← All questionstROAS Campaigns

How should I split tROAS goals between new and returning customers?

What to check before you touch anything

  • Does Google Ads actually receive information about whether a purchaser is new or returning? Without that signal, neither a separate goal nor separate reporting is technically possible.
  • How much does the real value of a new customer versus a returning one actually differ for the business (LTV, average order value, repeat purchase frequency)? If the gap is small, splitting goals may be over-engineering.
  • Is there enough data on new customers separately from returning ones for a campaign/group with its own goal to actually learn? On many accounts, returning customers are a small slice at any given moment, and splitting the goal further shrinks an already limited pool of data.
  • Are your campaigns using the New Customer Acquisition goal (available for Search, Shopping, and Performance Max)? If so, it's worth understanding exactly how it's configured (bid-only vs. value-based) and whether the current setup is introducing any distortion.
  • Are you conflating "acquire a new customer" with "optimize overall ROAS"? These are different business goals, and blending them into one campaign with one target can obscure the real economics of new-customer acquisition.

Possible approaches

  • Google provides a dedicated tool for exactly this: the New Customer Acquisition goal, which either simply tracks the share of new customers (bid-only mode), or explicitly boosts the value of conversions from new customers during optimization (value-based mode, which assigns new customers a higher relative value than returning ones). This is a built-in, Google-supported mechanism, not a workaround you're constructing yourself.
  • If the LTV gap between a new and a returning customer is substantial and measurable, you can pass different conversion values for these two groups through New Customer Acquisition value-based bidding, rather than trying to split them into entirely separate campaigns with different tROAS targets.
  • Splitting into fully separate campaigns (each with its own tROAS for "new" and "returning") makes sense when these two audiences also differ in keywords, landing pages, or offer — that is, when there's a substantive reason for separate structure, not just a difference in value.
  • If data volume for new or returning customers is too small to support a separate goal, it's usually smarter not to fragment campaigns and instead use the value-based approach within one campaign (described above), to avoid creating a data-scarcity problem for learning.
  • It's worth periodically confirming that New Customer Acquisition is correctly identifying "new" customers for this specific business (the data source — hashed customer lists or your own purchase data), since errors in how that base is defined can distort both reporting and the optimization itself.