New vs. returning customer ROAS: reconcile the revenue mix

An improving account total can conceal a weaker acquisition mix

Account ROAS can rise while the business acquires fewer new customers. Returning buyers can contribute more attributed revenue, covering a decline in first-customer orders. To understand the movement, split the revenue numerator into mutually exclusive customer groups before changing the campaign budget.

For this analysis, the new-customer revenue component equals attributed revenue from a customer’s first order divided by the full included ad spend. The returning component uses repeat-order revenue over that same spend. The components add to account ROAS. They are diagnostic shares of the account result; neither represents the standalone return of an independently funded campaign.

Classify the customer before aggregating revenue

Create one row per order with order ID, stable customer key, first completed order date, order date, revenue after your stated adjustments, and attribution status. Assign the first completed order to the new-customer group and later orders to returning. Define how cancellations, merged customer accounts and incomplete history are handled.

Google Ads conversion counting can record one or every conversion after an interaction. The selected counting rule does not establish whether a transaction came from a first-time buyer. Use the customer/order history for that classification. [Google Ads: conversion counting options]

A returning buyer may place the first order you have observed in a recently migrated system. Mark that identity as unresolved when earlier history is unavailable. Assigning every unresolved buyer to new acquisition makes a migration look like a marketing improvement. Keep unknown revenue as a third component so the total can still reconcile.

A completed two-period comparison

The following fictional USD figures use two complete, equally defined reporting periods. Both have $2,000 in included ad spend. Revenue is already adjusted under one consistent policy, and all example orders have resolved customer identities.

Account ROAS rose from 4.0× to 5.0×, but the first-customer component fell from 3.0× to 2.0×. The extra $4,000 in returning revenue more than covered the $2,000 decline in first-order revenue. That explains the account movement without claiming that retention caused it.

Check new-customer count as well as first-order revenue. Forty first-time customers spending $150 each generate $6,000. Twenty customers spending $200 each generate $4,000. In this illustration, both the number of acquired customers and their combined first-order revenue fall even though their average order value rises.

Keep scope changes out of the diagnosis

Do not divide first-customer revenue by prospecting spend unless the analysis has a defensible way to match that spend to the cohort. A buyer can encounter several campaign types. Reassigning costs after seeing the outcome can manufacture a favorable acquisition figure.

• Hold the revenue boundary constant: product sales, taxes, shipping, discounts and refunds should follow the same rule in both periods.
• Use equivalent maturity cutoffs if refunds or delayed purchases are still arriving.
• Document the attribution model and lookback scope rather than treating all observed orders as incremental.
• Reconcile new, returning and unresolved revenue to the same account total.

ROAS is also separate from profit. Returning customers may buy a heavily discounted product with a low contribution margin. A higher revenue ratio can coexist with less contribution after advertising. Customer mix explains where the revenue sits; margin analysis answers whether that revenue supports the business.

Choose the next review from the component that changed

If first-customer count falls, examine reach, offer, landing experience and qualified acquisition volume. If count holds while first-order revenue falls, look at product and order-value mix. If unresolved customer revenue expands, repair identity coverage before presenting a new-versus-returning verdict.

I would report the account ratio beside first-customer count, first-order revenue and unresolved coverage. That short set makes it harder for a strong repeat-purchase month to become an unsupported claim about acquisition efficiency.

• Check the revenue-to-spend arithmetic
• Review the acquisition and retention funnel
• Connect acquisition cost with customer value
• Organize customer cohorts

Sources

AI-assisted writing and an AI-generated editorial cover. Source links and checked dates are listed below; illustrative examples are labeled in the text.

• Google Ads: conversion counting options — checked 2026-10-18