First, confirm the events are really missing
Plenty of "missing" events are present and simply out of view. Events Manager shows one dataset at a time, so check that the one open is the one the site uses; businesses with several portfolios often look in the wrong place. Widen the date range to include recent full days, since the overview can lag behind live activity, and switch from the summary to the per-event view so that a single absent event, such as the purchase, is not hidden among page views that do arrive.
Then test deliberately rather than browsing. Open the Test Events view, enter the site's address, and load the pages that should fire each event in a browser profile with no extensions. Meta's Pixel Helper extension reports what fired on each page and flags errors. If you deploy through Google Tag Manager, run the container in preview at the same time: the guide to Tag Manager debugging and preview explains how to read which tags fired on which trigger. With those three views open, each cause below can be confirmed or ruled out in minutes.
Diagnosis table
Diagnosis
| Symptom | Check | Likely cause | Fix |
|---|---|---|---|
| Events arrive from the homepage but not from product, checkout or confirmation pages | Pixel Helper on one page of each template | Base code missing from some templates or a separate checkout domain | Add the base code to every template, including hosted checkout pages where allowed |
| Pixel Helper shows a request, yet the dataset stays empty | Identifier in the request against the dataset ID in Events Manager | Wrong pixel or dataset ID | Replace the identifier in the tag or plugin with the correct one |
| Events fire in Tag Manager preview but never for real visitors | Container version history; trigger conditions in preview | Trigger not firing for real traffic or container never published | Correct the trigger, then publish the container |
| Events appear only after someone clicks accept | Test with consent declined, then accepted | Consent banner holding the pixel until consent | Expected behaviour; make sure the accepted branch fires and plan a server route |
| Event counts sit well below sessions in analytics | Same pages in a clean browser versus one with a blocker | Ad blockers and browser restrictions dropping browser events | Add the Conversions API so blocked actions still reach Meta |
| Events arrive with no value, or the purchase event is absent while page views arrive | Order of data layer push and tag in preview | Event code running before the page or data is ready | Fire the tag on the data layer event that carries the data |
Fix, verify and prevent each cause
Base code missing from some templates
Sites are rarely one template. A theme update, a new landing-page builder, or a checkout hosted on a separate domain can each leave a set of pages without the base code, so events from those pages never start.
Fix: add the base code, or the Tag Manager container that loads it, to every template that should report, and check whether the checkout platform offers its own pixel field.
Verify: Pixel Helper reports the pixel once on a page from each template, and Test Events receives the expected event from each.
Prevent: keep a list of templates and the events each should send, and re-run it after every theme or platform change. The Meta Pixel setup tutorial has a verification list that covers every page type.
Wrong pixel or dataset ID
An agency's test pixel, an old dataset from a previous owner, or a typo in a plugin field all send events successfully, just to a destination nobody is watching.
Fix: copy the dataset ID from Events Manager in the business portfolio that owns it and paste it into the tag or plugin.
Verify: the identifier in the network request matches, and Test Events in the right dataset records the visit.
Prevent: store the identifier as a single Tag Manager constant rather than typing it into each tag, and record which portfolio owns the dataset, following the ownership advice in the pixel setup tutorial.
Tag Manager trigger not firing or container unpublished
A trigger built on a URL that changed, a click class that a redesign renamed, or a custom event that the developer named differently will never fire. Equally common: the tag works in preview, but the version containing it was never published.
Fix: correct the trigger condition against what preview shows, then publish.
Verify: outside preview, the event arrives in Test Events.
Prevent: name versions with what changed and who approved it, and review triggers after site releases, following the patterns in the guide to tags, triggers and variables.
Consent banner holding the pixel
Where a consent banner gates marketing tags, the pixel stays silent for visitors who decline or who never answer. This is the intended behaviour, not a fault, but it explains a gap between sessions and events.
Fix: confirm that the accepted branch fires every planned event and that the declined branch fires none; anything else is a configuration error.
Verify: test both branches in a fresh session and compare what Test Events receives.
Prevent: treat consent as part of the tag specification, as the privacy and consent guide recommends, and retest whenever the banner vendor or its categories change.
Blockers and browser restrictions
Extensions and some browsers block requests to known advertising domains, so a share of real actions never produces a browser event at all. You cannot remove this from the browser side.
Fix: add the Conversions APIServer-side event sharing with Meta. so the server reports the same actions, deduplicated against the pixel.
Verify: the dataset overview lists the key events with both browser and server connections, and the server count sits closer to the site's own records.
Prevent: build the server route into every new measurement plan rather than adding it after a gap appears; the Conversions API guide explains how the two routes share the work.
Event firing before the data exists
On many sites the order value, product IDs and currency arrive in the data layer a moment after the page loads. A tag that fires on page view reads empty variables, sends an event with nothing in it, or fails its own condition and sends nothing.
Fix: fire the tag on the custom event that the data layer push announces, and read values from data layer variables.
Verify: preview shows the push before the tag, and Test Events shows the value and currency populated.
Prevent: agree the event names and fields with developers in writing, using the contract described in the data layer guide.
A prevention routine: test events after every site release
Most missing-event problems start on a specific day: a theme update, a checkout migration, a new consent banner, a container publish. A short routine after each release finds them before a week of optimisation runs on bad data. Keep a test script listing each event, the page that fires it, and the parameters it should carry. After any release, run the script in a clean browser with consent accepted and again with it declined, reading Test Events and the Pixel Helper side by side. Check the dataset's diagnostics tab the next morning for new warnings. Once a week, compare purchase counts in Events Manager with orders in the shop over a closed period, remembering that AttributionAssigning conversion credit to touchpoints. windows mean the two will not match exactly. Put marketing on the site's release list so that the person who owns the tags hears about changes before they ship, not after the numbers fall. If you also see conversions falling on a setup that used to work, the guide to troubleshooting performance drops covers measurement breaks in that wider context. When the release is a new shop platform or a new domain, the tutorial on moving a website without breaking Meta Ads plans the events, server route, URL rules and catalog before launch instead of repairing them afterwards. If the browser events now arrive but the server copies do not, the diagnosis for Conversions API events that are not showing runs the same kind of check on the server route.
A homeware shop with an empty purchase row
A homeware shop moves its checkout to a hosted payment page on a new subdomain. A week later the dataset shows 4,800 PageView events and 610 AddToCart events, but only 3 Purchase events, against 214 orders in the shop. Pixel Helper finds the pixel on product pages but nothing on the confirmation page, because the hosted checkout uses its own template. The developer adds the container to that template through the checkout platform's script field, and the marketer moves the purchase tag from a page-view trigger to the order-complete data layer event. The next day Test Events records a purchase with value and currency, and over the following week the dataset logs 188 purchases against 201 orders, with the remainder explained by visitors who declined consent.
Common mistakes
Testing in a browser full of extensions and concluding the pixel is broken. Looking at the wrong dataset. Publishing nothing after a successful preview. Treating a consent-related gap as a fault and loosening the banner to close it. Adding a second pixel install to "make sure", which creates the double counting covered on the duplicate conversions page. Ignoring event quality once events arrive; the low event match quality page picks up from there.
