as free at launch. Do not interpret that as a guarantee of zero total selling costs. Confirm current program terms and any payment-processing, integration, shipping, or fulfillment costs separately.
Can I use my existing Google Shopping feed?
Use it as a starting point for catalog preparation. Confirm the receiving integration's accepted format, required fields, delivery method, and validation rules before submitting it. This guide does not assume that Google CSV files, SFTP delivery, or a specific API are universally supported by Perplexity.
How should Shopify merchants approach Perplexity Shopping?
Check which catalog-sharing arrangements apply to your store, markets, and products. Then review product details, identifiers, variants, availability, and landing pages. Do not treat Shopify catalog access as proof that every item has been included or will be recommended for a particular question.
Who is the merchant of record in Perplexity's PayPal shopping flow?
Perplexity's November 2025 announcement says participating retailers remain the merchant of record in its PayPal arrangement. Do not generalize that statement to every historical Buy with Pro flow or another checkout partner. Check the terms of the integration you use.
Does a complete product feed guarantee more recommendations?
No. It supplies information; it does not prove that an item will be selected for a given shopper. Test relevant questions, inspect the results, and work on observed gaps rather than treating feed completion as a ranking guarantee.
Start with the gap you can verify
Choose a handful of commercially important products. Check their information, test realistic questions, and save the answers and sources. Fix the clearest problem, then repeat the test.
The goal is not to add “AI-ready” to every page. It is to make your products easier to understand, your claims easier to verify, and your progress easier to measure.
