Google Merchant Center Disapproved Products: What to Fix Before Black Friday

20 min read · Updated Nov 17, 2026

Contents

When Merchant Center disapproves a product, it stops showing in Shopping ads, and the account’s spend gives you no sign of it. The budget simply moves to the products that can still show, so campaigns keep spending as usual on different items. A few weeks before Black Friday that makes disapprovals the most expensive problem a feed can have, because on the account’s charts they look like a quiet week. This guide covers how to read disapprovals, which causes to fix first, which “not eligible” products need no fix, how to fix a landing page error, a price mismatch and image problems, what to check in robots.txt, how to measure a fix, and what has to be done before Black Friday. The rules at the end are ready to drop into your own playbook.

This is how we review product feeds for the stores we manage. The method doesn’t depend on any particular tool: every step works by hand in Merchant Center and the Google Ads UI, and an AI assistant takes over the lookups and the math. The video above shows that work on a store we built as an example. Every problem in its feed comes from a real account, and we scaled the numbers to a single store with 4,000 products.

What the AI handles and what you decide

A disapproval review is mostly lookups, and with a few thousand products they take hours by hand. An AI assistant with access to Merchant Center and Google Ads gets through them in minutes. It:

  • reads the status of every product, separately for each destination,
  • groups disapprovals by cause and weighs each cause by the clicks those products used to get,
  • separates disapprovals from products that aren’t showing for other reasons,
  • opens a sample of product pages the way Google’s crawler does and compares the page price with the feed price,
  • finds the day the drop began in the click history,
  • sets out what to check after each fix, and when.

What no lookup can settle stays with you: the conversation with the store’s developer, the changes to the feed, the request for a feed re-fetch, and the decision on which fixes have to land before the sale. In the video, the session only reads. Nothing on the account changes.

Step 1: Separate Merchant Center status from eligibility in Google Ads

Merchant Center judges the product. Google Ads tells you whether the product can show in a campaign. Those are two different questions, and mixing them up gives you a false picture of the feed. “Not eligible” in Google Ads has many causes, and a disapproval is only one of them.

Start the review with a summary of the whole feed: how many products there are, how many are in stock, how many are approved with no issues, how many are limited, how many are disapproved and how many are pending. Against that baseline you can see how big the problem is.

  • Google evaluates each product separately for each destination: Shopping ads, dynamic remarketing and free listings. A product can be approved for ads and disapproved for free listings, or the other way around.
  • A disapproval in dynamic remarketing alone doesn’t stop Shopping ads. Common causes are personalized advertising policies or Merchant Center add-ons that aren’t active.
  • “Limited” means the product is approved in some destinations only. A limit on personalized ads alone doesn’t affect Shopping ads, so count those products as serving. Still check which destinations disapprove the product, because a limited product can be approved only in free listings and disapproved in every ad destination.
  • A yellow warning on a product is for information only. A red disapproval is what affects campaigns. The exception is an account-level warning with a deadline, such as one about price mismatches: it announces a suspension and goes to the top of the list.
  • The count of products “disapproved everywhere” covers every destination a product is sent to. A product sent only to free listings gets that status even though it never reached ads, which is why this count can run higher than the number of products disapproved for Shopping ads.
  • If Merchant Center approves products and Google Ads doesn’t see them, check the marketing methods in Merchant Center (Shopping ads and free listings). A common mistake is opting in to free listings only, without ads.

Take the product list and the disapproval reasons from Merchant Center. Read the number of products that can’t show in ads from the Shopping ads row of the per-destination breakdown, because the overall disapproved count doesn’t give you that number. Check it in Google Ads too: its product report covers a narrower set, and disapproved counts can drift in either direction. In one account we manage, Google Ads saw 1,270 products against 1,345 in Merchant Center. In another, Google Ads showed the landing page error code on 2,358 products while the Merchant Center disapproved count showed 239, and it was the Merchant Center number that was too low. When the two disagree, compare them by issue code and line both up against the products’ clicks before you tell a client how many products are broken.

With several feeds, keep in mind that the same product ID can legitimately live in more than one feed, under different feed labels. A table keyed on the ID alone merges them into one row and undercounts the catalog.

Step 2: Rank causes by the clicks they cost

Merchant Center sorts disapprovals by product count. What should set the order of your fixes is how much traffic each cause has taken. A product that never got a click can wait. A hundred products that used to get clicks can’t.

For each cause, pull two numbers:

  • the weekly average clicks of the affected products over the four full weeks before the drop,
  • the clicks of the same products over the last seven days.

The gap between them is the traffic that cause took, and the day the daily series starts falling tells you when the problem began. In the video, a landing page error on 410 products cut them from 1,321 clicks a week to 104, and it started seven weeks ago. A price mismatch on 380 products began six days ago: 1,384 clicks a week before it, 455 in the last week. Image overlays on 60 products took nothing, because those products never had a click.

  • Rate each product’s urgency by its Shopping ads clicks over the last 90 days: a product with traffic is urgent. The comparison of the four weeks before the drop with the last seven days answers a different question: how much traffic the whole cause took, and since when. Skip conversions and conversion value from Merchant Center for this, because they come from a different model and often show zeros. Take conversions from Google Ads.
  • A product with no traffic is low priority. That doesn’t mean the store dropped it: it may never have been in ads at all.
  • One product can be disapproved for several reasons at once, so the product counts in a table of causes can add up to more than the number of disapproved products. Say so under the table.
  • You can also put the damage in money terms. In one account, products with a landing page error carried 13.2% of the account’s conversion value over 90 days. In three others, they carried 39%, 69% and 74% of 30-day cost.
  • A misrepresentation disapproval or a suspension affects the whole account or the whole feed. Those go ahead of any cause that affects individual products.
  • With a misrepresentation disapproval, Merchant Center doesn’t tell you the cause. Check the business details and contact information, terms, privacy policy, shipping and returns policies, and whether the business name is the same everywhere. Appeal once, after fixing everything together, because appeals filed mid-fix risk another block.
  • A suspension for circumventing systems on a clean site can come from a redirect or a script that a plugin injects into the generated feed file. Open the raw contents of the file and read it line by line.

Then check one more pattern. If campaigns are spending and products that used to sell have shown cost with zero conversion value for several days, suspect a broken page. Seasonality doesn’t look like that.

Step 3: Out of stock and no campaign, two traps

Some products show as “not eligible” in Google Ads even though their feed entries are fine. In the video that’s 720 products, almost as many as are disapproved.

  • An out-of-stock product doesn’t show in ads, and it isn’t supposed to. There’s nothing to fix in the feed. Before you blame the account’s settings for a drop, check inventory, because out-of-stock products can look like the account’s best sellers in historical data.
  • A product with no campaign carries the reason “no campaign is advertising this product” in Google Ads, or it’s excluded in a campaign’s product tree. That’s a targeting decision, made at the campaign level. The feed is fine.
  • “Approved” in Merchant Center ignores inventory. An approved product that’s out of stock still doesn’t show.
  • Don’t use out of stock to hide a product while the page is being worked on. To pause a product, use the pause attribute, and take discontinued products out of the feed.
  • Keep products you don’t want to advertise in the feed. Exclude them in the campaign’s product tree or with a label in the feed, and they can keep collecting free listings traffic in Merchant Center.

Before you report that disapprovals are costing sales, read the reason on every “not eligible” product. Twice within eight days we diagnosed a “feed failure” from the status alone, and both times the cause was campaign targeting: once the products were excluded in a campaign’s product tree, once no campaign advertised them.

Step 4: Landing page errors. Open the pages the way Google’s crawler does

A landing page error does its damage quietly. For the first few days the product keeps showing and collecting clicks, conversions drop to zero right away, and Google disapproves the product only days later. In one account a product went eight days in a row without a conversion and was still getting 164 clicks on the eighth day. Nothing flagged the error the whole time.

Diagnosis starts with a sample of the affected product pages. For each one, check the response code, the page title and whether the page shows a price. In the video, 10 of 10 pages on the old URL path answer with a 403 and a “Just a moment...” screen. That’s a bot check, and it stops Google’s crawler too. The usual causes:

  • A firewall or bot check blocks Google’s crawler while the page works fine for people. The fix is an exception for Googlebot and for Storebot-Google, the crawler that checks product pages for Google Shopping. Set the exception for verified Google crawlers, for example with the firewall’s verified bots option, because a crawler’s name alone can be spoofed.
  • The server responds too slowly. After waiting, the crawler marks the page unavailable, while you see a working page. This one tends to come in waves, a few products a day.
  • Redirects from the feed URL to another page or another domain. The link in the feed should open the product page directly, on the domain verified in Merchant Center.
  • Old URLs after a store migration or a change to the URL structure.

A test from your own computer won’t settle it. A page can return 200 even with Googlebot’s user agent while the firewall still blocks Google’s IP addresses. The store’s server logs settle it.

When the same error shows up at once in several unrelated stores, on different platforms, while their pages work, hold off on blaming the store. Check the pages and give Google one to three days to recrawl. Sometimes, after a week of searching for the cause, Google announces the bug was on its side.

Once the page is fixed, ask Merchant Center to re-fetch the feed. That speeds up the fetch itself, but the products come back only after the crawler revisits the pages. Keep product IDs as they are while you fix things, because a new ID wipes the product’s history in Merchant Center and in your campaigns.

Step 5: Price mismatch. The feed has to keep up with the site

Google compares the feed price with the price on the page and at checkout. Any gap can end in a disapproval, and before a sale, prices change faster than the store fetches its feed. In the video, the page shows sale items at a price 20% below the feed.

  • The feed should mirror what the page shows, shipping costs included. In the US, prices go in without sales tax, the way the page shows them. Where prices include VAT, the feed includes it too.
  • A discount goes in sale_price, and price keeps the regular price. The sale window goes in sale_price_effective_date: two dates with time and time zone, separated by a slash. Without a time, the sale starts at 00:00 and ends at 23:59, and without a time zone Google uses UTC. A malformed value doesn’t trigger a disapproval. You get a notice that the information can’t be used, which is easy to miss.
  • During the sale, schedule a daily fetch for a time after the store updates its prices. Put discounts with known times into sale_price_effective_date ahead of time, so the feed doesn’t have to chase the site. A price change during the day goes through with a manual fetch. When prices change more than once a day, Google recommends sending products through the API instead of fetching the file again and again.
  • Automatic item updates in Merchant Center correct the price and availability in your ads from the page’s structured data. The crawler does this with a delay, so treat them as a fallback for a mismatch, most useful for feeds maintained by hand. Send sale prices through the feed, which you control. A separate matter is the data source Merchant Center creates on its own by crawling the site: it can put products you don’t want to advertise into your ads, so it’s worth pausing.
  • A price you correct by hand in Merchant Center lasts only until the next feed fetch overwrites it. A lasting fix goes into the feed’s source or into a feed management tool that sits between the store and Merchant Center.
  • Different variant prices on a single URL produce a price mismatch. Each variant’s link should open a page showing its price.
  • Prices visible only after login, while the feed lists a price, count as misrepresentation, and Google can take down the whole feed. A price that’s only available with a discount code, and differs from the price in the ad, puts the account at risk of suspension.
  • Keep promotional text out of titles (“10% off with code”). That earns a warning from Google.

You can’t force the sale price annotation, the struck-through price. It appears when the discount is over 5%, the previous price was active for at least 30 of the last 200 days, and the feed has a sale price. A missing sale price annotation on a product that shows the reduced price correctly is not an error.

Step 6: Image overlays. A small fix when they cost no clicks

Promotional text, borders and logos on the main image lead to a disapproval. Watermarks pass in one account and get disapproved in another, so don’t count on them.

  • Put the new image at a new URL. Google recrawls it within 24 to 72 hours, while an image replaced at the same URL can take up to six weeks to update.
  • If the disapproved products never had clicks, this is a small fix for a slower day. Before Black Friday, replace images on the products that sell first.
  • If you can’t replace the images before the sale, turn on automatic image improvements in Merchant Center. Google then removes some promotional overlays on its own, which can prevent a disapproval.
  • The recommended image size is 1500 × 1500 pixels. With other proportions, Google crops the image in the ad automatically.

Step 7: Check robots.txt after every change, including AI-bot blocks

More and more stores block AI companies’ crawlers in robots.txt. One line too many stops Google’s crawler as well. In the video, a block for AI bots ended with two lines: User-agent: * and Disallow: /. The first applies to every crawler that has no section of its own in the file, and the second bars it from the whole site. If Googlebot and Storebot-Google have no sections of their own, that rule blocks them along with the AI bots.

In one account we manage, a rule like that stopped Google Shopping traffic for four days: daily clicks fell from about 540 to 65–77 and came back to 565 the day after the store removed the two lines.

  • After every robots.txt change, check that Google’s crawlers can reach the product pages and the image URLs.
  • When clicks drop across every product overnight, compare the date with the last change to robots.txt and to the firewall settings.
  • Keep notes on site changes next to the click history. Without them, the cause of a drop like this surfaces weeks later.

Step 8: Measure whether the fix worked

The store’s total clicks won’t show the recovery, for the same reason they didn’t show the loss. The budget moved to other products, and after the fix it moves back, so the total barely changes. That’s why every fix gets two checks.

  1. The number of products disapproved for that cause, one to three days after the fix, once the crawler has revisited the pages. We count it as a pass when it falls below 2% of the count on the day of the review. At 410 products, that’s fewer than 8.
  2. The fixed products’ share of the whole store’s clicks in the first full week after the fix, compared with their share in the four weeks before the drop. In the video, the fix passes when the share returns to 11.5% for the landing page error and to 17.4% for the price mismatch.

A few rules keep the measurement honest:

  • Save the list of disapproved products on the day you read it. Without it, a week later nobody knows which products were fixed.
  • Start the “after” window on the day the fix reached the site and the feed. The date of the call with the developer doesn’t count.
  • A day-by-day click series shows the day of the failure and the day of the recovery. Weekly totals blur both.
  • Judge conversions and ROAS later, once the fixed products have collected conversions. After a disruption, Smart Bidding needs 50 to 90 new conversions to stabilize.
  • Don’t confirm an approval by reading the status a minute after the change. Merchant Center can show stale data right after an update, so check again later.

Step 9: Before Black Friday. What has to land in time

A fix to the feed or the site needs time before it shows up in ads. Count back from the sale:

  • a recrawl after a page fix takes one to three days, and images at a new URL take 24 to 72 hours,
  • upload sale prices with dates before the sale starts, and set the feed to fetch daily for its full run; if sale prices go in a separate supplemental feed, schedule fetches for it as well, because a primary feed fetch doesn’t refresh supplemental feeds,
  • for a sale that lasts a few days, set up promotion assets and new ads at least 48 hours ahead (ideally three to four days) with a scheduled start, because ad approval slows down at peak season.

A sale doesn’t call for rebuilding Shopping campaigns that work. You don’t add assets to a feed-only Performance Max campaign. The feed carries the sale price, and sitelinks to the sale pages and a promotion asset in your existing Search campaigns carry the promotional message. Adjust the target ROAS and budget of your always-on campaigns to how far prices actually drop, and spread the Black Week budget unevenly, with the most money on the last days, when most purchases come in. The strength of the offer decides what happens to the target ROAS:

  • for a strong promotion, lower or remove the target ROAS for its duration and raise the budget, then return to the previous target afterward. You can tag sale items with a label in the feed and run them in a separate campaign, excluded from their usual one;
  • for a weak promotion, leave structure and targets alone and support the sale with remarketing.

Don’t set the target ROAS for the weeks that follow on Black Friday data. ROAS during sale days is inflated, and a target set on it will choke the campaign once the sale ends. Set the new target only once the campaign has collected at least 50 conversions after the sale.

Check the feed every week, and in the weeks before Black Friday also after every change to the store’s site. Between reviews, a script that alerts you to new disapprovals helps. Test any ready-made script you find to make sure it still works.

Finally, split the work. In the video, the review ends with three tasks for three people: the developer lets Google’s crawlers through the bot check on the old URL path, whoever owns the feed adds sale prices with dates and sets a daily fetch, and the agency requests a feed re-fetch and checks the result a week later.

Rules at a glance

  • Read disapprovals in Merchant Center, separately for each destination. Take the count of products out of ads from the Shopping ads row, and check it in Google Ads too. Read “not eligible” in Google Ads together with its reason.
  • Clicks set the order of fixes: the weekly average before the drop against the last seven days. Suspensions and misrepresentation come first.
  • Out of stock and no campaign are not feed errors. Leave what works alone.
  • For a landing page error, open a sample of pages the way Google’s crawler does. Firewalls, slow servers and old URLs are the usual causes. Keep product IDs unchanged.
  • For a price mismatch: sale_price with dates and a daily fetch during the sale. Keep automatic item updates as a fallback. No login-only prices, no discount codes in place of the price, no promotions in titles.
  • Replace images at a new URL, starting with the products that sell.
  • After every change to robots.txt or the firewall, check that Googlebot and Storebot-Google can get in.
  • Measure every fix twice: the disapproved count after one to three days and the fixed products’ share of clicks after a week. The store’s total clicks won’t tell you anything here.
  • Count back from Black Friday: fixes to the feed and the site need a few days, and promotion assets and new ads need at least 48 hours.
  • Don’t set a target ROAS on sale data. Set a new one only after 50 conversions collected after the sale.
BDOS Club

Keep learning with free tutorials

BDOS Club is our free community on Circle. Join to get tutorials on using AI in Google Ads, from writing useful prompts to running full campaign workflows.

Join BDOS Club for free