Why Your Converlay Data Health Score Was Wrong — and How We Fixed It

Converlay's dashboard shows every merchant a Data Health Score — a single number meant to answer one question at a glance: is your tracking actually working? Underneath that score sit four dimensions, and one of them, Signal Completeness, had been reporting a false problem to nearly every shop using Converlay. Fleet-wide, it matched zero of 18,655 checkout events it checked. Not because the data was missing — because the check itself was looking in the wrong place.

Converlay v2.27.0 fixes that. If your Health Check has been showing a "missing customer identifiers" warning or a lower-than-expected score, this is why, and it is resolved automatically — no action needed on your end.

What the Data Health Score Measures

The Health Check on your dashboard breaks tracking quality into four dimensions: Purchase Coverage (are your purchases reaching your ad platforms), Consent Rate (how much of your traffic is trackable under your consent settings), Destination Delivery (are your routing attempts succeeding), and Signal Completeness — whether the events you send carry the customer identifiers, like hashed email and phone, that ad platforms use to match a conversion back to a real person.

Signal Completeness matters because match quality is not optional infrastructure — it is the difference between Meta, Google, and TikTok correctly attributing a sale to your campaign or discarding it as unmatched. The Health Check report weights this single dimension at 20% of your overall score, and it also feeds directly into the "missing email identifiers" recommendation banner merchants see when something looks off.

The Bug: A Field Name That No Longer Existed

The Signal Completeness query filtered checkout events on payload.user_data.em and payload.user_data.ph — abbreviated field names. But neither Converlay's pixel nor its SDK has ever written data under those keys. Both send full email and phone fields, plus SHA-256 hashed copies in a separate user_hints_hashed object for platforms that require pre-hashed identifiers.

The mismatch meant the query matched nothing, on any shop, ever. Across the fleet it checked 18,655 checkout events and found identifiers on exactly zero of them — signalCompleteness was pinned at 0% regardless of what a merchant's pixel was actually sending. That single stuck-at-zero dimension could depress an otherwise healthy Data Health Score by up to 25 points, and it fired the "missing email identifiers" recommendation for shops whose tracking had no problem at all. The same bug understated match rate in the separately-weighted PDF Health Check report by up to 50 points, since that report scores the identical dimension.

What Changed in v2.27.0

The identifier check now accepts either the plain field (email, phone) or the hashed field (user_hints_hashed), using NULLIF so an empty string is correctly treated as absent rather than as a present-but-blank value — the pixel writes email: '' when no customer is attached to an event, which the old logic would have miscounted either way. In short: the check now looks for what your store actually sends, not for a field name that was never in use.

There is nothing to configure. The fix applies automatically to every shop, and your Data Health Score and Signal Completeness dimension will recalculate on your next dashboard load or scheduled report.

What to Expect on Your Dashboard

If your Signal Completeness score jumps and the "missing email identifiers" recommendation disappears, that is expected — it means your pixel was sending identifiers correctly the whole time, and the score simply couldn't see it. Your Data Health Score overall may rise by a meaningful amount if Signal Completeness had been dragging it down.

If Signal Completeness stays low after this fix, that is now a real signal worth acting on: it means your storefront genuinely is not attaching email or phone to a share of your checkout events, and improving your consent and checkout capture flow would measurably improve match quality with your ad platforms.

Also in v2.27.0: Tracking Survives Store Domain Renames

The same release fixes a second reliability issue. When a merchant renames their store's .myshopify.com domain — common during a rebrand or after a store consolidation — Converlay's stored shop record used to keep pointing at the old domain while session tokens carried the new one, silently breaking authenticated calls and making the app appear dead. Converlay now follows domain renames automatically, keeping the old domain on file so already-collected events and cached pixel settings keep working through the transition instead of dropping data.

Accurate Reporting Is the Whole Product

A tracking tool that reports its own health incorrectly undermines the one thing merchants install it for: knowing whether their conversion data can be trusted. This fix does not add a feature — it makes an existing number tell the truth, which matters more.

Converlay connects your Shopify store to 26 destinations — including Meta, Google Analytics 4, TikTok, Pinterest, Klaviyo, and more — via server-side event forwarding, with a Data Health Score that now reflects what your pixel is actually sending.

Install Converlay from the Shopify App Store and check your Health Check score — if it just went up, now you know why.