What is identity resolution in retail?
The process of matching data from different sources and devices to the same customer, creating a single unified profile.
In short
Identity resolution is the process of recognising that records from different channels, devices and systems belong to the same person and combining them into one customer profile. Fashion retailers use it to connect online, in-store, app and loyalty data for personalisation, analytics and service.
How does it work in practice?
Systems use deterministic matching, based on exact identifiers such as email addresses, customer numbers or loyalty IDs, and sometimes probabilistic matching, which estimates likely matches from signals such as name, address or device. When a customer logs in online and later shows a loyalty card in store, both records are linked. The result is stored in a CDP or customer master, with rules for how conflicting data is resolved.
Why does it matter for fashion businesses?
Without identity resolution, a single loyal customer can look like several occasional ones. This distorts metrics such as customer lifetime value and retention, and leads to irrelevant marketing, such as promoting items already bought. Unified profiles also make it easier to respond to data access or deletion requests, because all records for a person can be found. In B2B, a similar process matches retailer accounts, stores and buyers across systems.
How is AI changing it?
Machine learning improves probabilistic matching by learning which combinations of signals reliably indicate the same person. AI also helps clean and standardise names and addresses before matching. As third-party identifiers decline, accurate resolution of first-party data becomes more important for advertising and measurement.
Common pitfalls
- Merging two different people into one profile through weak matching rules.
- Combining data in ways customers did not consent to.
- Missing store data because customers are not identified at the till.
- No process for correcting wrong matches once they are found.
Frequently asked questions
What is the difference between deterministic and probabilistic identity resolution?
Deterministic matching links records using exact shared identifiers, such as the same email address. Probabilistic matching estimates whether records belong together based on similar signals, which covers more cases but is less certain.
Why is identity resolution important for personalisation?
Personalisation relies on knowing a customer's full history. Linking all their interactions gives a more accurate picture of preferences, sizes and purchases across channels.
