{
  "name": "Meta CAPI & Pixel Checker",
  "publisher": "Sterling Lab",
  "url": "https://sterling.company/meta-capi-pixel-checker",
  "summary": "The Meta CAPI & Pixel Checker is a free self-audit of a Meta Pixel and Conversions API setup. You answer 14 yes, no or not-sure questions and get a score out of 100 and a list of likely gaps, biggest first, each linked to a fix. It does not scan your website or read your Meta account, and your individual answers are not sent anywhere.",
  "description": "Free self-audit of a Meta Pixel and Conversions API setup: 14 yes/no/not-sure questions, a weighted score out of 100 and a list of likely gaps, each linked to a fix.",
  "license_note": "Weights and band cut-offs are Sterling Lab's own definitions, not industry statistics or a benchmark.",
  "privacy": "Nothing is scanned and individual answers are never sent anywhere. Analytics records only that a check was finished, with the score, the band and the number of gaps. The score is computed from the answers alone, so an agent can compute it from this file without calling any server.",
  "answers": {
    "yes": "full weight",
    "unsure": "half weight (label in the UI: Not sure)",
    "no": "zero",
    "na": "check left out of both sums; allowed only for questions that have na_label"
  },
  "total_weight": 27,
  "formula": "score = round((sum of weights answered yes + 0.5 × sum of weights answered not sure) ÷ sum of weights of the applicable checks × 100). A check answered na is left out of both sums.",
  "core_rule": "The checks with weight 3 are core checks. A no or not sure on any of them keeps the band out of solid, even when the score is 80 or higher.",
  "bands": [
    {
      "id": "solid",
      "range": "80 to 100",
      "min": 80,
      "max": 100,
      "meaning": "Solid on paper. The remaining items are tidy-up work.",
      "next_step": "Fix the remaining items when convenient, then re-run the check after any change to tags, plugins, consent or the server."
    },
    {
      "id": "gaps",
      "range": "50 to 79",
      "min": 50,
      "max": 79,
      "meaning": "Gaps worth fixing. Some of what Meta merges and optimizes on is partial or guessed. A failed or unconfirmed core check holds you here even with a score of 80 or more.",
      "next_step": "Start with the first item in the fix list. If a core check is named, fix that one before anything else."
    },
    {
      "id": "unreliable",
      "range": "0 to 49",
      "min": 0,
      "max": 49,
      "meaning": "Hard to trust. Fix the top items before you change bids or budgets.",
      "next_step": "Do not change campaign budgets or optimization events yet. Fix the top three items, then run the check again."
    }
  ],
  "stages": [
    {
      "id": "reach",
      "title": "Events reach Meta"
    },
    {
      "id": "dedup",
      "title": "Deduplication and value"
    },
    {
      "id": "match",
      "title": "Matching quality"
    },
    {
      "id": "setup",
      "title": "Settings and upkeep"
    }
  ],
  "questions": [
    {
      "id": "both",
      "stage": "reach",
      "weight": 3,
      "core": true,
      "short": "Events sent by pixel and by CAPI",
      "question": "Your key events (a purchase, a lead) reach Meta from the browser pixel and also from a server through the Conversions API.",
      "how_to_check": "Events Manager → your dataset → Overview: each key event should list both Browser and Server as sources.",
      "na_label": null,
      "gap_if_no_or_unsure": {
        "title": "Meta may see only what the browser lets through",
        "text": "Browsers block scripts, drop cookies and close tabs before a request finishes. Sending the same event from a server as well fills some of those gaps. An event that comes from one source only has nothing to be merged with."
      },
      "fix_guide": {
        "label": "Meta Conversions API setup",
        "url": "https://sterling.company/guides/meta-conversions-api-setup"
      }
    },
    {
      "id": "onesource",
      "stage": "reach",
      "weight": 2,
      "core": false,
      "short": "One source per event",
      "question": "Each event has one pixel and one server source. A store app plus a Tag Manager pixel, or two CAPI plugins, are not both sending the same purchase.",
      "how_to_check": "Open your confirmation page with the Meta Pixel Helper extension: one Purchase (or Lead) fire is expected. Then list every app, plugin and container that sends events to the dataset.",
      "na_label": null,
      "gap_if_no_or_unsure": {
        "title": "The same conversion may arrive from several sources",
        "text": "Our dedup article describes a store with its native Meta app and another CAPI source both running: several streams for one purchase and no shared event ID. Pick one source of truth per event."
      },
      "fix_guide": {
        "label": "Meta CAPI installed, attribution still broken: the dedup key",
        "url": "https://sterling.company/blog/meta-capi-pixel-dedup-2026"
      }
    },
    {
      "id": "testevents",
      "stage": "reach",
      "weight": 1,
      "core": false,
      "short": "A test conversion seen in Test Events",
      "question": "After the last change to tags, plugins or the server, someone completed a test conversion and saw it in Test Events as a Browser row and a Server row.",
      "how_to_check": "Events Manager → Test Events: open your test URL, complete a purchase or a lead form, and look for two rows for the same conversion, one labelled Browser and one labelled Server.",
      "na_label": null,
      "gap_if_no_or_unsure": {
        "title": "Nobody has seen both copies of an event arrive",
        "text": "Events that show green in Events Manager have not been tested end to end. If only one side appears in Test Events, one of the two sources is not firing at all, and deduplication is not your first problem."
      },
      "fix_guide": {
        "label": "Conversions API and the duplicate-event trap",
        "url": "https://sterling.company/guides/facebook-conversion-api-gtm-server-side"
      }
    },
    {
      "id": "eventid",
      "stage": "dedup",
      "weight": 3,
      "core": true,
      "short": "Same event name and event ID",
      "question": "For each event the pixel and the server send the same event name and the same event ID, so Meta can merge the two copies into one.",
      "how_to_check": "Test Events: expand the Browser row and the Server row for the same conversion and compare the Event ID field character for character. A missing or different ID means the copies are not merged.",
      "na_label": null,
      "gap_if_no_or_unsure": {
        "title": "Meta may count each conversion twice, or drop one copy",
        "text": "Without a shared event ID Meta falls back to guessing from customer data. Our dedup article describes the result: the two copies are not merged, conversions look inflated or one copy is thrown away, and optimization learns from unreliable numbers."
      },
      "fix_guide": {
        "label": "Meta CAPI installed, attribution still broken: the dedup key",
        "url": "https://sterling.company/blog/meta-capi-pixel-dedup-2026"
      }
    },
    {
      "id": "idfield",
      "stage": "dedup",
      "weight": 2,
      "core": false,
      "short": "Event ID passed in the right field",
      "question": "The event ID is born once, on one side, and handed to both channels: as eventID in the pixel call and as the top-level event_id field of each server event, not inside user_data or custom_data.",
      "how_to_check": "In your tag template or code, find where the ID is generated and where each channel reads it. Test Events shows an event ID on the server row only when Meta read the field.",
      "na_label": null,
      "gap_if_no_or_unsure": {
        "title": "The event ID may be present but in a place Meta does not read",
        "text": "The pixel names it eventID, the Conversions API names it event_id, and every layer between your site, Tag Manager and server has to keep one string. If the browser and the server each generate their own, the two never match."
      },
      "fix_guide": {
        "label": "Meta CAPI installed, attribution still broken: the dedup key",
        "url": "https://sterling.company/blog/meta-capi-pixel-dedup-2026"
      }
    },
    {
      "id": "value",
      "stage": "dedup",
      "weight": 3,
      "core": true,
      "short": "Value and currency on Purchase and Lead",
      "question": "Purchase events carry the real order value and currency (a lead event carries the value you assigned to a lead), on the pixel and on the server copy.",
      "how_to_check": "Test Events: open a test purchase and compare the value and currency parameters of both rows with the real order.",
      "na_label": null,
      "gap_if_no_or_unsure": {
        "title": "Conversion value is missing or wrong",
        "text": "Without a real value and currency, value optimization and any revenue figure in Ads Manager rest on empty or invented numbers. Check the value on both copies of the event."
      },
      "fix_guide": {
        "label": "Facebook ads not converting: find the leak",
        "url": "https://sterling.company/guides/facebook-ads-not-converting"
      }
    },
    {
      "id": "delay",
      "stage": "dedup",
      "weight": 1,
      "core": false,
      "short": "Server events sent soon after the action",
      "question": "Server events are sent soon after the action happens, not in a late batch, and you know how long the delay is.",
      "how_to_check": "Ask whoever runs the integration when the server event is sent: at once from the order or form, or later from a queue or an export. Compare the event time in Test Events with the time of your test.",
      "na_label": null,
      "gap_if_no_or_unsure": {
        "title": "A late server event may miss its browser twin",
        "text": "Meta merges two copies of an event only when they arrive within the window it uses for matching duplicates. A server event sent much later may arrive as a second, separate conversion."
      },
      "fix_guide": {
        "label": "Meta CAPI installed, attribution still broken: the dedup key",
        "url": "https://sterling.company/blog/meta-capi-pixel-dedup-2026"
      }
    },
    {
      "id": "emq",
      "stage": "match",
      "weight": 2,
      "core": false,
      "short": "Event Match Quality checked per event",
      "question": "You know the current Event Match Quality of each key event and have looked at which customer details are missing when it is low.",
      "how_to_check": "Events Manager → Overview → open the event: Event Match Quality is shown per event, with the parameters that would improve it.",
      "na_label": null,
      "gap_if_no_or_unsure": {
        "title": "Nobody is watching how well server events are matched to people",
        "text": "Event Match Quality is Meta's own gauge of how well a server event can be tied to a real user. A thin set of customer details leaves a good server event hard to attribute."
      },
      "fix_guide": {
        "label": "Meta CAPI installed, attribution still broken: the dedup key",
        "url": "https://sterling.company/blog/meta-capi-pixel-dedup-2026"
      }
    },
    {
      "id": "cookies",
      "stage": "match",
      "weight": 2,
      "core": false,
      "short": "fbp and fbc sent with server events",
      "question": "The server events include the browser’s _fbp and _fbc cookie values in user_data.",
      "how_to_check": "Test Events → expand the Server row → user_data: fbp and fbc should be present as strings, not empty.",
      "na_label": null,
      "gap_if_no_or_unsure": {
        "title": "Server events may lack the click and browser IDs",
        "text": "Our dedup article explains that fbc is the Facebook click ID and fbp the browser ID. Without fbc the server event cannot confirm the click that began the session, and without fbp Meta loses linkage between visits."
      },
      "fix_guide": {
        "label": "Meta CAPI installed, attribution still broken: the dedup key",
        "url": "https://sterling.company/blog/meta-capi-pixel-dedup-2026"
      }
    },
    {
      "id": "hashing",
      "stage": "match",
      "weight": 2,
      "core": false,
      "short": "Customer details hashed before sending",
      "question": "Customer details sent through the Conversions API (email, phone, name, city, zip) are hashed with SHA-256, in lower case, before they leave your server.",
      "how_to_check": "Test Events → expand the Server row → user_data: hashed values look like long opaque strings. Plain text, or a warning icon, means a field is not hashed.",
      "na_label": null,
      "gap_if_no_or_unsure": {
        "title": "Customer details may be sent unhashed, or not matched",
        "text": "Our dedup article notes that Meta expects hashed customer details and that an event can arrive while its enrichment fails. Mainstream libraries hash for you; custom code often skips it."
      },
      "fix_guide": {
        "label": "Meta CAPI installed, attribution still broken: the dedup key",
        "url": "https://sterling.company/blog/meta-capi-pixel-dedup-2026"
      }
    },
    {
      "id": "consent",
      "stage": "match",
      "weight": 1,
      "core": false,
      "short": "Cookie banner and pixel agree",
      "question": "Your cookie banner and the pixel agree: the pixel starts after consent is given, and you know what the server sends for visitors who never use the banner.",
      "how_to_check": "Open your site in a clean browser, accept the banner and check that the pixel fires after Accept. Then ask who owns the server side what it sends when there was no banner choice.",
      "na_label": "No banner needed",
      "gap_if_no_or_unsure": {
        "title": "The pixel and the server may follow different consent rules",
        "text": "Our dedup article describes a consent bridge that holds back the pixel while the server event fires unconditionally, so Meta sees only one half of each conversion. The consent article covers the same pattern for Google."
      },
      "fix_guide": {
        "label": "Consent Mode v2 and missing GA4 conversions",
        "url": "https://sterling.company/blog/consent-mode-v2-ga4-attribution-gap"
      }
    },
    {
      "id": "domain",
      "stage": "setup",
      "weight": 1,
      "core": false,
      "short": "Domain verified in Meta Business",
      "question": "The domain your events come from is verified in your Meta business account.",
      "how_to_check": "Meta Business settings → Brand safety → Domains: your website domain should be listed and verified.",
      "na_label": null,
      "gap_if_no_or_unsure": {
        "title": "The domain may not be verified",
        "text": "Domain verification is a Meta setting that ties your site to your business account. If it is missing or unconfirmed, open the setting and verify it before you rely on the event setup below."
      },
      "fix_guide": {
        "label": "Meta Ads service: CAPI deduplication and account setup",
        "url": "https://sterling.company/services/meta-ads"
      }
    },
    {
      "id": "aem",
      "stage": "setup",
      "weight": 2,
      "core": false,
      "short": "Events prioritized in Aggregated Event Measurement",
      "question": "You have opened the Aggregated Event Measurement settings for your domain and the events you optimize for are in the priority list, with the most important one first.",
      "how_to_check": "Events Manager → Aggregated Event Measurement (or Web Event Configuration): compare the list and its order with the conversions your campaigns optimize for.",
      "na_label": null,
      "gap_if_no_or_unsure": {
        "title": "The event priority list may not match your campaigns",
        "text": "Aggregated Event Measurement is Meta’s setting for which events from a domain are available and in what order. If nobody has compared the list with the conversion your campaigns optimize for, the campaign may be aimed at an event that is not configured."
      },
      "fix_guide": {
        "label": "Meta Ads service: CAPI deduplication and account setup",
        "url": "https://sterling.company/services/meta-ads"
      }
    },
    {
      "id": "reconcile",
      "stage": "setup",
      "weight": 2,
      "core": false,
      "short": "Monthly check against real orders",
      "question": "In the last month someone compared the purchases or leads Ads Manager reports with real orders or leads in your shop, CRM or inbox.",
      "how_to_check": "Compare last month’s Meta-reported conversions with your own records. A steady gap is normal; a sudden drop, or a number that stops moving, means something broke.",
      "na_label": null,
      "gap_if_no_or_unsure": {
        "title": "Meta numbers are not checked against real orders",
        "text": "Our dedup article starts from accounts where Ads Manager reports fewer purchases than the shop or the CRM. A regular comparison with real orders is what shows that a setup has drifted."
      },
      "fix_guide": {
        "label": "Server-side tracking service",
        "url": "https://sterling.company/services/server-side-tracking"
      }
    }
  ],
  "output_order": "Sort the gaps by points lost, biggest first: a no loses the full weight of the check, a not sure loses half of it. Ties keep the order of the questions in this file.",
  "worked_example": {
    "scenario": "An online shop whose pixel and server both send events, with no cookie banner needed. consent = na. eventid = no (the server generates its own event ID). emq = unsure. hashing = unsure. domain = unsure. Every other check = yes.",
    "steps": [
      "Applicable weight: 27 - 1 = 26",
      "Lost: 3 + 1 + 1 + 0.5 = 5.5, so earned is 26 - 5.5 = 20.5",
      "Score: 20.5 / 26 x 100 = 78.8, shown as 79",
      "Band: 79 is in 50 to 79, so gaps"
    ],
    "gaps_in_order": [
      "eventid (no, weight 3): 3 points lost",
      "emq (unsure, weight 2): 1 point lost",
      "hashing (unsure, weight 2): 1 point lost",
      "domain (unsure, weight 1): 0.5 points lost"
    ],
    "how_to_compute": "1) Remove na checks from the total. 2) For each check lose its full weight for no, half for unsure, nothing for yes. 3) earned = applicable weight - total lost. 4) score = round(earned / applicable weight x 100). 5) Pick the band; if any core check is no or unsure, a score of 80 or more still gives gaps. 6) Sort the gaps by points lost, biggest first."
  },
  "human_page": "https://sterling.company/meta-capi-pixel-checker",
  "markdown_page": "https://sterling.company/meta-capi-pixel-checker/index.md",
  "paid_follow_up": "https://sterling.company/services/conversion-tracking-audit",
  "updated": "2026-11-21"
}
