Google Shopping Feed vs Website Checker

Catch feed-to-store drift before customers—and ad platforms—do.

Merchant Feed Consistency Checker reads a public Google Merchant product feed and compares each selected item with the live product page it links to. It detects price drift, availability disagreements, dead links, redirects, missing links, duplicate IDs, and pages whose product data cannot be extracted reliably.

Use it before submitting a feed, after catalog or storefront deployments, and on a schedule to catch quiet inconsistencies that can waste ad spend or trigger Merchant Center disapprovals.

Quick start

{
  "feedUrl": "https://shop.example/google-merchant-feed.xml",
  "maxItems": 100,
  "sampleEvery": 1,
  "onlyMismatches": true,
  "maxFeedMegabytes": 64
}

The feed must be reachable without a login. Supported formats include Google Shopping XML/RSS/Atom, CSV, and TSV; plain and gzip-compressed files are accepted.

What it checks

For every selected feed item, the Actor compares the feed with public product-page signals:

• Feed price versus live-page price
• Feed availability versus live-page availability
• HTTP errors and dead product links
• Redirected product URLs
• Missing product links
• Duplicate product IDs
• Product data that is absent from the live page

The page extractor prefers structured product data where available and normalizes prices and availability before comparing them.

Why feed drift is easy to miss

A feed generator and a storefront often update on different schedules. Each system can appear healthy while the values customers see no longer match the values sent to an ad platform. The result can be wasted clicks, disapprovals, or confusing landing pages without any obvious application error.

This Actor turns that silent failure into a deterministic dataset with feed values, page values, the numeric difference, HTTP evidence, and a per-run summary.

Input

Field	Description
feedUrl	Public XML, CSV, or TSV feed URL
maxItems	Maximum products compared with live pages
sampleEvery	Check every Nth item across the feed
onlyMismatches	Omit consistent item rows; omitted rows are not charged
maxFeedMegabytes	Download limit for unexpectedly large feeds
crawlerIdentity	Optional custom User-Agent

For a large catalog, start with maxItems: 100 or increase sampleEvery to sample the entire file rather than checking only its first products.

Output

The default dataset uses one Feed consistency results view. Check rowType to distinguish three records.

Item comparison

{
  "rowType": "item",
  "ok": true,
  "id": "SKU-1042",
  "title": "Trail Runner",
  "link": "https://shop.example/products/trail-runner",
  "position": 42,
  "mismatches": ["price"],
  "feedPrice": 89.99,
  "pagePrice": 94.99,
  "priceDelta": 5,
  "feedAvailability": "in_stock",
  "pageAvailability": "in_stock",
  "httpStatus": 200
}

Unreachable page

A free, ungraded row with the feed ID, product URL, position, HTTP status, and network error. A request timeout is not reported as a merchant mismatch.

Run summary

A free row with detected feed format, parsed items, checked items, products with problems, price mismatches, availability mismatches, dead links, duplicate IDs, and truncation status.

Pricing

Pay per event:

• $0.01 per Actor start at the default 512 MB memory
• $0.002 per delivered item comparison — $2.00 per 1,000
• Unreachable pages are free
• The run summary is free
• Products omitted by onlyMismatches are free
• Apify platform usage is included

Examples before user-level discounts:

Delivered comparisons	Estimated price
10	$0.03
100	$0.21
1,000	$2.01

The Actor honors the maximum total charge configured for a run and stops before continuing unpaid work.

Sampling large feeds

sampleEvery selects products throughout the feed instead of taking only the first page. For example, sampleEvery: 10 checks roughly 10% of the feed, subject to maxItems. This is useful for daily health checks on catalogs with tens of thousands of products.

Use maxFeedMegabytes as a safety limit. If the feed exceeds it, the Actor marks truncated: true in the summary so the result cannot be mistaken for a full audit.

Reliability

Requests to one host are rate-limited, body sizes and timeouts are capped, and HTTP 429 receives a backed-off retry. The feed URL and product URLs must be public. The Actor does not log in, bypass paywalls, solve anti-bot challenges, or upload anything to Merchant Center.

A product page that cannot be reached due to a network failure is returned as unreachable, free, and compared against nothing. HTTP statuses observed from a reached server remain useful evidence for dead-link checks.

Limitations

• The Actor compares public feed and page signals; it does not access your Merchant Center account or Diagnostics tab.
• No JavaScript rendering: values that exist only after client-side execution may not be visible in the initial HTML.
• Price comparison uses normalized numeric values. Taxes, shipping, market-specific prices, variants, or logged-in pricing can legitimately differ.
• Availability vocabularies are normalized, but custom text can remain unknown.
• CSV and TSV column names must be recognizable feed fields.
• Results are point-in-time checks and can change when the feed or storefront updates.

Automation

Schedule the Actor hourly, daily, or after feed-generation deployments. Use Apify webhooks and integrations to send mismatches to a spreadsheet, database, Slack workflow, ticketing system, or downstream Actor.

Responsible use and support

Check only feeds and public pages you are authorized to audit. Follow applicable website terms, laws, advertising policies, and organizational rules. The output is technical evidence, not a guarantee of Merchant Center approval and not legal or financial advice.

For support, open an Actor issue with the run ID, a sanitized feed sample, and one unexpected dataset row. Never post credentials, private feeds, or customer data.