Always-on taste curation — the free moodboard extension for creatives
Get the free extension
PortfolioPricingTasteBlog
Back to Hub

CAPI EMQ Benchmarks: Feeding the Meta Algorithm

CAPI EMQ Benchmarks: Feeding the Meta Algorithm

Why Advantage+ Shopping requires an EMQ score of 8.0+, and how to fix data deduplication.

G
Growth Engineering
Data InfrastructurePublished 2026-08-02Updated May 1, 2026

What is a good Event Match Quality (EMQ) score for Meta Conversions API?

In 2026, a 'Great' Event Match Quality (EMQ) score for the Meta Conversions API (CAPI) is between 8.0 and 10. A score of 6.0 to 7.9 is considered 'Good,' while anything under 4.0 is 'Poor' and requires immediate engineering intervention. High EMQ is critical because Meta's automated algorithms (like Advantage+ Shopping) rely entirely on server-side signals to find converting demographics. To achieve a score above 8.0, you must properly hash and pass high-fidelity PII (specifically hashed email and phone numbers) and ensure strict `event_id` deduplication if you are running the CAPI in parallel with the browser Pixel.

As third-party cookies face total deprecation and browser-level tracking degrades, the Meta Conversions API (CAPI) is no longer an optional upgrade—it is the foundational plumbing of your ad account.

However, simply "turning on CAPI" is not enough. The Meta Andromeda engine requires high-fidelity data to optimize delivery. The health of this data is measured by your Event Match Quality (EMQ) score.

Understanding the EMQ Tiers (2026)

Meta scores your EMQ on a scale from 0 to 10. This score reflects how successfully Meta can match your server-side conversion event to a specific user profile in its database.

EMQ ScoreStatusAlgorithmic Impact
8.0 – 10GreatOptimal delivery. Advantage+ AI has enough signal to lower CPAs efficiently.
6.0 – 7.9GoodAcceptable, but missing advanced attribution. (Often missing hashed phone numbers).
4.0 – 5.9OKWeak matching. Account will likely struggle to exit the learning phase.
Under 4.0PoorSevere data gaps. Often caused by broken deduplication or missing User Agent data.

Status

Optimal

The Hashing Hierarchy

  • Hashed Email (em) Impact+ 4.0 EMQ Points
  • Hashed Phone (ph) Impact+ 3.0 EMQ Points

Recommendation:To achieve an 8.0+ score on Purchase events, you must prioritize sending PII. All PII (email, phone, name) must be normalized (lowercase, trimmed) and SHA-256 hashed on your server before being passed to Meta. However, do NOT hash the 'client_ip_address' or 'client_user_agent'—those must be sent in plain text.

The Deduplication Warning: If you run the Pixel and CAPI concurrently (which is recommended), you MUST pass an identical 'event_id' from both sources. If you fail to do this, Meta will double-count the conversion, artificially inflating your ROAS dashboard while simultaneously feeding corrupt data to the bidding algorithm.

Signal Quality vs. Creative Quality

A common misconception is that a perfect EMQ score will save a bad ad. It will not.

EMQ simply ensures that when a good ad does convert, the algorithm learns exactly who bought it so it can find more people like them. If your creative is failing to stop the scroll, no amount of server-side tracking will lower your CPA.

This is why modern growth architecture requires a dual-pronged approach: flawless data infrastructure (CAPI at 8.0+ EMQ) paired with high-velocity creative testing. By using a programmatic platform like eonik to generate a massive diversity of hooks, you feed the Advantage+ engine the creative volume it needs. By maintaining a high EMQ score, you ensure the engine can actually learn from the winners eonik produces.

Related Essays

EssayJuly 25, 2026

The End of the Agency Retainer

Why the era of paying $15,000 a month for 30 video variations is over, and how programmatic assembly is shifting the balance of power back to the brand.

Read Essay
EssayJuly 25, 2026

Why AI Editing Fails Without Human Strategy

You can generate 1,000 video variations a minute, but if the foundational psychology is wrong, you just created 1,000 losing ads. Here is why the "human-in-the-loop" is mandatory.

Read Essay

What to read next

Continue with guides that match where you are — production, research, methodology, or shortlist.
  • Paid social creative blog hub

    Read guide
  • How to generate AI ads

    how to generate ai ads workflow

    Read guide
  • Technical playbooks

    paid social implementation guides hub

    Read guide
  • On brand AI ad workflow

    Read guide
  • Creative testing methodology

    creative testing methodology team ops

    Read guide
  • Ad variant testing workflow

    ad variant testing workflow steps

    Read guide

The Mac app for making ads

Your next ad, without the busywork.

Bring your footage and your AI clips. eonik puts together the finished, on-brand cut — and you approve every frame before it ships.

macOS 15+ · Apple Silicon · Free to start

1
eonik

Finished, on-brand ads without the busywork.

Product

  • Pricing
  • Creative Testing
  • MCP for agents

Knowledge

  • Ad Library
  • Knowledge Hub
  • Blog

Solutions

  • DTC Brands
  • Agencies
  • Growth Teams

Company

  • About
  • Community
  • connect@eonik.ai
PrivacyTerms

© 2026 eonik. All rights reserved.