DTC customer data audit checklist

1. Identity

• Is there a single customer identifier that survives across commerce, ESP, support and, where relevant, POS?
• How are guest checkouts resolved to known customers?
• What happens when a customer uses a second email address?
• Which system is the declared source of truth, and does everything actually defer to it?

2. Events

• Which behavioural events are captured today, and which are captured but never used?
• Are product and variant attributes attached to purchase events?
• Is browse and cart behaviour tied to the customer record or only to a session?
• Are returns, refunds and exchanges represented at all?

3. Zero-party data

• What has the customer told you directly: preferences, sizes, occasion, replenishment cycle?
• Where is it captured, and does it reach the segmentation layer?
• Is anything captured that no downstream system can read?

4. Consent and governance

• Is consent state stored per channel and per region?
• Can a suppression decision be reconstructed after the fact?
• Who can change segmentation logic, and is that change recorded?

5. Destinations

• Which tools can act on the profile today, and which only receive a subset?
• Where does data flow one way when it should flow both?
• What breaks silently, and how would you know?

6. The questions the data must answer

• Which customers are on a path to a second order, and which have stalled?
• What does a customer prefer, beyond what they bought once?
• Which cohort is worth spending margin on?
• Which lifecycle moment is underperforming right now?

Building a lifecycle event taxonomy

An audit ends in a taxonomy: the small set of events every system agrees on, named once and documented. Fewer, well-specified events beat a long list nobody trusts.

• Commerce: order placed, order fulfilled, order refunded, subscription started, paused, cancelled.
• Browse and intent: product viewed, collection viewed, cart started, checkout started.
• Capture: account created, list subscribed, consent changed, quiz or preference submitted.
• Loyalty: enrolled, points earned, reward redeemed, tier changed.
• Service and reviews: ticket opened, ticket resolved, review submitted.
• Retail: in-store purchase, staff-assisted enrolment, identity matched to an online profile.
• For each event: owner, source system, trigger definition, required properties, identity key, retention period.

When a brand needs a customer data audit

• Two systems report different revenue or customer counts for the same period.
• A migration, replatform or loyalty launch is planned in the next quarter.
• Segments have to be exported and rebuilt manually to be usable.
• Nobody can say confidently which system holds the authoritative consent record.