Skip to content
All field notes
Measurement
Published
Nov 19, 2026
Reading time
12 min read

Meta CRM Lead Coverage Is Low Even When Events Are Received: A Diagnostic Guide

Separate received Conversions API events from Meta CRM lead coverage. Check the native lead source, required fields, lead IDs, event stages, timestamps and report state without guessing at the cause.

AP
AdProof Research
Independent measurement desk
A CRM lead timeline passes through separate event receipt and lead-coverage checks, with one match left unresolved.
Receipt and lead coverage are separate checkpoints. Conceptual illustration; no advertiser results.
In this guide (9 sections)

If Meta says CRM events were received but its CRM Lead Coverage report remains low, treat those as different checkpoints. First confirm that the integration is using Meta's specific Conversions API for CRM payload for native Lead Ads (Instant Forms), then compare the actual raw-lead and later-stage events against the current payload specification. Inspect the lead identifier copied from the lead webhook, required source fields, event timestamps, and the exact reporting state. A successful request or an event visible in Events Manager does not, by itself, establish that Meta recognized the event for CRM lead matching or that it qualifies for funnel analysis.

That distinction is the practical issue in a recent question about CRM lead coverage remaining near one-third while events appear received. The post displayed as about eight hours old at review and had active follow-up. It is not a verified platform defect; replies suggest possibilities, but establish no cause. The official Meta developer documentation below supports the technical requirements.

First separate the measurement states#

The word “received” can describe the transport or event-processing view. Lead Coverage is a different report about leads with matching events uploaded to Meta. Sales-funnel setup is another validation phase. Do not treat these labels as synonyms.

StateWhat it can establishWhat it does not establish on its own
Your sender logged a requestYour integration attempted a requestThat Meta accepted every event or used it in the CRM integration
API response reports success or Events Manager shows eventsA request or event reached a documented receiving/diagnostic surfaceThat the event matched a native Meta lead or qualifies for CRM lead optimization
CRM Lead Coverage report shows matching eventsThe report's defined share of leads have matching events uploadedThat every later lifecycle stage is complete, or that campaign attribution equals a CRM total
Sales funnel can be configured or analyzedThe relevant setup has progressed to that product phaseThat a particular campaign's outcome is incremental or its reported lead total is complete

Meta's CRM developer implementation guide says its payload specification applies to Conversion Leads Optimization events and should not be used for other event types such as website leads. It also explicitly warns that an event can be accepted by the broader Conversions API and appear in Events Manager without being recognized for this CRM integration when it does not follow the specification or match a Meta lead ad. That makes a “request succeeded” indicator a necessary observation, not a sufficient coverage diagnosis.

Keep the denominator visible. Meta's Verify Your Data guide defines Lead Coverage as the percentage of leads that have matching events uploaded to Meta. It lists a running lead campaign generating 200 leads per month, at least 60% Lead Coverage, and correctly formatted required parameters as requirements for the sales-funnel validation phase. These are published setup criteria for that phase; they are not a guarantee that crossing a threshold will resolve every report state or improve campaign performance.

Related guide

Make the next tracking test reproducible.

Use the ungated QA checklist to record event rules, consent, duplicate sends and the report you expect to change.

Read the tracking QA checklist

Opens a practical guide. No email required.

Confirm the event type belongs to this integration#

Before troubleshooting individual fields, establish what kind of lead generated the event. Meta's payload specification currently limits this specification to the Conversion Leads optimization goal for native Lead Ads (Instant Forms) generated on Facebook or Instagram. Meta cautions against applying that payload to other lead-event types.

Record the lead origin, Meta dataset, campaign goal, sender, report range and observation time. If the lead did not originate from a supported native Lead Ad, do not assume the CRM lead-optimization fields make it equivalent. Follow the documentation for the actual event type and ask Meta which report applies; a CRM's total can be broader than the eligible population.

Compare one lead's raw event with the documented contract#

Meta's payload specification identifies the following required fields for the Conversion Leads integration:

FieldDocumented roleDiagnostic check
event_nameCRM stage name; send all stages as they update, including the initial leadVerify the raw lead and subsequent stages are named and sent according to the current specification
event_timeUnix timestamp in seconds for when the CRM stage update occurredCompare the actual sent value with lead creation and stage-update timestamps
action_sourceMust be system_generated for these eventsCheck the serialized payload, not only the sender's configuration screen
user_dataCustomer-information map; at least one supported parameter is requiredConfirm the chosen identifier is present and valid without exposing it in shared evidence
custom_data.lead_event_sourceName of the CRM or source toolCheck spelling, placement, and consistency in the actual request
custom_data.event_sourceMust be crmCheck the exact value and nesting in the sent request

For one recently generated native lead, compare the original record, stored record, serialized payload, response and corresponding Meta report in an authorized, access-controlled trace. Never put access tokens or raw identifiers in public evidence.

Meta identifies a Facebook-generated lead_id from the leadgen_id in the lead-generation webhook as the highest-priority match parameter, and says it is a 15–17 digit number included under user_data. The docs say an invalid lead_id causes event rejection. Copy the identifier from its source and preserve every digit through storage and serialization. Do not substitute an internal CRM row number or hash the platform-generated lead ID. If your implementation uses email or phone instead, Meta's specification says those fields must be hashed; hashing does not itself establish that sharing the data is authorized. Confirm the currently permitted data, access, notice and configuration with the responsible privacy and platform owners before sending any customer information.

Inspect the raw source and exact outbound value; a spreadsheet's display format does not prove the transmitted identifier is intact. The question's author reported that sampled IDs were stored as text and intact. That self-reported check neither proves all implementations safe nor identifies the cause, so a blanket spreadsheet-precision diagnosis would overstate the evidence.

Audit event order and timestamps#

The official payload specification says event_time is a Unix timestamp in seconds for when the lead stage was updated by the CRM. It must occur after the lead-generation time or the event may be discarded. The guide also says the timestamp can be up to seven days before the event is sent to Meta. These are separate checks: timestamp must be after lead creation, and the event must be within the documented lookback limit when sent.

For a small authorized sample, preserve lead creation time and timezone, CRM stage time, serialized Unix-second event_time, send time, response and report range in a private trace.

Check for a date-only field converted to midnight, a timezone conversion in the wrong direction, a value in milliseconds instead of seconds, out-of-order stages, or a retry that sends a stale timestamp. These are diagnostic hypotheses, not conclusions about a particular account. If the computed event time precedes lead creation, the official documentation supports investigating that condition. If the event is older than seven days at send time, compare it with the documented limit before resending. Do not change historical timestamps simply to make an event eligible; preserve the original event time and repair the transformation or retry logic for future events.

Send the raw lead and each stage as it happens#

The required raw-lead event is not interchangeable with sending only a later qualified or booked stage. Meta says to send all stages as updated, including the initial lead stage; its example funnel includes initial lead, marketing-qualified lead, sales opportunity and converted. Use the business's actual stage definitions, but preserve order and do not invent stages to satisfy a dashboard.

Create a private table with one row per lead and stage event, using internal references and keeping raw identifiers in the approved system. Record origin/eligibility, stage and timestamps, required-field checks, send attempt, response, diagnostics and report range. Classify the first unsupported step—selection, stage emission, payload, identifier, send, recognition or report reflection. “Unknown” is valid; aggregate reports do not necessarily expose row-level matches.

Interpret restrictions and coverage without guessing#

Some advertisers may see data-sharing restrictions or other dataset notices. Treat the exact label, scope and effective time as evidence to record, not as a diagnosis by itself. The reviewed question's author reported that the low coverage predated a later restriction notice; that timing argues against casually attributing the existing symptom to the new notice, but it still does not prove another cause. A commenter suggested restrictions might affect some fields; the author reported seeing several expected fields in sampled CRM events. Both are user observations, not a general statement of Meta policy.

The public developer pages cited here document the CRM payload and verification phases. They do not fully specify how every account-specific data-sharing restriction affects the CRM Lead Coverage calculation. If a notice appears, preserve its wording and first-observed time, check the current Meta account guidance available to your authorized administrator, and ask Meta support to clarify its effect on the exact integration. Do not try to evade a restriction by renaming fields, hiding the origin of data, or sending information the business is not authorized to share.

For “Analysis in progress,” record the exact surface, label, range and observation time. The verification guide describes Connect your CRM and Configure your sales funnel phases, but sets no universal duration after which the label proves a fault.

A hypothetical audit that leaves the cause open#

Suppose a lead-generation advertiser has 100 eligible native leads in a completed period. Its sender log records 100 raw-lead requests; a protected sample finds 94 with intact source IDs, while six lack a corresponding ID in the retained trace. The CRM also records 73 later stage updates, and the Lead Coverage report shows 42 matching leads.

These invented numbers do not prove 58 lost leads or a 42% implementation success rate. Confirm the eligible population, repair missing ID traces, check sample payloads and timestamps, then recheck the same cohort. The 73 later updates are stage events, not 73 additional leads. If evidence cannot locate the gap, leave it unresolved and escalate through an authorized support channel.

Practical checklist before changing the integration#

  • Confirm the source is a native Meta Lead Ad covered by this integration.
  • Save the report, population, date range, status and observation time.
  • Trace one raw lead through the exact serialized payload and response.
  • Check required fields and their nesting, including the exact webhook leadgen_id.
  • Verify Unix seconds, ordering after lead creation, and the seven-day send limit.
  • Send and audit the raw lead and subsequent stages separately.
  • Redact identifiers; treat restrictions as account-specific and unresolved unless confirmed.
  • Change one evidenced defect at a time and preserve before/after snapshots.

For broader measurement checks, the marketing tracking QA checklist separates configuration, browser execution and downstream reporting. The audience measurement guide explains how to keep delivery attributes and business outcomes distinct. Google's offline conversion upload guide covers a different platform's CRM upload, processing and attribution checkpoints; its API rules do not transfer to Meta.

AdProof's public materials focus on independent campaign delivery and impression measurement; they do not establish a native Meta CRM CAPI connector or access to Lead Coverage diagnostics. For a separate delivery-measurement question, contact AdProof to confirm supported sources and scope. Take CRM payload questions to Meta or the integration owner.

Frequently asked questions#

If Events Manager received every event, why can Lead Coverage still be low?

Because event receipt and CRM Lead Coverage are different checkpoints. Meta says events that do not follow the Conversion Leads payload or do not match a Meta lead ad may not be recognized for that integration even if the broader Conversions API accepts them and Events Manager shows them.

Does a low Lead Coverage percentage prove that Meta lost events?

No. The percentage describes leads with matching events uploaded to Meta. Compare the eligible native-lead population, raw-lead events, required payload fields, identifiers, timestamps and the exact report before deciding where a gap occurred.

Should I hash the Meta lead ID?

No. Meta describes lead_id as the Facebook-generated 15–17 digit number from the leadgen_id webhook. Preserve it exactly as provided. Meta's documentation requires hashing for email and phone parameters; those are separate identifier types.

What timestamp does the CRM integration use?

Meta specifies Unix time in seconds for the CRM stage update. The timestamp must be after lead creation or the event may be discarded, and Meta's payload page says it may be up to seven days before sending. Check both conditions against the original times.

Can data-sharing restrictions explain a low coverage report?

Do not infer that from a notice alone. Record the exact restriction and when it began, compare that with the symptom's timeline, and get current account-specific guidance from Meta. The public developer pages cited here do not establish the effect of every restriction on Lead Coverage.

Should I resend the whole funnel to raise coverage?

Not without identifying what was omitted or rejected and confirming the event remains within Meta's documented timestamp limit. First retain the original requests, responses, timestamps and diagnostic state. A resend is not a substitute for resolving a malformed, ineligible or out-of-window event.

Your next step

Make the next tracking test reproducible.

Use the ungated QA checklist to record event rules, consent, duplicate sends and the report you expect to change.

Read the tracking QA checklist

Opens a practical guide. No email required.

Back to all field notes