Conversions API Troubleshooting: Missing Purchases, Duplicates, Delays (and What to Do Next)
Your CAPI setup looks correct, but Events Manager tells a different story — missing purchases, double-counted conversions, or events arriving hours late. This guide walks you through the most common CAPI failure modes, a structured troubleshooting flow, and the decision table to diagnose and fix each issue.

Quick Answer: Why Your Conversions API Isn't Working (and How to Fix It)
The Conversions API (CAPI) creates a direct server-to-server connection between your website and Meta, bypassing browser limitations that cause the Pixel to miss events. When CAPI works, it recovers signal lost to ad blockers, iOS ATT opt-outs, and browser errors. When it breaks, it either drops events silently (missing purchases), counts them twice (broken deduplication), or sends them too late for Meta to use effectively.
Most CAPI problems fall into five categories. Here's what to check first:
- Missing Purchase events — CAPI events aren't arriving in Events Manager. Usually caused by a broken server-side connector, incorrect access token, or misconfigured event payload.
- Duplicate conversions — Both Pixel and CAPI fire the same event, but
event_idmatching isn't working. Meta counts every purchase twice, inflating your reported ROAS. - Delayed events — Server events arrive hours or days after the conversion. Meta discounts late events for optimization, so they contribute less to ad delivery decisions.
- Low Event Match Quality (EMQ) — CAPI events don't include enough customer parameters (hashed email, phone) for Meta to match them to real user profiles.
- Connectivity failures — The server-side integration stops sending events entirely, often after platform updates, token expiration, or hosting changes.
The fix depends on the failure mode. This guide gives you the diagnostic flow, decision table, and validation checklist to identify and resolve each one. If you want to skip the manual checks, Adfynx can surface CAPI health issues — missing events, deduplication gaps, EMQ drops — across all your connected Meta accounts with read-only access.
Why CAPI Problems Are Hard to Spot
Unlike a broken Pixel — which Pixel Helper can flag instantly — CAPI failures are invisible from the browser. Your website looks fine. Customers complete orders. But the server-side events that should reach Meta either don't arrive, arrive corrupted, or arrive late.
The damage compounds quietly:
Missing events mean Meta sees fewer conversions. The algorithm has less data to optimize against, so it delivers your ads to broader, lower-quality audiences. Your CPM rises and ROAS drops — but you can't tell if the cause is creative fatigue, audience saturation, or broken tracking.
Duplicate events inflate your metrics. If both Pixel and CAPI report every purchase separately, your Events Manager shows twice the actual conversions. Budget decisions based on a 4x ROAS that's really 2x lead to overspending on underperforming campaigns.
Delayed events weaken optimization. Meta's algorithm weighs recent signals more heavily. A Purchase event that arrives 6 hours after the conversion contributes less to the optimization model than one that arrives in seconds. Chronic delays reduce the algorithm's ability to find high-converting users.
The core challenge: you can't see CAPI events by browsing your site. You have to look in Events Manager, compare server event counts against your backend data, and test payloads directly. This makes CAPI issues harder to catch than Pixel issues — and more damaging when they persist.
If you manage multiple ad accounts, catching CAPI failures across all of them manually is impractical. A tool like Adfynx can monitor server event delivery and flag gaps across your connected accounts with read-only access — so you find out when CAPI breaks, not weeks later when performance has already degraded.
What to do next: Walk through the common failure modes below to identify which category your issue falls into.
Common CAPI Failure Modes
1. Missing Purchase Events
Symptoms: Your e-commerce platform shows completed orders, but Events Manager shows fewer (or zero) Purchase events from the "Server" source. Your Pixel may still be sending Browser events, masking the CAPI gap.
Likely causes:
- CAPI connector or integration is misconfigured or disconnected
- Access token has expired or been revoked
- Event payload is malformed (wrong event name, missing required fields)
- Server environment changed (hosting migration, SSL certificate issue, firewall rule blocking outbound requests to Meta's API)
- Partner integration (Shopify, WooCommerce plugin) was updated and broke the CAPI connection
How to verify: Open Events Manager → Data Sources → your Pixel → Test Events. Complete a test purchase on your site. Within a few minutes, you should see a Purchase event with "Server" as the source. If only "Browser" appears, CAPI isn't sending.
2. Duplicate Conversions (Deduplication Failure)
Symptoms: Events Manager shows roughly twice the number of purchases as your actual backend orders. Both "Browser" and "Server" sources show events, but each purchase appears as two separate events.
Likely causes:
event_idis not being sent by either Pixel or CAPI (or both)- Pixel and CAPI are sending different
event_idvalues for the same event event_idformat mismatch (e.g., one sendsorder_123, the other sends123)- Partner integration doesn't support
event_iddeduplication by default
How to verify: Compare Purchase event count in Events Manager against your actual orders for the same 7-day period. A ratio close to 2:1 confirms deduplication failure. For a detailed walkthrough of the deduplication ratio test, see our Pixel health check guide.
3. Delayed Server Events
Symptoms: CAPI events appear in Events Manager but with significant lag — hours or even days after the actual conversion. The events "count" but contribute less to real-time optimization.
Likely causes:
- Server-side processing queue is backed up or batching events too aggressively
- The
event_timeparameter in the CAPI payload doesn't match the actual conversion time - Third-party connector batches events and sends them in bulk at intervals instead of real-time
- Server timezone misconfiguration causing
event_timeto be off
How to verify: Complete a test purchase and note the exact time. Check Events Manager → Test Events to see when the Server event appears. If there's more than a few minutes of delay, investigate your server-side event pipeline.
4. Low Event Match Quality (EMQ)
Symptoms: Events Manager shows an EMQ score below 6.0 for your key events. Meta can receive your events but can't reliably match them to user profiles, which reduces optimization accuracy.
Likely causes:
- CAPI events don't include hashed customer parameters (
emfor email,phfor phone,external_id) - Customer parameters are sent unhashed (Meta requires SHA-256 hashing)
- Parameters are sent but contain placeholder or template values instead of real data (e.g.,
{{email}}) - Advanced Matching is not enabled for your Pixel
How to verify: Events Manager → Data Sources → your Pixel → Overview. Check the EMQ score for each event type. Click into the details to see which parameters are being received and their match rates.
5. Complete CAPI Disconnection
Symptoms: No Server events appear in Events Manager at all. Only Browser (Pixel) events are visible. This may happen suddenly after working for weeks or months.
Likely causes:
- API access token expired (tokens can expire or be invalidated by permission changes)
- Partner integration was disabled, uninstalled, or updated
- Server hosting changed (new IP, new domain, SSL reconfiguration)
- Meta's API endpoint changed and the integration wasn't updated
- Permissions were modified in Business Manager, revoking the token's access
How to verify: Events Manager → Data Sources → your Pixel → Test Events. Navigate your site and check for Server events. If none appear, check your CAPI connector's status dashboard or logs for error messages.
What to do next: Use the troubleshooting flow below to systematically diagnose your specific issue.
CAPI Troubleshooting Flow
Follow these steps in order. Each step eliminates a category of problems so you can focus your investigation.
Step 1: Check if Server Events Are Arriving at All
Open Events Manager → Test Events. Navigate your site and trigger key events (page view, add to cart, purchase). Look at the source column.
- Both Browser and Server events appear → CAPI is connected. Proceed to Step 2.
- Only Browser events → CAPI is disconnected or broken. Jump to Step 5 (connectivity).
- Only Server events → Pixel is broken, but CAPI works. This is a separate issue — check your Pixel installation.
Step 2: Check Event Counts Against Backend Data
Pull your Purchase event count from Events Manager for the last 7 days. Compare it to your actual order count from your e-commerce platform for the same period.
- Ratio ~1:1 → Deduplication is working. Proceed to Step 3.
- Ratio ~2:1 → Deduplication is broken. Both Pixel and CAPI are double-counting. Fix
event_idmatching (see Decision Table). - Ratio well below 1:1 → Events are being lost. Some purchases aren't being tracked by either source. Investigate which source (Pixel, CAPI, or both) is dropping events.
Step 3: Check Event Timing
Complete a test purchase and note the exact time. Watch Test Events for when the Server event appears.
- Within seconds to 1–2 minutes → Timing is fine. Proceed to Step 4.
- More than 5 minutes → Investigate your server-side event pipeline for batching or queue delays.
- Hours or days → Your integration is batch-sending events. Switch to real-time delivery or reduce batch intervals.
Step 4: Check Event Match Quality
Events Manager → Data Sources → Pixel → Overview. Find the EMQ score for your key events.
- EMQ ≥ 6.0 → Good match quality. Your CAPI setup is healthy.
- EMQ 4.0–5.9 → Needs improvement. Add more customer parameters to your CAPI payloads.
- EMQ < 4.0 → Poor. Prioritize adding hashed email, phone, and external ID to server events.
Step 5: Diagnose Connectivity Issues
If no Server events appear in Test Events:
1. Check your CAPI connector's admin panel or dashboard for error logs.
2. Verify the API access token is still valid (tokens can expire after password changes or permission updates).
3. Test the CAPI endpoint directly using Meta's Graph API Explorer with your token.
4. Check your server's outbound network — can it reach graph.facebook.com?
5. If using a partner integration (Shopify, WooCommerce), check if the plugin needs reconfiguration after an update.
What to do next: Match your specific problem to the decision table below for the precise fix.
Decision Table: Problem → Cause → Verify → Fix → Validate
| Problem | Likely Cause | How to Verify | Fix | How to Validate |
|---|---|---|---|---|
| No Server events in Events Manager | Access token expired or revoked | Check CAPI connector logs for authentication errors; test token in Graph API Explorer | Generate a new access token in Events Manager → Settings → CAPI; update token in your connector | Trigger a test purchase; confirm Server event appears in Test Events within 2 minutes |
| No Server events after platform update | Partner integration (Shopify/WooCommerce plugin) broke during update | Check plugin status; look for error messages in plugin settings or server logs | Reconfigure the CAPI plugin; reconnect to your Pixel; re-enter credentials if needed | Run Test Events; confirm Server events resume for all key event types |
| Purchase count ~2x actual orders | event_id not being sent, or Pixel and CAPI send different IDs | Compare Events Manager Purchase count to backend orders over 7 days; check event payloads in Test Events for event_id field | Ensure both Pixel and CAPI send an identical event_id per event occurrence (e.g., order_12345) | Recheck the 7-day ratio after fix; should be ~1:1 |
| Purchase count ~1.5x actual orders | Partial deduplication — some events have matching event_id, others don't | Inspect individual events in Test Events; look for events missing event_id | Audit all checkout paths (web, mobile, app) to ensure event_id is generated consistently for every purchase | Monitor the ratio for 7 days; it should converge toward 1:1 |
| Server events arrive hours late | Connector batches events instead of sending real-time | Check connector settings for batch interval; test a purchase and time the Server event arrival | Switch to real-time event delivery; reduce batch interval to under 1 minute if real-time isn't available | Time a test purchase; Server event should appear within 2 minutes |
event_time is wrong on Server events | Server timezone misconfigured; event_time not set to actual conversion time | Compare event_time in Test Events detail view against the actual purchase time | Fix timezone configuration; ensure event_time is set to the Unix timestamp of the actual conversion, not when the server processes it | Verify event_time matches within 60 seconds of actual purchase time |
| EMQ below 6.0 for Purchase events | Missing customer parameters in CAPI payloads | Events Manager → Pixel → Overview → click EMQ detail; check which parameters are received | Add hashed em (email), ph (phone), and external_id to your CAPI event payloads | Monitor EMQ score over 7 days; target ≥ 6.0 |
| EMQ parameters show "not matched" | Customer data sent unhashed or with wrong format | Inspect CAPI payloads in Test Events; check if em and ph values are SHA-256 hashed | Hash all customer parameters with SHA-256 before sending; remove whitespace and lowercase before hashing | Recheck EMQ detail; parameters should show "matched" status |
| Events arrive but with "Not a Standard Event" warning | Event name doesn't match Meta's standard event list | Check Test Events for event name warnings (e.g., "Purchased" instead of "Purchase") | Correct event names to match Meta's exact standard event names: Purchase, AddToCart, InitiateCheckout | Confirm no event name warnings in Test Events after fix |
CAPI events missing value or currency | Purchase event payload doesn't include revenue data | Check Test Events detail for Purchase event; look for value and currency fields | Add dynamic value (order total) and currency (e.g., "USD") to your Purchase event payload | Verify value and currency appear in Test Events for a new test purchase |
| Server events appear but Events Manager totals don't match | Attribution window mismatch between your comparison periods | Align the date range and attribution window (7-day click, 1-day view) in both Events Manager and your backend comparison | Ensure you're comparing the same time period and attribution window; use the same start/end dates | Recalculate with aligned windows; discrepancy should drop below 20% |
| CAPI stops working after Business Manager permission change | Token was invalidated by permission or admin changes | Check if the system user or admin who generated the token still has the required permissions | Regenerate the token with appropriate permissions; update the token in your connector | Confirm Server events resume in Test Events |
After working through this table, Adfynx can serve as an ongoing monitoring layer — it tracks whether Server events are arriving, flags deduplication ratio shifts, and alerts you to EMQ drops across all connected accounts, with read-only access only.
What to do next: Review the example scenarios below, then use the checklist to formalize your troubleshooting process.
Example Scenarios
Example 1: Shopify Store — Missing Purchases After App Update
A Shopify store spending $25K/month on Meta Ads notices ROAS dropped from 3.2x to 1.8x over two weeks. The media buyer suspects creative fatigue and pauses several ad sets. Performance doesn't recover.
Investigation:
- Events Manager shows Purchase events only from "Browser" source. No "Server" events for the past 12 days.
- The Shopify CAPI app was auto-updated 12 days ago — matching the performance drop timeline.
- Backend Shopify orders are stable at ~150/week. Events Manager shows ~90 Purchase events/week (Browser only).
Diagnosis: The CAPI app update broke the server-side connection. For 12 days, Meta received approximately 40% fewer Purchase events (missing all CAPI-only conversions from iOS users and ad-blocker users). The algorithm had less data to optimize, which degraded delivery quality.
Fix: Reconfigure the CAPI Shopify app — reconnect to the Pixel, re-enter the access token, and verify the event payload. Run a test purchase and confirm Server events appear in Test Events.
Validation: After 7 days, Purchase event count in Events Manager rises to ~145/week (matching backend orders at ~1:1). ROAS begins recovering as the algorithm receives complete conversion data again.
Key lesson: Performance drops that look like creative or audience issues can actually be tracking failures. Always check CAPI connectivity before making optimization changes.
Example 2: WooCommerce Agency — Double-Counting Across Client Accounts
An agency manages 10 WooCommerce client accounts. During a quarterly audit, they compare Events Manager Purchase counts to actual orders for each account. Three accounts show ratios between 1.8:1 and 2.1:1.
Investigation:
- All three accounts use the same WooCommerce CAPI plugin.
- The plugin sends Purchase events server-side but doesn't include
event_idin the payload. - The browser Pixel also fires Purchase on the confirmation page.
- Both sources report every purchase — without
event_idmatching, Meta counts each one twice.
Diagnosis: Deduplication failure. Reported ROAS for these accounts is approximately double the actual value. The agency has been making budget scaling decisions based on inflated numbers.
Fix: Switch to a WooCommerce CAPI plugin that supports event_id, or implement custom event_id generation that passes the same order ID to both Pixel and CAPI. Test with a sample purchase and verify only one Purchase event appears per order in Events Manager.
Validation: Monitor the Purchase ratio for 7 days. It should converge to ~1:1. Recalculate ROAS using backend revenue data for the period where double-counting occurred, and adjust budget decisions accordingly.
If you're managing multiple accounts, running this ratio check manually for each one is tedious. In Adfynx, you'd see deduplication ratios flagged across all connected accounts in a single view — with read-only access, so there's no risk to client campaigns while you diagnose.
For a broader view of how tracking platform choices affect these issues, see our guide on conversion tracking platforms.
CAPI Troubleshooting Checklist
Use this checklist when diagnosing any CAPI issue. Work through it in order — earlier items rule out foundational problems so you don't waste time on downstream issues.
Connectivity & Authentication
- [ ] CAPI connector is active and connected — Check your connector's admin panel (Shopify app, WooCommerce plugin, GTM Server-Side, or custom integration) for connection status.
- [ ] Access token is valid — Test the token in Graph API Explorer or check connector logs for authentication errors. Tokens can expire after password changes, permission updates, or admin removals.
- [ ] Server can reach Meta's API — Verify outbound requests to
graph.facebook.comare not blocked by firewalls, IP restrictions, or SSL issues. - [ ] Correct Pixel ID is configured — The CAPI connector should send events to the same Pixel ID that your browser Pixel uses. Mismatched IDs mean deduplication can't work.
Event Delivery
- [ ] Server events appear in Test Events — Open Test Events, trigger a purchase on your site, and confirm "Server" source events appear within 2 minutes.
- [ ] All key events are being sent — Check that PageView, AddToCart, InitiateCheckout, and Purchase all appear from the Server source, not just Purchase.
- [ ] Events arrive in real-time (not batched) — Time a test event. If Server events take more than 5 minutes to appear, investigate batching settings.
- [ ] Event names match Meta's standard list — Verify event names are exactly
PageView,AddToCart,InitiateCheckout,Purchase(case-sensitive). No variations like "Purchased" or "Add_to_cart".
Deduplication
- [ ]
event_idis present in both Pixel and CAPI events — Check Test Events detail view for theevent_idfield on both Browser and Server events. - [ ]
event_idvalues match between Pixel and CAPI — Both sources must send the identicalevent_idfor the same event occurrence (e.g., same order ID). - [ ] 7-day Purchase ratio is ~1:1 — Compare Events Manager Purchase count to backend orders. A ratio above 1.3:1 suggests deduplication problems.
Event Quality
- [ ] Purchase events include
valueandcurrency— Check Test Events detail for these parameters. Without them, Meta can't optimize for purchase value. - [ ] EMQ score is ≥ 6.0 for key events — Events Manager → Pixel → Overview. Check EMQ for Purchase and AddToCart.
- [ ] Customer parameters are hashed correctly —
em(email) andph(phone) should be SHA-256 hashed, lowercase, with whitespace trimmed before hashing. - [ ]
event_timeis accurate — Theevent_timein CAPI payloads should match the actual conversion time (Unix timestamp), not the server processing time.
Validation Steps After Any Fix
After making changes to your CAPI setup, run these validation steps to confirm the fix worked:
- [ ] Trigger 3–5 test purchases — Complete real or test orders and confirm all appear as Server events in Test Events.
- [ ] Verify deduplication — Each test purchase should appear as one event, not two, even though both Pixel and CAPI fired.
- [ ] Check EMQ after 48 hours — EMQ takes time to recalculate. Check 48 hours after changes to see if the score improved.
- [ ] Monitor 7-day Purchase ratio — Track the Events Manager vs. backend order ratio for a full week to confirm the fix holds under normal traffic conditions.
- [ ] Spot-check event parameters — Review 2–3 Purchase events in Test Events to verify
value,currency,event_id, and customer parameters are all present and correct.
For a complete tracking audit that covers both Pixel and CAPI health, see our Pixel health check checklist.
Common Mistakes When Troubleshooting CAPI
1. Blaming ad creative or audience targeting when the real problem is broken CAPI. A sudden ROAS drop often triggers creative changes or audience restructuring. But if the cause is missing CAPI events (fewer conversions reaching Meta), optimization changes won't help. Always check tracking health before making campaign-level changes.
2. Adding CAPI without configuring event_id deduplication. This is the single most common CAPI mistake. Teams add server-side tracking to improve signal coverage, but forget to implement event_id matching. The result: every conversion is double-counted, inflating ROAS and distorting budget decisions. Fix deduplication before doing anything else.
3. Assuming CAPI is working because Events Manager shows conversions. Events Manager increasingly uses modeled conversions to fill gaps. It can report purchase numbers even when your CAPI is completely disconnected. Always check the source breakdown (Browser vs. Server) in Test Events, not just the totals.
4. Not re-testing CAPI after platform updates. Shopify app updates, WooCommerce plugin updates, theme changes, and hosting migrations can all break CAPI silently. After any update, run a test purchase and verify Server events appear in Test Events. This takes 2 minutes and prevents weeks of degraded tracking.
5. Sending unhashed customer parameters. Meta requires customer data (email, phone) to be SHA-256 hashed before sending via CAPI. Sending unhashed data is a privacy violation and will result in poor Event Match Quality. Always hash on your server before including parameters in the event payload.
6. Ignoring event_time accuracy. If your event_time parameter reflects when your server processed the event (possibly hours later due to queuing) instead of when the conversion actually happened, Meta treats the event as stale. Set event_time to the actual conversion timestamp.
7. Testing CAPI only on desktop. Mobile traffic often follows different code paths than desktop. A CAPI integration that works on desktop checkout might not fire for mobile purchases, especially if you have separate mobile templates or app-based checkout flows. Test on both.
8. Not monitoring CAPI after the initial setup. CAPI isn't "set and forget." Tokens expire, plugins update, servers change. Build a recurring check (weekly or bi-weekly) to verify Server events are still arriving and deduplication is holding. For how to build this into your measurement routine, see our guide on measuring ROAS reliability.
FAQ
What is the Conversions API and why do I need it?
The Conversions API (CAPI) is a server-to-server connection between your website's server and Meta's systems. It sends conversion events (purchases, add-to-carts, page views) directly from your server, bypassing the browser. You need it because browser-based Pixel tracking loses signal from ad blockers, iOS ATT opt-outs, and browser restrictions. CAPI recovers that lost signal, giving Meta more complete data to optimize your ad delivery.
Does CAPI replace the Meta Pixel?
No. CAPI and the Pixel work together — they're complementary, not interchangeable. The Pixel captures browser-side interactions; CAPI captures server-side data. Running both together with proper event_id deduplication gives Meta the most complete picture of your conversions. Removing either one creates a blind spot.
How do I know if my CAPI is actually sending events?
Open Events Manager → Data Sources → your Pixel → Test Events. Navigate your site and trigger a conversion (add to cart, purchase). In the Test Events view, look for events with "Server" as the source. If you only see "Browser" events, your CAPI isn't sending data. Check your connector's status, access token, and server logs for errors.
What is event_id and why does deduplication matter?
event_id is a unique identifier you attach to each event occurrence. When both Pixel and CAPI send the same event with the same event_id, Meta knows it's one conversion and counts it once. Without matching event_id values, Meta counts the Pixel event and the CAPI event separately — doubling your reported conversions. This inflates ROAS and leads to wrong budget decisions.
What's a good Event Match Quality (EMQ) score?
An EMQ score of 6.0 or above is the general target. This means Meta can reliably match most of your server events to real user profiles. Below 6.0, matching accuracy drops and optimization suffers. Improve EMQ by sending more hashed customer parameters with CAPI events: email (em), phone (ph), and external ID (external_id).
How quickly should CAPI events arrive in Events Manager?
Server events should appear in Test Events within seconds to a couple of minutes after the conversion. If events consistently take more than 5 minutes, your integration may be batching events instead of sending them in real-time. Meta's algorithm weighs recent events more heavily, so chronic delays reduce optimization effectiveness.
Can CAPI track events beyond website purchases?
Yes. CAPI can send web events, app events, offline conversions, and messaging events. For web, the standard events include PageView, ViewContent, AddToCart, InitiateCheckout, Purchase, and more. You can also track offline events like in-store purchases and attribute them to your Meta campaigns. Custom events beyond Meta's standard list are also supported.
Why did my CAPI stop working after a Shopify/WooCommerce update?
Platform updates can change how CAPI plugins interact with your checkout flow, modify API endpoints, or reset configuration settings. Some updates require re-authentication or reconnection to your Pixel. After any platform or plugin update, immediately test a purchase and verify Server events appear in Test Events. Most CAPI outages from updates are fixable by reconfiguring the plugin.
How do I fix CAPI events that are missing value and currency?
Without value and currency on Purchase events, Meta can't distinguish a $10 sale from a $500 sale — which means it can't optimize for purchase value (ROAS optimization). Fix this by ensuring your server-side event payload dynamically populates value with the order total and currency with the correct currency code (e.g., "USD", "EUR"). Check Test Events to confirm both fields appear after the fix.
How often should I check that CAPI is still working?
At minimum, spot-check weekly by comparing your Events Manager Purchase count to your actual backend orders. Do a full CAPI health check monthly — or immediately after any server change, plugin update, or platform migration. Token expiration is another common cause of silent disconnection, so check your connector's authentication status monthly as well.
Conclusion
Conversions API troubleshooting comes down to five questions:
1. Are Server events arriving at all? If not, check connectivity, token, and connector status.
2. Are events being double-counted? If your Purchase count is ~2x actual orders, fix event_id deduplication.
3. Are events arriving on time? Delays beyond a few minutes weaken optimization. Switch to real-time delivery.
4. Is Event Match Quality high enough? EMQ below 6.0 means Meta can't match your events to users effectively. Add hashed customer parameters.
5. Does the data match reality? Your Events Manager numbers should be within 20% of your actual backend data. Larger gaps mean something is broken.
Next steps:
1. Run the troubleshooting flow from this guide to identify your specific failure mode.
2. Match your issue to the decision table for the precise fix.
3. After fixing, run the validation checklist: 3–5 test purchases, deduplication ratio check, EMQ review after 48 hours.
4. Set a weekly calendar reminder to spot-check the Purchase event ratio against backend orders.
5. After any platform update, plugin change, or server migration, immediately re-test CAPI delivery.
Try Adfynx — Monitor CAPI Health With Read-Only Access
If you want continuous monitoring of your Conversions API health across all your Meta ad accounts, Adfynx flags missing server events, deduplication failures, and EMQ drops with read-only access. No write permissions, no campaign modifications — just visibility into what's working and what's broken. There's a free plan to get started. Start here →
---Suggested Internal Links
- "Meta Pixel Health Check Checklist: Validate Events, Dedupe, and Match Quality" → /blog/meta-pixel-health-check-checklist-validate-events-dedupe-match-quality-2026 — complete Pixel + CAPI health audit
- "Best Advertising Platforms for Conversion Tracking (2026)" → /blog/best-advertising-platforms-conversion-tracking-2026 — how tracking platform choice affects CAPI implementation
- "Measuring ROAS in 2026: What's Noisier, What Still Works, and What to Do Next" → /blog/measuring-roas-2026-noisy-what-works-what-to-do-next — ROAS measurement reliability in the context of tracking gaps
- "Meta Pixel Signal Quality: Fix Duplication, Delay & Distortion" → /blog/meta-pixel-signal-quality-fix-duplication-delay-distortion-2026 — deep dive into signal quality issues related to CAPI
- "How to Find Your Facebook Pixel ID (and Verify It's Working)" → /blog/how-to-find-facebook-pixel-id-verify-working-2026 — verify your Pixel ID matches your CAPI configuration
- "Attribution-Driven Ad Measurement: How to Evaluate Adtech Firms" → /blog/best-adtech-firms-attribution-driven-ad-measurement-2026 — how CAPI data quality affects attribution accuracy
- "ATC vs IC vs PUR: The Real Optimization Logic Behind Meta Conversion Events" → /blog/meta-conversion-events-atc-ic-pur-optimization-guide-2026 — which events to send via CAPI and when
You May Also Like

Meta Pixel Health Check Checklist: Validate Events, Dedupe, and Match Quality
A broken Pixel doesn't always look broken — it just silently feeds Meta bad data. This checklist walks you through every health check item: event coverage, deduplication, Event Match Quality, and the diagnostic flow to catch issues before they inflate your CPM.

Best Advertising Platforms for Conversion Tracking (2026): How to Choose Without Breaking Attribution
Comparing conversion tracking platforms in 2026? This guide breaks down what reliable tracking actually means, compares Pixel vs CAPI setups, and gives you a decision table to pick the right platform for your team size and tech stack.

Meta Pixel Signal Quality: Fix Duplication, Delay & Distortion to Stop CPM Inflation (2026)
Your ads aren't getting expensive because of competition—it's because Meta doesn't trust your signals. Learn how duplication, delay, and distortion in Pixel/API data cause CPM spikes and how to fix them.
Subscribe to Our Newsletter
Get weekly AI-powered Meta Ads insights and actionable tips
We respect your privacy. Unsubscribe at any time.