Run This 15-Minute Shopify Audit to Find the 10% of Products Driving 75% of Your Sales

By Steve Merrill | April 30, 2026

A practical Shopify catalog audit for finding the small group of products that drives most revenue, then fixing the data AI shopping tools actually read.

Why start with the 10% of products driving 75% of sales?

Your best products deserve the first pass because they already have proof. If AI shopping tools misunderstand those products, you lose recommendations on the items most likely to convert.

I see this in audits all the time. The winners have weak data while slow movers get the same attention. The client note that kicked this off was blunt: roughly 10% of the catalog drove about 75% of revenue.

That is normal in ecommerce. The mistake is treating every SKU like it has the same upside. Same story. Different store.

AI shopping systems read product titles, descriptions, variants, availability, reviews, structured data, and feed fields. Google explains how product structured data helps search systems understand product details in its Product structured data documentation. Google Merchant Center also documents required product attributes in its product data specification.

What does a 15-minute Shopify catalog audit look like?

A fast audit starts with revenue concentration, then checks whether top products are clear to machines. Pull your top 20 products by revenue, open each product record, and score the fields an AI assistant would need before recommending it to a buyer.

Use three columns: revenue rank, product data score, and risk. Revenue rank comes from Shopify analytics. Product data score comes from your title, description, images, variants, availability, reviews, schema, and feed attributes.

Risk flags anything that could make an agent skip the product. Most stores find ugly gaps fast. A best seller with a vague title. A size variant marked out of stock by accident. A product description written like a brand poem.

Looks nice. Tells the crawler almost nothing.

Which product fields should you check first?

Check title, handle, description, availability, variant names, price, image alt text, review count, product type, vendor, GTIN or MPN, and structured data first. Those fields explain what the product is, who it fits, whether it can ship, and why a buyer should trust it.

Shopify stores often hide the real product answer in lifestyle copy. A human can infer the product. A crawler may not. The title should say the plain thing people buy.

The first paragraph should answer who it is for, what it does, and why it is different. I would rather see a boring clear title than a clever one that fails the feed test. "Waterproof Leather Dog Collar - Medium - Brown" beats "The Trail Companion" when an AI agent is comparing options.

Shopify's own product admin docs explain the product fields merchants control, including title, description, media, pricing, inventory, and organization fields. Those fields are not paperwork. They are machine-readable context.

How do you score each product without overthinking it?

Give each top product a score from 0 to 10. One point each for a clear title, direct first paragraph, complete variants, correct availability, useful images, review signals, product category, GTIN or MPN when relevant, clean schema, and feed-ready attributes.

Don't turn this into a month-long project. Fifteen minutes is enough to spot the problem. If your top revenue product scores 4 out of 10, that is your first fix.

If a low-volume SKU scores 9 out of 10, leave it alone for now. The point is sequence. Fix the products that pay the bills first. Then move into mid-tier products that have demand but weak data.

That second group is where AI recommendations can create a new breakout.

What AI visibility problems show up in this audit?

The biggest problems are vague descriptions, missing product identifiers, bad availability, thin variant data, and weak review context. These are boring issues. They also decide whether an AI shopping answer has enough confidence to recommend your product.

We had one audit where an inventory flag made a product look unavailable even though the store could ship it. Another where the best seller had no useful product type.

Another where variants were named "Option 1" and "Option 2." Zero. Nothing. Blank.

AI assistants need confidence. A complete product record gives them that. Google Search Central says merchants can add product markup for price, availability, ratings, and shipping details, which gives search systems cleaner product facts.

How should Shopify stores fix the winners first?

Fix your top products in three passes: clarity, completeness, and trust. Clarity means the product page says exactly what the item is. Completeness means every feed and variant field is filled. Trust means reviews, policies, shipping details, and return information are easy to read.

Start with the product title and the first 150 words. Then fix variants and availability. Then add missing identifiers and product categories.

Last, review the page the way an AI answer would quote it: can it pull a direct sentence that explains why this product fits a buyer?

I know that sounds basic. Good. Basic is where most of the money leaks out.

What should you do with the other 90% of the catalog?

The rest of the catalog still matters, but it gets a different plan. Group products into winners, possible breakouts, and maintenance SKUs. Winners get full fixes now. Possible breakouts get data cleanup next.

Maintenance SKUs get minimum accuracy so they do not pollute the feed. One ignored mid-tier product can become the answer if the data is clearer than competitors.

That is why I would not delete the long tail from the work plan. I would stage it. Priority first. Coverage second.

This is where most teams get stuck. They try to fix 800 products at once, finish none, then wonder why AI visibility did not move. Pick the 20 that matter. Finish them. Then repeat.

Check Your Store's AI Readiness ->