These checks review Event Match Quality (EMQ) and related Meta signals for your pixel / dataset. They can be run from the dashboard. Data is read through authorised Meta connections you or Littledata has set up.
| Outcome | Meaning |
|---|---|
| Error | Likely hurts ad performance or matching. |
| Warning | Investigate—often fixable. |
| Info | FYI only. |
| Success | No issue for this rule. |
| Unknown | No snapshot or not enough data yet. |
What it measures: Meta’s 0–10 score for how well customer information (email, phone, etc.) is sent with Purchase events. Higher scores usually mean better targeting and attribution.
What we compare: Your latest score vs roughly the last 90 days (typical range) for the same event, so you can see drift—not just a single pass/fail.
Rules: When we have 30+ days of history for that event, we only flag Warning or Error if the score is below the absolute band and more than 2 standard deviations below your own trailing average (so normal day-to-day jitter doesn’t flip the check). With less than 30 days of history, we use the absolute thresholds only.
| Outcome | Purchase EMQ (absolute bands) |
|---|---|
| Success | 6.5 or above (or below band but not far enough below your average—see above) |
| Warning | 5.0 – 6.49 (and 2σ rule when baseline exists) |
| Error | Below 5.0 (and 2σ rule when baseline exists) |
| Unknown | No Purchase data in the latest quality snapshot |
You might see: Which match keys are weak (for example coverage under 80% on a key Meta expects).
Same idea as Purchase EMQ, for InitiateCheckout.
| Outcome | EMQ (absolute bands; 2σ rule applies when 30+ days of history) |
|---|---|
| Success | 5.0 or above |
| Warning | 4.5 – 4.99 |
| Error | Below 4.5 |
| Unknown | No checkout event in the snapshot |
Same idea for AddToCart.
| Outcome | EMQ (absolute bands; 2σ rule applies when 30+ days of history) |
|---|---|
| Success | 4.0 or above |
| Warning | 3.0 – 3.99 |
| Error | Below 3.0 |
| Unknown | No add-to-cart event in the snapshot |
What you might see: If your Google / Meta ad account has more than one pixel, we may refresh quality data for those extra pixels and note how their EMQ compares to your main Littledata pixel in the message.
Important rule: Your pass, warning, or error for Purchase / Checkout / Add to cart is still based only on the primary (Littledata) pixel. Sibling comparisons are context, not a second grade.
Behind the scenes (plain English): Quality snapshots rotate through accounts regularly; extra pixels are included when applicable so comparisons stay current.
What we look at: The same kinds of pixel diagnostic warnings Meta shows in their own UI—missing or badly formed customer parameters, and similar matching issues.
Rules:
| Outcome | Rule |
|---|---|
| Success | No diagnostic warnings for that event. |
| Info | One or more Meta dataset-quality warnings—informational only (not Warning/ Error); detail text should mirror what Meta says. |
| Unknown | No web events in the dataset for that check. |
Caveats: Meta’s pipeline can lag; warnings can disappear after fixes propagate.
What they are: About every 30 days we refresh a baseline EMQ per event so slow drift is visible, not only “above / below a fixed number.”
Rule: If today’s score is meaningfully below that baseline (default: worse by more than ~15%), we raise the severity one step—for example from Success toward Info or Warning—even if you still clear the absolute threshold.
Caveats: New pixels need at least one baseline cycle before this extra logic applies.