What is a product feed audit?
A product feed audit is a diagnostic pass over the product data you send to shopping channels — Google, Meta, TikTok, and the rest — that flags missing required fields, thin optional coverage, and low-quality values that reduce reach, ranking, or ad eligibility. A good audit doesn't just list problems; it scores your feed against each channel's spec and ranks the fixes that will move performance the most.
How the product feed audit works
- Upload or paste a feed link. We accept CSV, TSV, and XML — including Google Shopping RSS with the g: namespace. If your feed lives at a public URL, paste it and we'll fetch it. No login, no export step.
- Auto-mapping. We match your column names to the canonical shopping feed schema. product_name → title, msrp → price, image[1].url → additional_image_link, and dozens more. Alternate formats and namespace prefixes are handled automatically.
- Per-channel scoring. Every SKU is validated against seven channel specs (Google, Meta, TikTok, Pinterest, Snap, Reddit, Applovin). Required fields, recommended fields, and value-quality rubrics all contribute to a single 0–100 score per channel.
- Ranked fixes with projected lift. The audit surfaces the highest-impact issues first — with an estimated point lift on each channel if you apply the fix. Sorted by total impact, not alphabetical.
What a product feed audit optimizes for
The Marpipe Feed Auditor evaluates each channel on three weighted dimensions. The exact fields and weights vary by channel, but the shape is consistent everywhere.
Required coverage (40% of score)
The non-negotiables. Every channel has a list of fields that must be present and valid for a product to be eligible — id, title, description, link, image_link, price, availability, and channel-specific extras like condition on TikTok or brand on Meta. Each SKU is validated (URLs must resolve, prices must include a currency, availability must match the allowed enum), and the coverage percentage across your catalog rolls up into the required score.
Recommended coverage (30% of score)
Optional fields that unlock better matching, more surfaces, and richer ads: brand, gtin, color, size, additional_image_link, google_product_category, custom_label_0–4, and more. Higher-weight fields matter more — for example, google_product_category is weighted heavily on every channel because it drives category matching, while sale_price is scored presence-only (some catalog has it, or none does — not every SKU is on sale).
Field utilization quality (30% of score)
Even present fields can be under-utilized. This dimension grades the content of your feed:
- Title optimization. Length, keyword coverage, category signals, structured attributes like size and color inline.
- Description quality. Length, uniqueness, and channel-appropriate tone.
- Google product category depth. Deeper GPC categories (e.g., "Apparel > Women's > Dresses > Maxi Dresses") outperform shallow ones on discovery.
Product feed optimization by channel
The auditor scores your feed against seven separate specs. Each channel gets its own scorecard so you can see exactly where you stand — and where the biggest gains are.
Google Shopping feed
Google Merchant Center is the strictest spec, and it's the one most other channels model themselves on. The auditor checks required fields (id, title, description, link, image_link, availability, price, brand, condition), promotes gtin to a hard requirement because Google penalizes missing identifiers, and grades google_product_category depth. RSS 2.0 XML feeds with the g: namespace are parsed natively — no need to convert.
Meta and Facebook product feed
Meta's DPA (Dynamic Product Ads) spec overlaps heavily with Google but weights brand and google_product_category more aggressively for catalog matching. Meta comment metadata (# ref_application_id, # ref_asset_id) at the top of CSV exports is skipped automatically. The auditor also checks item_group_id for variant grouping — required if your catalog has color/size variants.
TikTok product feed
TikTok Shopping is a full Google-clone spec plus a hard brand requirement. Custom labels (custom_label_0–4) are grouped and scored together because TikTok's discovery algorithm leans on them for audience segmentation. google_product_category carries the highest quality weight on this channel.
Pinterest product feed
Pinterest treats title and description as ranking signals in visual search, so title optimization and description quality are weighted higher here than on ad-first channels. Rich attributes like color, size, and material boost eligibility for filtered surfaces.
Snapchat product feed
Snap's spec is close to Meta's but with lighter recommended-field requirements. The auditor grades required-field completeness strictly and treats sale_price as presence-only — Snap surfaces promotional pricing when any part of the catalog is on sale.
Reddit product feed
Reddit's product catalog for ads is newer and less prescriptive, but the auditor evaluates the same core coverage and quality dimensions to keep results comparable across channels.
AppLovin product feed
Applovin's Marpipe-native spec is designed for programmatic creative — it emphasizes google_product_category depth, additional_image_link coverage, and sale_price presence for dynamic pricing overlays.
Ranked fixes with projected lift
Every audit ends with a prioritized fixes panel. Each fix shows:
- What to change. A concrete instruction — "populate gtin on the 4,231 SKUs missing it" — with example rows from your actual feed.
- Projected lift. How many points your score goes up per channel if you apply the fix, with the total lift across channels rolled up at the top.
- Which channels benefit. Fixes are channel-aware. gtin moves Google's score dramatically; it barely nudges Pinterest.
- Effort signal. Presence-only fixes (add one field) are quick wins. Rubric-driven fixes (rewrite titles) are higher effort but higher ceiling.
Fixes are sorted by total lift, so the top of the list is always your best next move.
Product feed audit FAQ
Why should I optimize my product feed?
Your feed is the source of truth every shopping channel uses to decide what to show, when, and to whom. Missing required fields disqualify products entirely. Thin recommended fields shrink your addressable audience. Low-quality titles and descriptions lose ranking auctions to competitors with better metadata. An optimized feed is the highest-leverage change you can make in shopping performance — it compounds across every campaign, every channel, every day.
Is the product feed audit free?
Yes. The audit runs entirely in your browser and there's no paywall on results. You get the full scorecard, every fix, every channel.
What are the most common product feed errors?
The recurring ones, in rough order of frequency:
- Missing gtin on identifier-required channels (especially Google).
- Shallow google_product_category — using a top-level category like "Apparel" instead of "Apparel > Women's > Dresses > Maxi Dresses."
- Titles that skip key attributes — no brand, no color, no size in the title itself.
- Empty additional_image_link — many feeds ship one hero image and nothing else.
- Missing item_group_id on variant catalogs, which flattens variants into separate unrelated products.
- Invalid availability values — free-text like "yes" or "available" instead of the required enum (in stock, out of stock, preorder, backorder).
- Prices without currency — 19.99 instead of 19.99 USD.
What's the difference between a feed audit and feed optimization?
An audit is diagnosis — it tells you what's wrong and what to fix. Optimization is treatment — actually making the changes. The Marpipe Feed Auditor does the audit for free. If you want the optimization done for you — transformed feed output, ongoing monitoring, and per-channel variant generation — that's what the full Marpipe platform handles.
Privacy and data handling
Your feed never leaves your browser. The audit runs entirely client-side: parsing, column matching, scoring, and fix generation all happen in JavaScript on your device. We don't upload your catalog, we don't store it, and we don't share it. If you paste a public feed URL, we fetch it through a lightweight proxy solely to bypass browser CORS restrictions — the response is streamed straight to the parser and discarded when you close the tab.
We use standard product analytics (page views, tool interactions) to improve the product, but the contents of your feed are never included in that data.