Guide · Updated · 12 min read

Meta restricted a health store. Here is what happened, and the plan.

A real store, five weeks, one rejected review. What went wrong, what we found, and what to do on day one.

The short answer

A Shopify maternity store lost its Meta conversion tracking in five weeks. Meta warned, then blocked standard events, then rejected a review on the same day and added a block for Europe. The cause was not one thing. It was the data Shopify sent, the words on the site, and a review requested too early. This is what happened, what we found, and the plan that gives a store like this the best chance.

DataCops is a tool that stops a health or wellness store sending Meta the detail that gets it restricted, and sends a clean, neutral signal from the server instead.

How DataCops helps a restricted store:

  • Health mode. Every conversion goes through a fixed list of allowed fields. Paths, query strings, referrer, page titles, product names, cart items and form answers never leave the site.
  • Neutral events. A purchase goes to Meta as P_1, other steps as E_1. The name says nothing about what was bought.
  • Server-side, with matching. Click ID and hashed email and phone still go, so Meta can match the sale without knowing the product.
  • Proof. The delivery log marks every cleaned send, so you can show what left your store.
  • Your own count. The click ID is kept on the server for up to 90 days, so you can count Meta-sourced orders even when Meta cannot.
  • Ads Warmup. Upload past customers (up to 20,000 rows) and send them to Meta, Google Ads and TikTok to give new campaigns something to learn from.

Best for: health, wellness, supplement, telehealth, clinic and maternity stores on Meta, and for any store that has just had a warning.

What happened, day by day

The store sells maternity clothing on Shopify. Most of its buyers are in the UK and Northern Europe. It ran Meta ads and leaned on them. We changed the details so it cannot be identified. The order of events is real.

DayWhat happened
Day 0Meta sends a warning that the site may fall in a restricted category. Nothing is changed.
Day 31Meta blocks standard events. Purchase and Add to cart stop arriving. Ads keep running with no conversion signal.
Day 35A review is requested in Events Manager. It is rejected the same day, and Meta adds a block for visitors in Europe.

That is the part nobody tells store owners. The warning feels like noise. The block comes weeks later. The review is a single button, and pressing it too early can make things worse.

The owner was also ill during part of that month. That matters, because this is the usual shape of it: a small team, a long gap, then a sudden stop. If the warning had been acted on at Day 0, the story would likely be different.

Why a clothing store counts as health

Meta put this store in a "health and wellness condition" category. For Meta, pregnancy is a health condition. So a store that sells maternity dresses sits next to stores that sell pregnancy aids.

Meta's own help page lists the kinds of sensitive health information it does not want sent: conditions, sexual and reproductive health, mental health, medical devices, procedures, medications and supplements, and places that point to treatment. Reproductive health is on that list.

The lesson is wider than maternity. If your store touches a body, a condition, a treatment or a life stage, assume you are exposed. Supplements, weight loss, braces, skincare for conditions, fertility, mental health and telehealth are all in the same position.

What we found on the store

We went through the events, the store and the Shopify settings. Six things were feeding Meta health detail while the review was pending.

  1. Shopify's Meta channel sent everything. Every customer event carried the product name, the product link, email, phone and name. That is the detail Meta reads as health information, and it arrived on every event.
  2. Advanced matching sent gender and date of birth. For a maternity store, those two fields make the link to pregnancy obvious.
  3. Product titles carried the words. Titles used "pregnancy" and "postpartum" throughout.
  4. The meta description carried them too. It is the text Meta's crawler reads first.
  5. One claim sentence. A line about easing pressure on the lower back, hips and abdomen. That is a health claim, whatever the product is.
  6. The alt text. This is the one the owner never saw. The photos were fine. But the hidden accessibility text behind them held the full product descriptions. The home page alone had 75 images whose alt text said "pregnant" or "postpartum".

Vendor reports say Meta's classifier crawls site content. Alt text is site content. The photos did not need to change. Only the hidden text behind them did.

Why the review failed the same day

While Meta was reviewing the site, it was also receiving the evidence against it. Every event still carried the product name and link. Every page still carried the words. The review had nothing clean to look at.

Vendor reports describe the review as automated: a single button, a few days for a decision, no way to attach evidence, and a 30 day wait after a rejection. We cannot verify those numbers from Meta. What we saw matches the shape: no human, no conversation, a fast no.

The rule that follows is simple. Clean first, then ask once. A review pressed before the clean-up is a wasted attempt.

The three levels, and where this store sat

LevelWhat Meta doesCan neutral server events work?
Core setupStrips URL detail and custom parametersYes
Standard eventsCan block Lead, Add to cart and PurchaseNeutral custom events are the route to try
FullCan block all event sharingNo. No tool gets through

This store was at the middle level for most of the world, and Meta's notice said Europe could be at the full level. Most of its buyers were in Europe, so that is where it hurt.

The levels are explained in full in Meta health and wellness restrictions, explained.

That is the honest map. Outside the blocked region, a clean neutral signal can still reach Meta. Inside it, nothing can.

Best case, real case, worst case

An owner in this spot needs to see all three, not only the hopeful one.

Best case

You clean Shopify's data sharing, the matching fields, the titles, the description, the claim and the alt text. You wait for four weeks of clean data. You request one review. Meta lifts the category or relaxes the level. Standard events return. This can happen, and it is what to aim for. Nobody can promise it.

Real case

Standard events stay blocked. Neutral events from the server reach Meta for visitors outside the blocked region. In the blocked region, you run ads on broad targeting with your creative doing the work, and you count sales yourself. Revenue continues, with less signal and a higher cost to learn. This is where most stores live for a while.

Worst case

A second review is rejected. The level stays at full for the region where your buyers are. Meta can no longer learn from your sales there. If Meta was most of your sales, the store is in danger of shutting. The test is blunt: if Meta ads are the main way new customers find you, a full block is a business problem, not an ads problem.

That is why the first weeks matter more than any trick later. And why a store should never have only one channel.

The plan: today, this week, next month

Today

  1. Stop Shopify's Meta data sharing. It is the source of the product names, links and customer details.
  2. Turn off gender and date of birth in advanced matching. Two clicks.
  3. Do not request a review.
  4. Find your level. In Events Manager settings, look at which category and level your data source shows.

This week

  1. Make the word list. Product titles, the meta description, headings, claims and image alt text. Put the old text on the left and the new text on the right.
  2. Change the text, not the brand. Photos, products, testimonials and design stay. Only the words Meta reads change.
  3. Turn on Health mode and send neutral events from the server. In DataCops, Health mode sends P_1 for the order and E_1 for the other steps, with hashed email and phone, value and currency, and nothing about the product.
  4. Create a custom conversion in Meta from the neutral order event, and point the campaign at it, where the level allows it.
  5. Run ads again. Where there is no conversion signal, optimize for traffic or engagement and let creative carry it.

Next month

  1. Check the delivery log. It shows which sends reached Meta and which were held.
  2. Request one review, with about four weeks of clean data behind it.
  3. Plan for no. Decide in advance what you will do if it is rejected.

Where Meta reads your words

Run this checklist on any health-adjacent store.

  • Product titles and collection names
  • Page titles and the meta description
  • Headings and body text, including claims
  • Image alt text, including text added by apps or themes
  • URL paths and handles
  • Custom audience names and custom conversion names
  • Event names and event parameters
  • Form fields and what is typed into them

The DataCops restriction checker helps find the words before Meta does.

Tracking while Meta is blind

A block on Meta's side does not stop you counting your own sales.

  • Your store records every order. Shopify is the source of truth.
  • The click is kept. DataCops stores the Meta click ID on the server, so an order can be tied to a Meta ad even if Meta never sees the sale.
  • Judge on your numbers. Compare revenue by source from your store, not only Meta's report.
  • Test outside Meta. A geo test compares places with and without ads.
  • Build your own audience. Ads Warmup uploads past customers (with their consent) so new campaigns start with real buyers.

The last resort, and what never to do

If a second review fails and the level stays at full, there is one option some stores take: a separate, real store on a new domain, with clean content, used for Meta traffic.

Be careful. This is not a fix. It is a new start, and it has real risks.

  • It has to be a real store, with real products and pages Meta can read.
  • The content must pass the same word checks before launch.
  • It needs its own pixel and dataset.
  • Meta can still link two domains through the same ad account, business account or payment method. We cannot promise it works.

Never do these:

  • Redirect between a restricted domain and a new one. Meta sees one page and the customer sees another. That is the practice most likely to get an ad account closed.
  • Show Meta a clean page and customers a different one.
  • Hide the health pages so the crawler cannot read them.

Keep your existing domain for organic search, Instagram, email and returning customers. It still works for them.

What we would tell any owner on day one

  1. The warning is the deadline. Act at Day 0.
  2. Find where the words are. Titles, description, claims, alt text.
  3. Stop sending health detail. Shopify sharing and matching fields first.
  4. Send less, from the server, with neutral names.
  5. Ask for review once, after the clean-up.
  6. Never depend on one channel. Build organic and email while you can.
  7. Keep your own count. Your store data beats any platform report during a block.

What else the same install does

  • Real people only. Every visit gets a bot verdict, with a Real people only switch per ad platform (off by default), so junk traffic does not feed a weak signal.
  • First-party collection. One script and one DNS record put collection on your own domain.
  • Edge capture. With your DNS on Cloudflare, the free Worker reads the click at the edge. It captures; it does not block.
  • The click is kept. Click IDs stay on the server for up to 90 days, and a signed server-set cookie lasts up to 400 days where enabled.
  • The sale after the form. Booked, showed and won stages from a CRM go back to your ads, for clinics and services that close by phone or in a consultation.
  • Consent. A first-party consent manager on IAB TCF v2.2 from the same script.

What we cannot do

  • We cannot lift a restriction. Meta decides, and it does not explain or negotiate.
  • We cannot make Meta count sales in a region that is fully blocked.
  • We cannot promise a review will pass.
  • This is not legal advice. Check your own rules.
  • The ending of this case is not written yet. The clean-up is under way, and we will update this page with what Meta does, good or bad.

Sources

FAQ

Can a normal clothing or accessories store be restricted as health?

Yes. Meta classifies the data source by what it reads on the site and in the events. A maternity store, a supplement shop or a brace seller can all fall under a health condition category even when the owner never thought of the store as health.

What should I do first if Meta warns me?

Do not request a review yet. First stop the sources of health detail: Shopify's Meta data sharing, advanced matching fields such as gender and date of birth, product titles, the meta description, health claims and image alt text. Review once, after the clean-up.

Is a review always rejected?

No. But one vendor that reviewed 125+ ad accounts reports a review is automated, takes a few days, and that no genuine health seller in its sample got the category removed. In this case the review was rejected the same day, with the store still full of health words.

What if Meta blocks all events?

At the full level no tool can send conversion events from that domain. You can still run ads, using traffic and engagement goals, and count sales yourself from your own store data and click IDs.

Should I move to a new domain?

Only as a last resort and only as a real, separate store with clean content. Never redirect between a restricted domain and a new one, and never show Meta a different page than the customer sees. We cannot promise it works, and Meta can still link the two.

Can DataCops remove a Meta restriction?

No. Meta decides. DataCops can stop your store from sending health detail to Meta, send a clean neutral signal from the server, and show what was sent so you can prove the clean-up.

Is this a real customer?

Yes. The details are changed enough that the store cannot be identified, and nothing here is an invented result. The ending is not written yet, and this page says so.

Get ahead of the warning

Send Meta a clean, neutral signal and keep your own count.

Setup in 5 minutes. No credit card.

Live traffic quality

Updated just now

Visits · last 24h

487
Real users
35873.5%
Bots · auto-filtered
12926.5%

Without filtering, 26.5% of your reported traffic is bot noise inflating dashboards and draining ad spend.

Don't trust your analytics!

Make confident, data-driven decisions withactionable ad spend insights.

Setup in 2 minutes
No credit card