Skip to main content
EcommerceIndustry solution

Conversion tracking for ecommerce brands, measured against your orders

An ecommerce brand has one unambiguous source of truth — the orders in its store — and almost every measurement problem is a discrepancy against it. Purchases lost when a wallet payment skips the thank-you page, the same order counted twice by two channel apps, refunds that never reverse so reported revenue only climbs. The build sends purchase from the order record, deduplicated on the order ID, and reconciles per order over a closed week rather than comparing totals that hide offsetting errors.

Free30 minutesNo obligation

Ecommerce · what breaks

What ecommerce brands usually run into

  1. 01

    Four dashboards, four revenue figures

    Summing to more than was banked, with nobody able to say which to use.

  2. 02

    Reported ROAS better than the bank statement

    Which is how the problem usually gets noticed.

  3. 03

    Refunds never reversed

    So the reported number drifts further from reality with your return rate.

Ecommerce · the fix

What changes — Ecommerce

  • One order, one conversion

    Deduplicated on the order ID at every destination.

  • Revenue that reconciles to the store

    Per order, over a closed window.

  • Refunds and cancellations reversed

    So the number can go down as well as up.

Ecommerce · architecture

The conversion map for ecommerce brands

Five events, one of which decides the money. The others exist so bidding has enough volume to learn from.

  1. 1

    Product view

    High volume, browser-side is fine. Feeds retargeting and catalog matching.

  2. 2

    Add to cart

    Browser-side. The first real intent signal and useful for audience building.

  3. 3

    Checkout started

    Browser-side. Where the drop-off analysis lives.

  4. 4

    Purchase

    Server-side, from the order record. This is the one that has to be right.

  5. 5

    Refund

    Server-side, from the refund webhook. Almost never implemented, and it's why revenue only climbs.

Ecommerce · diagnose

Where ecommerce brands tracking usually breaks

  1. 01

    Wallet payments skipping the confirmation page

    Shop Pay and PayPal can be abandoned after payment. The order exists and no platform hears about it.

  2. 02

    Two apps both sending purchase

    A channel integration plus a container is two conversions for one sale.

  3. 03

    Non-web orders in the comparison

    POS, draft and phone orders belong in your total and not in a platform reconciliation.

  4. 04

    Refunds never reported

    Reported revenue only goes one way.

Ecommerce · prove

How the work is proven — Ecommerce

Source of truth

your store's order records, web orders only

  • Order ID present in every destination's event log.
  • Purchase count within a few percent of web orders for a closed week.
  • No order ID appearing twice anywhere.
  • A real refund tested and confirmed reversing.

Ecommerce · free audit

What we check in the free audit

  • Wallet payments skipping the confirmation page
  • Two apps both sending purchase
  • Non-web orders in the comparison
  • Refunds never reported

Checked against

your store's order records, web orders only

Get my free conversion tracking audit

Free30 minutesNo obligation

Ecommerce · when not to buy

What lands in your monthly report — Ecommerce

One page a month, in language you can put in front of a founder or a board.

  • Orders in your store against conversions reported by each platform, for the closed month.
  • The gap, split into attribution overlap, blocked events and consent denial — so you can see which part is a fault and which is expected.
  • Refunds reversed, and the net revenue figure after them.
  • Anything that broke during the month, with the date and the cause.

Ecommerce · what we deliver

The flow we build — Ecommerce

  • Purchase from the order webhook

    Not from a thank-you page a customer can skip or refresh.

  • Order ID as the deduplication key

    Stable and identical on every path without coordination.

  • Customer parameters hashed from the order

    Normalized before hashing, which is where match quality is usually lost.

  • Refund webhook wired as a reversal

    At every destination that accepts one.

Written guarantee

95%+ tracking accuracy

Measured against your store's order records, web orders only. Miss it and we keep working at no cost.

Get my free conversion tracking audit

From $400 · See packages

FAQ

Ecommerce questions, answered

No. Each counts only what it can attribute, so a permanent gap is correct. What matters is that it's explainable and stays stable month to month.

Free · 30 min

Get your ecommerce numbers reconciled

Free30 minutesNo obligation