Facebook Ads Conversion Tracking Problems: Why a ‘Healthy’ Pixel Still Misses Conversions

Last updated: 28 September 2026

Your Facebook Pixel is installed. Events Manager shows it firing. No errors. You check your Ads Manager and conversions look roughly credible.

Then you check your CRM. The numbers are nowhere close.

A Facebook Pixel can look perfectly healthy and still miss a third of your conversions. Below: how to tell whether yours has a real problem, how to read your Event Match Quality score, why you need the Conversions API as well as the Pixel, and what the two 2026 attribution changes did to your numbers. Server-side tracking — which we set up for clients — is what closes the gap.

A healthy Pixel does not mean complete conversion tracking. These are two different things.

The Pixel fires for the sessions it can see. iOS restrictions, ad blockers, and cross-domain checkout flows mean a significant portion of your actual customers are invisible to it — even when every status indicator says it is working correctly.

If you’re here because your Facebook and Google Ads numbers simply don’t agree with each other — rather than a genuine shortfall against your CRM — that’s a structural difference, not a tracking failure.

Why Meta shows fewer conversions than you actually got (and when that’s normal)

Some gap between Meta and your real sales is normal. Meta only counts conversions it can link to someone who clicked or saw your ad within your attribution setting, and iOS privacy settings, browsers and cookie consent hide some of those people from it. A gap is a problem when it’s large, sudden or growing.

Why Meta under-reports:

• Attribution setting. Meta only credits conversions inside the window you’ve chosen. Since March 2026, only link clicks count for click-through attribution; other engagement moved to engage-through attribution (see the 2026 changes below).
• iOS and browser restrictions. Opted-out iPhone users, Safari’s tracking prevention and ad blockers stop the Pixel seeing some visits and purchases.
• Consent. Visitors who decline your cookie banner shouldn’t be tracked by the Pixel.
• Pixel only, no Conversions API. Without a server-side copy of each event, anything the browser blocks is lost for good.
• Deduplication. If the Pixel and the Conversions API send the same event without a shared event ID, Meta can’t tell them apart. That usually inflates the count, so a sudden jump is as much a warning sign as a drop.
• Cross-device journeys. Someone who clicks on their phone and buys on a laptop is only linked if Meta can match them as the same person. Sending hashed email and phone through the Conversions API helps.

Is my gap normal or is tracking broken?

1. Compare like with like. Take your CRM, form or call records for the same dates, and count only the enquiries or sales you can reasonably put down to Meta.
2. Look at the trend, not one week. A steady gap is usually attribution and privacy. A gap that opens suddenly, especially after a site, checkout or cookie-banner change, is usually broken tracking.
3. Check Events Manager. Open your dataset’s Diagnostics tab for active issues, and the Overview to see whether events are arriving from both the browser and the server.
4. Check Event Match Quality. A low score means Meta is struggling to match conversions to people. Look at the parameter breakdown for what’s missing rather than chasing a single number. See our guide to improving Event Match Quality with CAPI.

If Meta showing half your real conversions is new, treat it as broken and use the checklist below. If it has always been that way, the usual cause is Pixel-only tracking, and server-side tracking is the fix.

Meta conversion tracking suddenly stopped: a 10-minute checklist

When Meta conversions drop to zero overnight, something changed on your site or in its connection to Meta. Work through these seven checks in order. Each one takes a minute or two in Events Manager or your browser.

1. Is the Pixel firing? Open your thank-you page with the Meta Ads Data Advisor Chrome extension (formerly Meta Pixel Helper), or use the Test Events tab in Events Manager and complete a test conversion. No Purchase or Lead event means the Pixel code or the event has gone from that page.
2. Is the Conversions API still sending? Check your dataset’s Overview for server events. If they’ve stopped, check whatever sends them: a partner integration, the Conversions API Gateway or your own server. On a manual setup, an access token that has been regenerated or removed in Settings stops events arriving.
3. Is your domain still verified? Business Settings > Brand Safety > Domains. Check this especially after a domain move or a DNS change.
4. Does the event you optimise for still fire? If the ad set optimises for Purchase or Lead, that event has to arrive. A custom conversion built on a URL rule fails silently when the thank-you page URL changes.
5. What changed on the site? A theme update, a new checkout, a page-builder change or a new form plugin is how Pixel code or thank-you pages usually disappear. Compare the date conversions stopped with your change log.
6. Did the cookie banner change? A new banner or new settings can stop the Pixel loading until a visitor accepts, or stop it loading at all.
7. Is a partner integration disconnected? Shopify, WooCommerce plugins and booking platforms connect through their own Meta integration. Check it’s still connected, and to the right dataset rather than an old Pixel.

How to fix missing Facebook ad conversions

To fix missing Facebook ad conversions, make sure the Pixel fires where the conversion happens, add the Conversions API to catch what the browser misses, deduplicate the two with a shared event ID, and check your attribution setting before you compare numbers.

1. Fix the Pixel. The conversion event must fire on the confirmation page or on form submission, once per conversion. Confirm it in Test Events.
2. Add the Conversions API. Send the same events from your server, through a partner integration, the Conversions API Gateway or a manual setup (your dataset’s Settings tab in Events Manager). This recovers conversions that ad blockers and browser restrictions hide.
3. Deduplicate. Send the same event_id and event_name from the server as the Pixel sends as eventID. Meta only deduplicates events received within 48 hours of the first event with that ID.
4. Send customer details with server events. Hashed email and phone number raise Event Match Quality and let Meta attribute more of your conversions.
5. Verify your domain in Business Settings > Brand Safety > Domains.
6. Check the event setup. The ad set’s optimisation event must be the one that fires. Rebuild any URL-based custom conversion after a thank-you page URL changes.
7. Check attribution before you compare. Look at the ad set’s attribution setting and your report columns. Since 2026, only link clicks count as clicks, and the 7-day and 28-day view windows are no longer returned by the Ads Insights API, so older benchmarks overstate what you should expect.

How can you tell if Facebook has a tracking problem?

Compare Facebook reported conversions with your CRM sales for the same period, counting only the sales you can reasonably put down to Meta. There is no universal healthy ratio: it depends on your attribution setting, your sales cycle and how much of your business comes from other channels. Zero conversions reported means the Pixel is broken or no event is set up for the conversion you track.

What matters more than any single ratio is change. Set a baseline from a period when you trust the tracking, then watch for a gap that suddenly widens. As a rough guide from our own accounts, Meta reporting well under half of the sales you can trace to it, with no attribution change to explain it, usually points to a Pixel or Conversions API fault rather than normal under-reporting.

Note: healthy Facebook ratios shifted lower in 2026 following two platform changes (covered below). If you’re benchmarking against pre-2026 data, establish a fresh post-March 2026 baseline before diagnosing.

Before troubleshooting anything else, check your Event Match Quality score.

How do you check your Event Match Quality score?

Event Match Quality (EMQ) is Facebook’s score for how accurately it can match your conversion events to real users on the platform. The lower the score, the less reliably Facebook can identify who converted, which affects lookalike audience quality, optimisation and the value of your conversion data. Find it in Meta Events Manager under your Pixel’s Overview tab.

Read the score as a direction, not a pass or fail. It varies by event type: a Purchase event that carries an email address will usually score higher than a PageView that can’t. The useful questions are whether your key conversion events are improving, and which customer parameters Events Manager says you could be sending but aren’t. A score that is low for your Lead or Purchase event is worth fixing whatever the exact number.

What lowers EMQ:

• Missing customer identifiers in the event payload — email, phone, name, city, country
• Running Pixel only, without CAPI (browser Pixel sends limited identity data)
• Events not deduplicating correctly

The fix: ensure CAPI is sending all available first-party identifiers alongside each purchase or lead event. Each additional identifier meaningfully improves the score.

Most accounts have never opened the EMQ breakdown, and it is one of the quickest places to find recoverable conversions.

Pixel or Conversions API — why do you need both?

The Facebook Pixel tracks browser-level events. The Conversions API (CAPI) sends conversion events server-to-server, so ad blockers and browser restrictions can’t stop them in transit. Running both together, with proper deduplication, maximises event coverage and EMQ. The Pixel handles upper-funnel signals; CAPI makes purchase and lead events far more resilient to ad blockers and browser limits. It doesn’t remove the need for consent: if a visitor declines tracking, you shouldn’t send their data through either route.

In plain terms:

• Pixel: JavaScript running in the user’s browser. Ad blockers stop it. iOS privacy restrictions limit it. How much it misses varies widely by audience, device mix and checkout setup.
• CAPI: Your server sends the event directly to Facebook’s API, so browser blockers don’t affect it. It can send every transaction your server processes, subject to your consent rules. Meta still has to match each event to a person before it can attribute it to an ad.

Running both requires deduplication. If both the Pixel and CAPI send a Purchase event for the same transaction, Facebook counts them twice unless you configure deduplication. Both events must include an identical event_id for the same transaction. Configure this in Meta Events Manager and verify the Deduplication column shows activity.

For service businesses and clinics using third-party booking platforms — where the booking confirmation page is on a different domain — the browser Pixel frequently misses conversions entirely because the Pixel code is not present on the third-party confirmation page. CAPI, configured to pull purchase or booking events from your backend, catches those regardless of where the booking completed.

What did iOS 14.5 take — and how does CAPI recover it?

Apple’s iOS 14.5 App Tracking Transparency (April 2021) removed the cross-app tracking identifier Facebook used to link ad impressions to post-click conversions on iPhone. Most iPhone users opted out. Facebook lost visibility into a large portion of iOS-originated conversions. CAPI recovers this by matching server-sent customer data (hashed email and phone) to Facebook account identity, bypassing the need for the blocked identifier.

In business terms: before April 2021, Facebook could see that an iPhone user who clicked your ad then purchased on your website. After iOS 14.5, for users who opted out, Facebook often cannot connect the ad click to the purchase. That conversion is invisible to the algorithm — affecting both reporting and campaign optimisation.

This gap is still present in 2026 for any account not running CAPI. It is not a temporary problem that will resolve itself.

Aggregated Event Measurement (AEM): Facebook’s framework for handling iOS conversions. Meta has been simplifying it: from 2025 it began removing the old requirement to rank eight priority web events, and many accounts no longer have the Aggregated Event Measurement tab at all. If your Events Manager still shows it, keep your most valuable conversion events at the top. See Facebook’s AEM settings in order of business value. If not configured, Facebook’s ability to attribute iOS conversions is further reduced. Find this in Events Manager > Aggregated Event Measurement.

What are the two 2026 Facebook attribution changes — and what do they mean for your numbers?

Two separate changes hit Facebook accounts in 2026. Both reduced reported conversions. Neither indicates a decline in actual campaign performance. Check your CRM — if actual customer acquisition is stable, do not cut budgets or pause campaigns in response to lower Ads Manager figures.

January 12, 2026 — Attribution window removal

Facebook permanently removed the 7-day view and 28-day view attribution windows from the Ads Insights API. Only 1-day view remains for view-based attribution. Advertisers running awareness campaigns, video campaigns, or any campaigns where users typically convert more than 1 day after seeing (but not clicking) an ad saw reported conversions drop overnight.

If your Facebook numbers dropped sharply around January 12, 2026, this is the cause.

March 3, 2026 — Click definition change

As confirmed in Facebook’s official announcement, Facebook changed its click attribution definition. Previously, any interaction with an ad — likes, reactions, shares, saves — qualified as a click for click-through attribution. From March 2026, only link clicks that bring a user to your site count.

Non-link interactions moved to a new category called engage-through attribution with a 1-day window. The critical detail: the previous engagement-based click-through attribution used a 7-day window. Engage-through only has 1 day. This means conversions where a user engaged with your ad (without clicking a link) and then converted on days 2–7 now fall outside any attribution window entirely — they did not simply move to engage-through, they left Facebook’s attribution entirely.

What this means for your reporting: click-through conversions dropped. A new engage-through column appeared. Some conversions that previously appeared under click-through are genuinely gone from Facebook’s reporting — not reclassified.

What this means for your business: actual customer acquisition is unchanged. Check CRM.

An event firing is not the same as a conversion attributed to an ad

Events Manager tells you whether Meta received an event. Ads Manager tells you whether Meta credited that event to one of your ads. An event can arrive perfectly and still never appear as an ad conversion, and that is often correct.

Two examples show the difference:

• Fired, not attributed. Someone finds you on Google, fills in your form, and the Lead event arrives in Events Manager. They never saw or clicked a Meta ad, so Ads Manager rightly shows nothing. Nothing is broken.
• Should be attributed, isn’t. Someone clicks your Meta ad on their phone, books a call on the same visit, and the Lead event arrives, but Ads Manager shows no conversion. Now check the ad set’s attribution setting, whether the optimisation event matches the event that fired, and whether the event carries enough customer information for Meta to match it to the person who clicked.

So when numbers look low, check both screens. If events aren’t arriving, it’s a tracking problem: use the checklist above. If they arrive but aren’t credited, it’s usually attribution or matching, covered in the sections below.

How do you diagnose tracking with Test Events in Meta Events Manager?

Facebook’s Test Events tool lets you trace live conversion events in real time — you can see whether the event fires, what data is being sent, and whether Facebook can match it to a user. Use this before assuming your tracking is broken.

How to use it:

1. Meta Events Manager → your Pixel → Test Events
2. Enter your live site URL
3. Complete the conversion action you want to test (submit a form, complete a test booking, or add to cart)
4. Watch the event appear in the right panel — confirm the Purchase or Lead event fires on the correct page

Also install the Meta Ads Data Advisor Chrome extension (formerly Meta Pixel Helper) to confirm browser-side Pixel firing. Note: Data Advisor only confirms the browser Pixel, not CAPI. Use both tools together.

Common failure patterns Test Events reveals:

• Purchase event fires on the add-to-cart page instead of the order confirmation page
• No Purchase event fires at all — Pixel code missing from thank-you page
• Third-party checkout (PayPal Express, Klarna hosted checkout, booking platform) redirects to its own confirmation page — Pixel never fires

What causes zero conversions in Facebook?

1. Pixel code missing from the thank-you page — Events Manager may show the Pixel as active (because it fires on other pages) but it has never seen the confirmation URL
2. Third-party checkout bypasses the website — PayPal Express, hosted payment pages, and third-party booking platforms redirect users to their own confirmation screen, not yours
3. Purchase event fires too early — triggered on add-to-cart or checkout initiation, not on purchase confirmation
4. CAPI not configured — Pixel alone is missing iOS conversions by default; without CAPI there is no server-side recovery
5. Custom conversion rule broken — URL-based custom conversion set up with a rule that no longer matches the actual thank-you URL after a site update

Want us to fix your Meta pixel and attribution?

Frequently asked questions

Bons & Frazer runs Meta Ads management for UK businesses on a fixed monthly fee, with server-side tracking built in. You deal directly with the two founders, with no contracts and no account managers. Based in Norfolk? See Meta Ads in Norwich.