E-commerce

E-commerce overview

Track whether AI assistants list your products in "what should I buy" answers, and whether they could recommend and sell them at all.

A brand named in a paragraph is not the same as a product offered as a clickable option. The E-commerce screen measures the second: whether AI answers to shopping questions list your products, at what rank, through which sellers and on which engines. It also tells you whether an assistant could recommend and sell your products — the feed, page markup, crawler access and checkout path it needs.

Use it if you sell physical products, on your own site, on marketplaces or both.

Where to find it

Open Analytics › E-commerce in the sidebar. The screen has four tabs:

TabWhat it answers
Product visibilityWhich shopping questions are tracked, and how often each product is listed in the answers.
Ad loadWho advertises on your category questions, and on what share of them.
ProductsThe catalogue being measured.
ReadinessIf an assistant wanted to recommend your product, could it? Includes the AI shopping feed and the agent test.

The buttons at the top right add products: Import from store, Import from file and Add product.

Step 1: Add your products

Everything on this screen starts from your product list. There are three ways to fill it:

  1. Import from store — pulls the catalogue from a connected Shopify, ikas or WooCommerce store, using the read-only connection you set up under Integrations. The import runs in the background; the products appear shortly after you press the button. If no store is connected, the screen tells you so.
  2. Import from file — for any cart, marketplace or spreadsheet that can export a product file. It accepts CSV, TSV or JSONL with Google product feed columns (Turkish column headers are recognised too). You can download a sample file from the dialog.
  3. Add product — type a product by hand. The name is required; an SKU or model code and a product URL make matching in answers stronger.

How the file import behaves:

  • Press Preview first. It shows how many rows are new, will be updated or are unchanged, which rows have something missing or wrong, and which columns were not recognised.
  • Rows without a product name are skipped.
  • Products that are not in the file are not deleted.
  • At most 5,000 rows at a time; split a larger catalogue. A file that is too large is refused whole, so you never end up with half a catalogue.

Removing a product on the Products tab also stops measuring it.

Step 2: Track shopping questions

A product can only be listed in an answer if some question asks for it. The Product questions card at the top of the Product visibility tab writes those questions for you:

  1. Press Generate questions.
  2. Rixe groups your catalogue into buying groups (up to 12) and writes shopping questions for each one in your market's language — "best …", budget, comparison, gift and use-case questions.
  3. Every group also gets one question that names no brand. Those carry the real measurement: they show whether you are mentioned at all, and who comes up instead.

New questions are added to your tracked prompts, so they count against your plan's prompt limit. If the limit is reached, the remaining questions are left out with a message. Their answers appear after the next daily measurement. Read Prompts for how tracked questions work in general.

Step 3: Read product visibility

The Product visibility table covers shopping answers from the last 30 days:

ColumnMeaning
ProductYour product, with the favicon of its page.
ListingThe share of shopping answers that listed this product, with the raw count (for example 12/40).
Average rankIts average position in the answers that listed it (#1 is first).
SellerWhere the answers sent people to buy it: your site or marketplaces.
EngineThe engines that listed it.

A product with too few shopping answers shows not enough data instead of a rate. A rate built from a handful of answers would move with chance, not with your work.

Marketplace rankings

In some markets a footnote names marketplaces whose own search rankings are not measured, because those marketplaces offer no third-party search API. Which marketplaces appear depends on your market.

Ad load

The Ad load tab shows the share of your category questions that trigger ads, the ad load, and the advertisers. An advertiser is a competitor whether or not it is on your competitor list.

Only surfaces with a results page (Google AI Overviews and SERP measurement) carry ads. Engines measured as plain language models show none, so ad load cannot be measured there.

Readiness

The Readiness tab is the audit behind recommendations: product feed completeness, crawler access, page markup, reviews, which AI shopping surfaces you can enter from your country, the AI shopping feed for ChatGPT, and the agent checkout test. It is covered in two pages:

  • AI shopping feed — your catalogue in ChatGPT's product feed format, and how to submit it.
  • Agent readiness — the readiness score, the fix list and the agent test.

Common questions

My products are mentioned in answers, but the table is empty. The table counts shopping answers to tracked questions. Generate product questions (Step 2), then wait for the next measurement.

Why does a marketplace show as the seller instead of my site? Assistants link to wherever they found the product. Many assistants cannot read marketplace pages at all, so a product page on your own site matters — see Agent readiness.

Does importing a file delete products I removed from it? No. Remove products by hand on the Products tab.

Still stuck? Write to [email protected] and a person will answer.