serkankaya.
Türkçe sürüm
Let’s work together ↗

How to Improve Event Match Quality (EMQ): Step-by-Step Guide

What Meta's Event Match Quality score means, why it stays low and how to raise it. Customer information parameters, fbc, external_id and privacy limits.

Event Match Quality (EMQ) is a score out of 10 that shows how well Meta may be able to match an event received through Conversions API to a person with a Meta account. You raise it by adding more correctly formatted customer information to every server event. The biggest gains usually come from the real IP and user agent, the fbc cookie, a phone number with country code and a stable external_id.

In this post I explain what the score means, how to read it in Events Manager and the most common reasons it stays low. Then I share the checks I use to raise EMQ step by step, along with the privacy limits under KVKK.

What does the Event Match Quality score tell you?

The Event Match Quality score summarizes how likely Meta is to connect a server event to a real person. Meta’s best practices documentation defines the score as “out of 10” and states that higher matching quality produces better results.[1] An unmatched event is still counted in Events Manager, but it cannot be attributed to an ad and contributes little to optimization.

The logic behind the score is simple. The Meta Pixel in the browser sits naturally close to the user’s Meta cookies and session. A server, on the other hand, only knows the user through the information sent to it. That is why the user_data object in a server event is the only basis for matching.

Form or checkout Server: user_data Meta matching Event attributed to ad
A server event can only be linked to a Meta account through its user_data fields; EMQ summarizes the strength of that link.

EMQ should be read as a data quality indicator, not a success metric. A higher score does not automatically lift ad performance. But as the share of unmatched events falls, Meta’s optimization is fed with more real results.

Where do you find Event Match Quality and how do you read it?

Event Match Quality appears in Events Manager separately for each event sent through Conversions API.[1] When you open an event’s details in the pixel’s Overview screen, you see the match score and the list of customer information parameters that came with the event. That list is the fastest way to see which field is missing.

Meta’s documentation also states a clear limit: EMQ is currently available only for web events. For offline and physical store events, app events and conversion leads, Meta suggests contacting your Meta representative.[1] If you send sales from a CRM with the physical_store source, you may not see this score; the matching rules are still the same.

When I read the score, I look at three things:

  • Event type: PageView and Purchase are not judged by the same yardstick. A page view has no email; a purchase does. Events that collect identity details should score higher.
  • Parameter coverage: Event details show what share of events carries each field. A field that arrives only in part of the events shows that data drops on one path of the flow.
  • Change over time: A sudden drop in the score usually points to a site update, a form change or a consent setting cutting off data.

Meta’s developer documentation does not give a single “good score” threshold. So I compare each event with its own history and with similar events in the account.

Which customer information determines EMQ?

The fields that determine EMQ are the customer information parameters in Conversions API. Meta’s best practices documentation lists email (em), IP address (client_ip_address), first and last name (fn, ln) and phone (ph) as examples of high-quality parameters.[1] The same document recommends sending external_id and event_id with all events.

The fields fall into two groups. Contact and identity details are hashed; browser and click details are sent unhashed.

FieldWhat it carriesHashing
emLowercase email with spaces trimmedSHA-256 required
phPhone with country code, no symbolsSHA-256 required
fn, ln, ct, zp, countryFirst name, last name, city, zip code, countrySHA-256 required
external_idThe person’s ID on the site or in the CRMRecommended
client_ip_addressThe user’s real IP addressNot hashed
client_user_agentThe user’s browser user agentNot hashed
fbcThe fbclid from the ad clickNot hashed
fbpThe Pixel’s browser IDNot hashed

Meta’s customer information documentation says that sending the IP address and user agent together may help improve event matching. For website events, client_user_agent is required anyway.[2]

The fbc value has the format fb.1.<creation time>.<fbclid> and is stored in the _fbc cookie. fbp is the browser ID in the _fbp cookie.[3] Because fbc links the event to a specific ad click, it is especially valuable for ad traffic. I explain how these cookies are created in the Meta Pixel setup post.

Field combinations that are not enough on their own

Some fields cannot identify a person on their own. According to Meta’s documentation, from Graph API v13.0 onward an event is invalid if it only contains subsets of city, country, state, zip code, gender and user agent.[1] Events with only date of birth and user agent, or only a name and gender, are in the same situation. An event must carry at least one strong identifier.

How do you improve Event Match Quality?

Short answer

How do you improve Event Match Quality?

Event Match Quality improves when you add more correctly formatted customer information to the server event. According to Meta, sending additional customer information parameters may help increase the score. The biggest gains usually come from the real IP and user agent, the fbc and fbp cookies, a phone with country code and a stable external_id.

  1. Add the user's real IP address and user agent to the server event.
  2. Read the _fbc and _fbp cookies on the landing page and carry them to the server with the event.
  3. Normalize the phone with its country code and the email in lowercase, then hash them with SHA-256.
  4. Add a stable external_id to every event.
  5. Enrich events beyond the form or checkout with details from the CRM record.
  6. Monitor parameter coverage in Events Manager and restore any field that drops.

Meta’s documentation sums up this approach in one sentence: sending additional customer information parameters may help increase Event Match Quality.[1] In practice I work in the order below. It starts with the step that fixes the most fields with the least effort.

  1. Carry the real user IP and user agent

    When the server event is created, take the IP address and user agent of the browser that made the request. If you are behind a proxy or CDN, read the IP from headers such as X-Forwarded-For in the correct order.

  2. Get fbc and fbp to the server

    Read the _fbc and _fbp values on the landing page. Pass them to the server in a hidden field on form submit or with the order record. If _fbc is missing and the URL has an fbclid, build the value in Meta's format.

  3. Normalize contact details

    For email, trim spaces and lowercase the letters. For phone, remove symbols and the leading zero, and always add the country code. Then hash with SHA-256.

  4. Add a stable external_id

    Pick a value that does not change, such as a member number, CRM record number or a visitor ID stored in a first-party cookie. Send the same value for the same person with every event.

  5. Enrich events with record data

    If checkout only collects email, add the phone, name and city from the logged-in user's account as well. Use of this data must be permitted.

  6. Test and monitor

    Check incoming fields in the Test Events screen. After launch, monitor parameter coverage in Events Manager weekly and find the cause of any field that drops.

Why does a phone number with country code matter so much?

Meta’s customer information documentation says a phone number must include a country code to be used for matching. The rule applies even if all your data comes from one country.[2] Forms in Turkey often collect phones as 0532 …. If this value is hashed without converting it to 90532…, the resulting hash matches no record at Meta.

The same problem shows up with email. An email with capital letters or a trailing space produces a different hash. Because hashing cannot be reversed, Meta does not notice the mistake; the field looks sent, but it is useless.

// Normalization before hashing (Node.js)
import { createHash } from 'node:crypto';

const sha256 = (v) => createHash('sha256').update(v).digest('hex');

function normalizeEmail(email) {
  return email.trim().toLowerCase();
}

function normalizePhoneTR(phone) {
  let digits = phone.replace(/\D/g, '').replace(/^0+/, '');
  if (!digits.startsWith('90')) digits = '90' + digits;
  return digits;
}

const user_data = {
  em: [sha256(normalizeEmail(' [email protected] '))],
  ph: [sha256(normalizePhoneTR('0532 000 00 00'))],
};

The phone function in the example assumes Turkish numbers only. For forms that collect international patients or customers, it is safer to take the country code from a separate form field.

Keep fbc and fbp values fresh

Meta recommends regularly refreshing fbc and fbp values if you store them as user parameters.[1] Keeping an fbc from months ago in a lead record and resending it with every event links the event to an old click. In the CRM I store the first click’s fbc in one field and the last known value in another.

What are the most common reasons for a low EMQ?

When EMQ stays low, the cause is usually wrongly sent data rather than missing data. The comparison below shows the differences I see most often in setup audits.

Setup that lowers the score

  • The IP field carries the server's or CDN's address
  • The server library name is sent as the user agent
  • fbc and fbp never reach the server
  • Phone is hashed without a country code
  • An external_id that changes every session is used

Setup that supports the score

  • The user's real IP is read from the correct header
  • The browser's own user agent is carried over
  • _fbc and _fbp are passed with the form or order record
  • Phone is normalized to a format starting with 90
  • The same external_id goes with every event for the same person
Accuracy matters as much as the number of fields; a wrong value weakens matching as much as an empty one.

Meta asks you to always send the real IP address.[2] If you use a tagging server such as server-side GTM, verify in the Test Events screen that the client reads these values correctly from the incoming request.

Another common cause is cookie consent. When a user rejects advertising cookies, _fbp and _fbc are not created and some fields are not sent. This drop is not a bug; it is the result of the consent choice. I cover the consent flow in detail in the Consent Mode v2 post.

Deduplication problems also affect the score indirectly. If the Pixel and the server send the same event with different event_id values, the event is counted twice. According to Meta’s deduplication documentation, the event name and event_id must be the same in both events.[4]

How should EMQ be handled in health tourism and e-commerce?

In health tourism and e-commerce, the most valuable event often happens in the CRM or order system, not on the site. A confirmed consultation, a deposit or a delivered order happens after the user has left the site. For these events to match a person, the identifiers collected at the first touch must be stored in the record.

The structure I use works like this: when a lead form is submitted, fbc, fbp, IP, user agent and the phone with country code are written to the lead record. When a stage changes in the CRM, the server builds the event from that record and sends it to Meta. I cover the Meta side of this flow in the Meta Conversions API post and the CRM side in the health tourism offline conversions post.

E-commerce is a little easier. Email, phone and address are already collected at checkout. A common mistake here is that, in guest checkout, these details reach the Pixel but are not added to the server event. Even with ready-made integrations on platforms like Shopify, you need to check in Events Manager which fields are sent.

In both sectors, EMQ can serve as a health indicator for the measurement side of performance marketing work.

EMQ checklist

The list below collects the points I check before launching a Conversions API setup and whenever EMQ drops.

Event Match Quality checklist0/8

Conclusion

Event Match Quality is a data quality score that shows how well the server events you send to Meta can be linked to a person. What raises the score is not more fields but the right fields sent in the right format: the real IP and user agent, fbc and fbp, a phone with country code and a stable external_id.

As a next step, open the details of your most valuable event in Events Manager and note which parameter is missing or arrives with low coverage. Starting there and applying the steps above in order also makes the change in the score measurable.

Key terms

external_id
external_id is the unique ID an advertiser assigns to a person in its own system, such as a member number, CRM record number or cookie-based visitor ID. Hashing is recommended in Conversions API, and it strengthens matching by linking events that belong to the same person.
fbp
fbp is the browser ID the Meta Pixel writes to the _fbp cookie in the format fb.1.<creation time>.<random number>. It is sent unhashed in Conversions API requests and links a server event to the Pixel event coming from the same browser.
Customer information parameters
Customer information parameters are the fields added to the user_data object of a Conversions API event that are used to match the event to a person. Fields such as email, phone, name and city are hashed with SHA-256, while IP address, user agent, fbc and fbp are sent unhashed.

Frequently asked questions

What is Event Match Quality?

Event Match Quality (EMQ) is a score out of 10 that shows how effective the customer information in a server event sent through Conversions API may be at matching that event to a Facebook or Meta account. The score appears per event in Events Manager.

How do you improve Event Match Quality?

Start by sending the real user IP address, user agent, fbc and fbp with every server event. Then normalize email and phone with its country code, hash them with SHA-256 and add a stable external_id to every event. Meta states that sending additional customer information parameters may help increase the score.

What is a good EMQ score?

Meta's developer documentation does not give a single target score; it only says the score is out of 10 and that higher matching quality produces better results. That is why it is more reliable to compare each event with its own history and with other events in the same account. Events that collect identity details, such as purchases, should score higher than page views.

Is EMQ shown for offline and app events too?

According to Meta's best practices documentation, Event Match Quality is currently available only for web events. For offline, physical store, app and conversion leads events, Meta suggests contacting your Meta representative. Strengthening matching with the same parameter rules still helps for these events.

Is sending hashed data compliant with KVKK?

Hashing does not stop data from being personal data. Before sending email or phone to Meta, the privacy notice, explicit consent where required and cookie consent steps must be in place. Decide which fields to send together with your legal team.

Sources

  1. Meta for Developers: Best Practices for Conversions API
  2. Meta for Developers: Customer Information Parameters
  3. Meta for Developers: ClickID and the fbp and fbc Parameters
  4. Meta for Developers: Handling Duplicate Pixel and Conversions API Events
  5. Meta Business Help Center: About event match quality