---
title: "Product matching"
description: "Product matching data: get GTINs, model numbers, titles, pack sizes and images from Amazon, Walmart and other retailers to match one item across stores."
canonical: https://www.zenrows.com/use-cases/product-matching
updated: 2026-10-19
---

# Product matching

Product matching is identifying the same product across retailers, even when titles, codes and photos differ, so prices and assortment can be compared. Zenrows collects the evidence a matcher needs (GTINs, model numbers, brand, pack size, specs, images) from protected retailer pages and returns it as structured fields per candidate.

## Sample output: one product, three retailers

```json
{
  "anchor": {
    "sku": "INT-55021",
    "brand": "Example Home",
    "model": "EH-K17",
    "gtin": "5012345678900",
    "title": "Example Home 1.7L Kettle, Brushed Steel"
  },
  "candidates": [
    {
      "retailer": "retailer-a.example",
      "url": "https://retailer-a.example/p/eh-k17",
      "title": "Example Home Kettle 1.7 Litre Steel",
      "gtin": "5012345678900",
      "match": "exact_gtin"
    },
    {
      "retailer": "retailer-b.example",
      "url": "https://retailer-b.example/item/88123",
      "title": "EH-K17 Electric Kettle 1.7L (Brushed)",
      "gtin": null,
      "model": "EH-K17",
      "capacity_l": 1.7,
      "match": "model_and_spec"
    },
    {
      "retailer": "retailer-c.example",
      "url": "https://retailer-c.example/dp/X7731",
      "title": "Example Home Kettle Twin Pack",
      "pack_size": 2,
      "match": "rejected_pack_size"
    }
  ]
}
```

Illustrative: an exact identifier match, a model and capacity match, a two-pack rejected.

## What product matching is used for

Every cross-store price or range comparison rests on matching.

| Job | Data you collect | Typical sources |
|---|---|---|
| Competitor price comparison | GTIN, model, brand, size, current price | Amazon, Walmart, Target, Best Buy, Currys, MediaMarkt |
| Assortment overlap | Your SKUs each competitor carries or lacks | Category and search pages |
| Private-label equivalents | Closest store brand by size, ingredients, specs | Tesco, Boots, Walgreens |
| Comparison and affiliate sites | One card, many merchants' offers | Merchant pages, Google Shopping offers |
| Cross-border matching | Regional codes and titles | Amazon .com, .co.uk, .de; local retailer sites |
| Duty-free and travel retail | Brand SKUs under each shop's own codes | Airport e-shops, travel retail sites |

## Why product matching data is hard to collect

Most errors are missing evidence: without the GTIN, matchers fall back on titles, where retailers differ most.

| What breaks | What it looks like | How Zenrows handles it |
|---|---|---|
| Identifiers not visible | The GTIN sits in metadata or a script | Extract returns fields plus raw HTML, keeping markup identifiers |
| Pack size and units | "6 x 330ml" matched to "330ml" | Separate CSS selector fields for size and quantity |
| Variants behind one URL | Four colours, four codes, one page | Adaptive Stealth Mode renders until variant data loads |
| Search pages blocked | GTIN or model searches return a CAPTCHA | Zenrows' stealth gets through on search and product pages |
| Retailers in many countries | Codes change by market | Geotargeting fetches each market's catalog locally |

## How to collect product matching data with Zenrows

1. **Start from identifiers**: GTIN or model per retailer, then brand plus title.
2. **Keep the candidate URLs** with a CSS selector links field.
3. **Fetch the candidates as JSON** in Batch with Extract: identifiers, size, specs, price.
4. **Match in order of strength**: GTIN, then model plus specs, then review or a model.

```python Python
# pip install requests
import json
import requests

GTIN = "5012345678900"  # from your own catalog

# Example selectors: match them to each retailer's search results page
params = {
    "apikey": "YOUR_ZENROWS_API_KEY",
    "url": f"https://retailer-a.example/search?q={GTIN}",
    "mode": "auto",
    "css_extractor": json.dumps({
        "titles": ".product-card .title",
        "links": ".product-card a @href",
    }),
}
results = requests.get("https://api.zenrows.com/v1/", params=params).json()
candidates = results["links"][:5]
print(candidates)
```

```bash cURL
curl -X POST "https://async.api.zenrows.com/v1/jobs" \
  -H "X-API-Key: YOUR_ZENROWS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "type": "regular",
    "status": "closed",
    "zenrows_params": { "mode": "auto", "extract": "auto" },
    "tasks": [
      { "url": "CANDIDATE_URL_1", "external_id": "INT-55021:retailer-a" },
      { "url": "CANDIDATE_URL_2", "external_id": "INT-55021:retailer-b" }
    ]
  }'
```

```bash CLI
npx -y @zenrows/cli init
# Candidate links from a retailer's search results for one GTIN
zenrows extract "https://retailer-a.example/search?q=5012345678900" \
  --css '{"titles":".product-card .title","links":".product-card a @href"}'
# candidates.jsonl: one {"url": "...", "external_id": "SKU:retailer"} per line
zenrows batch create candidates.jsonl --wait
zenrows batch results <job-id> --out candidates-results.jsonl
```

```text Claude Code (MCP)
# Connect once
claude mcp add zenrows -e ZENROWS_API_KEY=YOUR_ZENROWS_API_KEY -- npx -y @zenrows/mcp

# Then ask:
Use Zenrows to open this Currys product page and this Argos product page. Compare
brand, model number, capacity, pack size and colour, and tell me whether they are
the same product. Quote the field that decides it.
```

## Which Zenrows product fits

| If you need to | Use | Why |
|---|---|---|
| Search a retailer by GTIN or model | [Fetch](/products/fetch) | CSS selectors return titles and links |
| Get identifiers and specs across many retailers | [Extract](/products/extract) | One schema-free request per page |
| Check a whole catalog's candidates at once | [Batch](/products/batch) | Thousands per job, your SKU as external ID |
| Resolve the hard pairs with a model | [MCP Server](/mcp) | An agent fetches and compares both pages |

## What success looks like

| Measure | What good looks like | How to check it |
|---|---|---|
| Match coverage | Most SKUs matched wherever stocked | SKUs with a match, per retailer |
| Match precision | Matches share product, size and pack | Weekly review of a random sample |
| Identifier capture | A GTIN or model on most candidate pages | Empty identifier fields per retailer |
| Time to match | New listings matched within one cycle | First seen to matched |
| Rematch rate | Matches survive URL changes and relistings | Matches lost between runs, and why |

## Product matching FAQ

### What is product matching?

Linking one product's listings across stores to one record. Titles, codes and images differ, so a matcher compares identifiers (GTIN, brand, model number), then attributes (size, colour, pack count), enabling item-level price, assortment and availability comparison.

### How does AI product matching work?

Rules plus a model: code accepts exact identifier matches; a model judges the rest from titles, specs and sometimes images, with a reason. Feed it structured fields and clean text, not raw HTML.

### What if a retailer does not show the GTIN?

It is often in structured data or scripts, which Extract and raw HTML surface. Otherwise match on brand plus model, confirmed by two or three specs: capacity, pack size, wattage.

### How do I match products across countries?

Geotargeting fetches each market locally: catalog, currency, titles. GTINs often hold across regions while retailer codes change, so key on GTIN or manufacturer model; translate titles only as fallback.

### How often should matches be refreshed?

Less often than prices: when a URL stops resolving, an identifier changes or a product is relisted. Search weekly for newly added products.

### What is the difference between product matching and product mapping?

Matching decides two listings are one item. Mapping stores the result: each SKU's URL and code at every competitor, kept current. Pricing, assortment and MAP teams query the map; matching runs for new or broken entries.

### How do I do a GTIN lookup across retailers?

Search each retailer by GTIN (many index barcodes), fetch top results, and accept only an exact page GTIN. Google Shopping helps when a retailer's search returns nothing.

## Related use cases

- [Web scraping for price comparison](https://www.zenrows.com/use-cases/web-scraping-for-price-comparison.md): Merchant offers, prices, shipping and identifiers for comparison sites and apps, from catalogs without a feed.
- [Competitor price scraping](https://www.zenrows.com/use-cases/price-scraping.md): Competitor prices, promotions, shipping and stock per SKU from retailers and marketplaces, refreshed on your schedule.
- [Product data enrichment](https://www.zenrows.com/use-cases/product-data-enrichment.md): Specs, images, descriptions and GTINs from manufacturer and retailer pages, written into the gaps in your PIM.
- [Product data scraping](https://www.zenrows.com/use-cases/product-data-scraping.md): Whole product catalogs, variants, specs, images and sellers from marketplaces, normalized into one feed.
- [Counterfeit detection](https://www.zenrows.com/use-cases/counterfeit-detection.md): Listing prices, seller history and product images from marketplaces and resale apps, scored to flag likely fakes.
- [Digital shelf analytics](https://www.zenrows.com/use-cases/digital-shelf-analytics.md): Search rank, share of shelf, content compliance, ratings and availability for CPG brands, per retailer and region.

Start free: https://app.zenrows.com/register. Docs: https://docs.zenrows.com

## Structured data



<script type="application/ld+json">
{"@context":"https://schema.org","@graph":[{"@type":"WebPage","@id":"https://www.zenrows.com/use-cases/product-matching#webpage","url":"https://www.zenrows.com/use-cases/product-matching","name":"Product matching","description":"Product matching data: get GTINs, model numbers, titles, pack sizes and images from Amazon, Walmart and other retailers to match one item across stores.","isPartOf":{"@id":"https://www.zenrows.com/#website"},"inLanguage":"en","datePublished":"2026-10-19","dateModified":"2026-10-19","about":{"@type":"Thing","name":"product matching"},"keywords":"product matching, product matching ai, product mapping, ecommerce product matching, cross-retailer product matching, sku matching, gtin lookup","publisher":{"@type":"Organization","name":"Zenrows","url":"https://www.zenrows.com/"}},{"@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://www.zenrows.com/"},{"@type":"ListItem","position":2,"name":"Use cases","item":"https://www.zenrows.com/use-cases"},{"@type":"ListItem","position":3,"name":"Product matching","item":"https://www.zenrows.com/use-cases/product-matching"}]},{"@type":"FAQPage","@id":"https://www.zenrows.com/use-cases/product-matching#faq","mainEntity":[{"@type":"Question","name":"What is product matching?","acceptedAnswer":{"@type":"Answer","text":"Linking one product's listings across stores to one record. Titles, codes and images differ, so a matcher compares identifiers (GTIN, brand, model number), then attributes (size, colour, pack count), enabling item-level price, assortment and availability comparison."}},{"@type":"Question","name":"How does AI product matching work?","acceptedAnswer":{"@type":"Answer","text":"Rules plus a model: code accepts exact identifier matches; a model judges the rest from titles, specs and sometimes images, with a reason. Feed it structured fields and clean text, not raw HTML."}},{"@type":"Question","name":"What if a retailer does not show the GTIN?","acceptedAnswer":{"@type":"Answer","text":"It is often in structured data or scripts, which Extract and raw HTML surface. Otherwise match on brand plus model, confirmed by two or three specs: capacity, pack size, wattage."}},{"@type":"Question","name":"How do I match products across countries?","acceptedAnswer":{"@type":"Answer","text":"Geotargeting fetches each market locally: catalog, currency, titles. GTINs often hold across regions while retailer codes change, so key on GTIN or manufacturer model; translate titles only as fallback."}},{"@type":"Question","name":"How often should matches be refreshed?","acceptedAnswer":{"@type":"Answer","text":"Less often than prices: when a URL stops resolving, an identifier changes or a product is relisted. Search weekly for newly added products."}},{"@type":"Question","name":"What is the difference between product matching and product mapping?","acceptedAnswer":{"@type":"Answer","text":"Matching decides two listings are one item. Mapping stores the result: each SKU's URL and code at every competitor, kept current. Pricing, assortment and MAP teams query the map; matching runs for new or broken entries."}},{"@type":"Question","name":"How do I do a GTIN lookup across retailers?","acceptedAnswer":{"@type":"Answer","text":"Search each retailer by GTIN (many index barcodes), fetch top results, and accept only an exact page GTIN. Google Shopping helps when a retailer's search returns nothing."}}]}]}
</script>
