Skip to content

Guide

How to Check if Your Meta Pixel and Conversions API Are Working

  • meta pixel checker
  • check meta pixel
  • facebook pixel test
  • conversions api
  • event_id
  • fbp cookie
  • fbc cookie
  • pixel helper

You open your store, the Meta Pixel Helper icon turns green, and you move on. A month later your Purchase numbers in Ads Manager do not match your orders, retargeting audiences look thin, and nobody can say why. The icon was telling the truth about one thing only: a pixel loaded on that page. It says nothing about whether the right events fire with the right data, whether your server events can be matched to the browser ones, or whether the cookies Meta relies on survive longer than a week.

This guide covers the five checks that actually decide whether your Meta tracking works, how to run each one by hand with free tools you already have, and how to run all of them at once with our free Meta Pixel and CAPI checker.

Why “Pixel Helper is green” is not enough

Pixel Helper inspects the page you are looking at, in your own browser, at that moment. That is useful, and it also leaves gaps:

  • It shows presence, not quality. A green icon with a ViewContent event that has no value or currency is still green, yet Meta cannot use that event to optimise for value or build a catalogue audience.
  • It cannot see your server. If you run the Conversions API, those events never pass through your browser, so Pixel Helper does not know they exist, let alone whether they carry the same event_id as the browser events.
  • It does not show cookie lifetimes. Your browser may be Chrome, where the _fbp cookie lives for 90 days. Your customers on Safari keep the same cookie for 7 days. Pixel Helper will never flag that.
  • It is easy to misread a warning. “Duplicate pixel” or “Pixel activated multiple times” warnings are often dismissed as noise, while they usually mean every visit is counted twice.

The 5 checks that matter

# Check What “working” looks like What breaks when it fails
1 The pixel loads once One pixel ID, one init, one PageView per page Every visit is counted twice, cost per result looks better than it is
2 PageView fires A request to facebook.com/tr with ev=PageView on every page No traffic signal, no website audiences
3 ViewContent carries product data ev=ViewContent on the product page with value, currency, content_ids, content_type No dynamic product ads, weak value optimisation
4 Events carry an event_id An eid parameter on every browser event Browser and server copies cannot be deduplicated
5 The _fbp and _fbc cookies exist and last _fbp after the first page, _fbc after a visit with fbclid Weaker matching, conversions after an ad click lose their link to that click

Each of these can be checked by hand. Here is how.

Check 1 and 2: the pixel loads once and sends PageView

Open your home page in a private window with DevTools open (right click, Inspect, then the Network tab). Type facebook.com/tr in the filter box and reload.

  • You should see one request with ev=PageView. Click it and look at the id parameter: that is your pixel ID.
  • If you see two PageView requests with the same id, the pixel is installed twice. The usual causes are a theme snippet plus a platform app, or a platform app plus a Google Tag Manager tag. Pick one installation and remove the other.
  • If you see two PageView requests with different IDs, you have two pixels. That can be intentional (an agency pixel, for example), but it is worth confirming someone still needs the second one.
  • If you see requests without ev=PageView, or no requests at all while Pixel Helper shows the pixel as loaded, the pixel may be initialised but held back by a consent banner. That is expected before consent. Accept the banner and reload before drawing conclusions.

In Pixel Helper, the same problem shows up as a warning that the pixel was activated more than once. Treat it as a real issue until you have confirmed otherwise.

Check 3: ViewContent carries value, currency and product IDs

Open any product page and keep the same facebook.com/tr filter. You should see a request with ev=ViewContent. Click it and check the payload:

  • cd[value] and cd[currency]: the product price and its ISO currency code (EUR, USD, PLN).
  • cd[content_ids]: the product or variant ID, in the same format your catalogue uses.
  • cd[content_type]: usually product (or product_group if you send parent IDs).

A ViewContent without these fields still counts as an event. It just cannot do its main job: telling Meta which product someone looked at, so that dynamic product ads and “viewed but did not buy” audiences have something to match. If the product ID in the event does not match the ID in your catalogue, the event is recorded but never joined to a product.

If there is no ViewContent at all, turn on the standard product view event in your platform’s Meta integration, or fire it from GTM on the product template.

Check 4: every event has an event_id

Click any of the facebook.com/tr requests from the previous checks and look for an eid parameter. It holds the eventID you passed to the pixel:

fbq('track', 'ViewContent', { value: 49.0, currency: 'EUR' }, { eventID: 'vc-8a1f2c' });

If you only run the browser pixel, a missing eid costs you nothing today. The moment you add the Conversions API, it becomes the most important field you send. Meta matches a browser event and a server event as the same action when they share the same event_name and the same event_id. Without it, one purchase reported by both the pixel and your server can be counted twice. We explain the mechanics, and the six ways it breaks, in Meta CAPI deduplication: how to test a pair.

Two things to look for beyond “is it present”:

  • The ID changes per event. If every event on a page carries the same eid, or the same ID is reused for PageView and ViewContent, deduplication can merge events that are not duplicates.
  • The server sends the same value. You cannot see the server call in DevTools. Use Events Manager, Test Events, and trigger an action: when both copies arrive and match, Meta shows the event as deduplicated. You can also paste both payloads into our free CAPI deduplication checker, which compares them in your browser and uploads nothing.

Check 5: the _fbp and _fbc cookies

In DevTools, open the Application tab (Storage in Firefox and Safari), then Cookies, then your domain.

  • _fbp should appear after the first page view. Its value looks like fb.1.1727600000000.1234567890: a version, a subdomain index, a timestamp and a random number. It identifies the browser across visits and is one of the keys Meta uses to match your events to a person, which lifts your Event Match Quality.
  • _fbc should appear when the visitor lands from a Meta ad. To test it, open your store with a fake click ID in the address bar, for example https://yourstore.com/?fbclid=test123. After the page loads, _fbc should hold a value ending in test123. If it does not appear, the usual causes are a redirect that drops the query string (for example to a country or language version of the store) or a consent banner that blocks the pixel until consent, which is expected.

Now look at the Expires column. In Chrome, _fbp is set for about 90 days. But the pixel writes it with JavaScript, and Safari limits cookies written by JavaScript to 7 days, and to 24 hours when the visit comes from an ad click with tracking parameters. A Safari customer who clicks your ad on Monday and buys the following Wednesday arrives with a fresh _fbp and no _fbc, so Meta sees a stranger. Chrome will never show you this, which is why our checker computes the Safari lifetime for every cookie. The full story is in Safari deletes your tracking cookies after 7 days.

To pass cookie values to your server events, read _fbp and _fbc on your server and send them in user_data.fbp and user_data.fbc. That keeps the browser and server copies of the same visitor linked. If you would rather not build that yourself, the AxiTrace Conversion API sends each order with fbp, fbc and a matching event_id.

How to check all of this in 40 seconds

Running these checks by hand takes about half an hour per store, and you need to know where to look. Our free Meta Pixel and CAPI checker runs them in one pass:

  1. It opens your home page in a real Chromium browser, finds a product page on its own, and loads it too.
  2. It records every Meta request, decodes the payload (event names, eid, value, currency, content IDs) and counts pixel initialisations.
  3. It opens your store with a synthetic fbclid to check whether _fbc is stored, then computes how long Safari would keep each cookie.
  4. It stops every tracking request before it leaves the scanner, so the scan never adds visits or events to your Meta account.

You get a score, the issues sorted into Fix today and Improve, an estimated fix time for each one, and the steps to fix it. No signup is needed to see the summary. For a wider view that includes GA4, TikTok, consent behaviour and ad-blocker exposure, use the website tracking checker.

What any browser-based checker cannot see

Being honest about limits saves you a wrong conclusion:

  • Checkout and Purchase. No automated checker should place orders in your store. Purchase events have to be checked with a real test order and Events Manager, Test Events.
  • Your server events. A scanner sees what the browser sends. If your platform already sends Conversions API events from its server, the scanner cannot see them, and a “no server-side signal” finding can be ignored once you have confirmed it in Events Manager.
  • Logged-in data. Advanced matching fields such as a hashed email appear only after a customer logs in or checks out, so their absence on an anonymous page is normal.
  • Your ad account settings. Aggregated event measurement, domain verification and attribution settings live in Meta, not on your site.

What to fix first

If the checks turn up several problems, fix them in this order, because each one affects the ones after it:

  1. Duplicate pixel. Everything else is measured wrong until PageView is counted once.
  2. Missing or incomplete ViewContent. Product audiences and dynamic ads depend on it.
  3. Missing event_id. Required before you turn on the Conversions API, or your conversions will double.
  4. Cookie lifetime. The longest fix, and the one that matters most for Safari-heavy traffic. See why server-side tracking matters for the bigger picture.

For the matching side of the story, our guide on improving Event Match Quality covers which customer fields to send and how to hash them.

Frequently asked questions

Does Pixel Helper work on mobile?

No. Pixel Helper is a desktop Chrome extension. To check mobile behaviour, use Events Manager, Test Events with a test code, or open your store on your phone and watch the events arrive in Test Events.

Why does Events Manager show more PageViews than my analytics?

The most common cause is a pixel installed twice. Other causes are single-page apps that fire PageView on every route change, and bots that analytics tools filter out while Meta does not. Check 1 above rules out the first.

Will running your checker add fake events to my Meta account?

No. The scanner blocks every request to Meta, Google, TikTok and other collection endpoints before it leaves the browser. It uses a synthetic fbclid to test cookie storage, and that value is never sent to Meta. If your store sends events from its own server when a page loads, your server may still record the visit; the scanner identifies itself with the header X-AxiTrace-Scanner: 1 so you can filter it. Details are on how our scanner works.

Is the Conversions API still worth it if my pixel is working?

Yes, if you care about accurate numbers. A working pixel still loses the events that ad blockers, Safari and slow pages remove before they reach Meta. Meta’s Conversions API sends the same events from your server, and with matching event_id values Meta keeps one copy of each.

Written by

Chris Gaudy

Founder & CEO

Building accurate conversion tracking for e-commerce teams.

Get tracking insights in your inbox

Tips on server-side tracking, conversion optimization, and product updates. Delivered weekly.

No spam. Unsubscribe anytime.