Supplement ad tracking: separate new customers from repeat revenue
Connect supplement campaigns to initial orders, repeat purchases, subscription payments, and refunds with a clearly defined revenue basis.
By Mika Garcia · Published
In this article
Define whether you are measuring acquisition or repeat business
A supplement campaign can produce an initial order, an upsell, a later reorder and a recurring subscription payment. Those outcomes are commercially useful, but they are not four newly acquired customers.
Start by separating the questions. Acquisition reporting asks what it costs to bring in a new paying customer. Order reporting counts transactions. Cohort reporting follows the revenue from the customers acquired in a period. Each needs its own denominator and observation window.
Keep the customer definition explicit when people use different emails or checkout methods. A source system’s new-customer flag and your own cross-store identity rule may not describe the same population. Record the distinction rather than silently combining them.
Choose the source of truth for each money movement
Use the successful order or payment source for revenue, with a stable transaction identifier and currency. Treat the confirmation-page visit as a journey event, not independent proof that a new payment occurred.
DATALYR lists Shopify and CheckoutChamp among its commerce connections. Follow the documentation for the source you actually use and verify one transaction from checkout through the recorded event. Do not assume two connected sources describing the same sale will automatically deduplicate each other.
For subscription orders, determine which system owns the recurring charge. Write down how initial orders, renewals, refunds and cancellations arrive. A cancellation stops or changes a relationship; it is not necessarily a refund of money already collected.
Follow one buyer through the ledger
Consider a hypothetical buyer acquired through Campaign A. Their first order is $60. A separately charged upsell adds $20, a later recurring order adds $50, and a successful partial refund returns $10.
The ledger contains three positive payments and one refund, producing $120 after the stated refund adjustment. It still contains one newly acquired buyer. Whether each later payment receives Campaign A credit depends on the chosen attribution and renewal policy; do not assume it automatically.
This example excludes product costs, shipping, tax treatment and processor fees. Its $120 is a defined receipt measure, not profit. Add the costs relevant to a profitability decision before using it to scale spend.
One buyer is not three acquired customers
1. Initial order + upsell
$60 + $20
Two successful payments in this example.
2. Recurring order
+$50
A later payment from the same customer.
3. Refund-adjusted receipts
$120
$60 + $20 + $50 − $10.
Show first-order and cohort returns side by side
Suppose a campaign spends $1,000 and acquires 20 new customers whose first orders total $1,200. Its initial-order revenue ROAS is 1.2× and media cost per new customer is $50.
If the same cohort later produces another $600 in receipts and $100 in refunds within 60 days, refund-adjusted cohort receipts reach $1,700. On that defined basis, the ratio to original media spend is 1.7×. It is a cohort revenue ratio, not proof that every later order was caused by the original campaign.
Wait until comparable cohorts reach the same age before comparing their 60-day result. A fresh launch should not appear worse merely because it has had less time to reorder.
Two valid views of one acquisition cohort
Initial orders
$1,200 / $1,000 = 1.2×
Revenue from the first purchase stage.
60-day adjusted receipts
$1,700 / $1,000 = 1.7×
Includes $600 later receipts and $100 refunds.
Keep product context out of unintended destinations
Supplement product names, quiz responses and landing-page URLs can reveal more than a generic purchase amount. Audit those fields before sending events to an advertising destination. A tracking implementation should not collect detailed personal answers merely to make a campaign report look more complete.
DATALYR documents health-related filtering and platform restriction controls. Review them when the business or destination requires them, and separately inspect scripts installed outside DATALYR’s managed paths.
A clean revenue ledger and permitted ad delivery are different requirements. Do not relabel a restricted event or move it server-side to conceal its meaning.
Verify a purchase, a repeat payment, and a refund separately
Trace one permitted test journey with a recognizable campaign label. Confirm the landing visit, successful transaction, customer identity, value and currency. Then inspect the source’s treatment of a repeat payment and a refund.
Check whether a one-click upsell reuses an order or creates another transaction in your actual checkout. That detail changes deduplication and order-count expectations. Compare the measurement record to the source ledger rather than assuming one checkout session equals one money movement.
Finish with a report contract: new-customer definition, payment source, repeat-revenue policy, refund basis, currency and cohort window. That makes it possible to distinguish an acquisition problem from a reorder or margin problem.