Customer Data Unification and Segmentation

Customer Data Unification means taking all the different pieces of information a business has about a person and stitching them together into one complete record. Imagine a customer named Maria. The sales team has her name, email, and phone number from a purchase. The support team has her name, company name, and a support ticket. The marketing team has her email and the fact she clicked three different emails. These are all separate ‘data sources’ — different computer systems that each hold a piece of Maria. Without unification, the business treats Maria as three different people. With unification, all that data becomes one profile that says: Maria, email maria@example.com, phone 555-1234, works at Contoso, bought product X, opened three emails, and submitted one support ticket.

Microsoft Customer Insights achieves this through a process called ‘match and merge’. First, it pulls data from source systems — these could be Dynamics 365 Sales, Dynamics 365 Customer Service, a custom website database, or even a spreadsheet. This ingested data is raw and messy: names might be spelled differently (‘Maria’ vs ‘M. Smith’), emails might be missing, phone numbers might have different formats. Customer Insights then runs ‘matching rules’ that look for common identifiers like email address or phone number. If it finds a match — ‘maria@example.com’ in the sales data and ‘maria@example.com’ in the marketing data — it deduces that both rows belong to the same human. It merges them into one ‘unified customer profile’.

What happens to conflicting data? If sales says Maria lives in London and marketing says she lives in Manchester, the system uses ‘priority-based’ rules. The business decides which source is more trustworthy — maybe sales data is considered more reliable because it came from a verified purchase. So the profile keeps the address from sales. You can also set up ‘deduplication’ rules to prevent a single customer from appearing twice in the unified view.

Once you have unified profiles, segmentation is the next step. A ‘segment’ is a group of customer profiles that share certain characteristics — for example, “all customers who bought a product in the last 30 days” or “all customers who have not opened an email in 90 days and are at risk of leaving”. Customer Insights lets you build these segments using rules based on any data in the unified profile: demographics, purchase history, support interactions, website behaviour, or custom fields you add.

Segments can be ‘static’ — you define the rules once, and the segment captures exactly the people who meet those criteria at that moment. They can also be ‘dynamic’ — the segment updates automatically as new data comes in. For instance, a dynamic segment for “customers with an open support ticket older than 7 days” will refresh every day so the marketing team always has the current list.

Why does this replace older methods? Before unified customer data platforms, businesses would export lists from each system and try to merge them in Excel — a process that was slow, error-prone, and outdated the moment new data arrived. Customer Insights automates the matching, merging, and updating, giving a real-time, reliable single view that powers personalised campaigns through Dynamics 365 Marketing or other engagement channels.

Key terms you must know for MB-910:

An IT professional working with Dynamics 365 often starts by identifying where customer data lives. Suppose a mid-sized company called Northwind Traders has three systems: a Dynamics 365 Sales instance storing leads and opportunities, a separate Dynamics 365 Customer Service instance tracking support cases, and a legacy SQL database holding old purchase orders from a discontinued website. The IT person’s job is to connect these to Customer Insights so that a single customer view is created.

Here is the step-by-step reality: The IT pro logs into Customer Insights and first sets up data sources. They choose the ‘Dynamics 365’ connector for Sales and Service, and the ‘SQL Server’ connector for the legacy database. They test the connections to ensure authentication works. Next, they review the schema — which fields exist in each table? Sales has ‘emailaddress1’, Service has ‘primarycontactemail’, and the SQL database has ‘Email’. These are all the same concept (email) but with different names. The IT pro creates a ‘mapping’ that tells Customer Insights to treat all three as the same attribute.

Then they configure match rules. They decide that a match on email address is the strongest signal. If two records share the same email, they are definitely the same person. They also set a secondary rule: if first name, last name, and postcode match, that is also a match, because sometimes customers have different email addresses. The IT pro runs a preview match to see how many duplicates are found. They notice some false positives — ‘John Smith’ might match with ‘John A. Smith’ on name and postcode but actually be different people. They add a rule to exclude matches where the last purchase date differs by more than five years, to reduce errors. After tuning, they run the full unification process.

Now the IT pro needs to create segments for the marketing team. The marketing manager wants: “Customers who purchased in the last 60 days and live in London”. The IT pro builds a segment rule: purchase date greater than 60 days ago AND city equals London. They save it as a dynamic segment. They also build a second segment: “Customers with an unresolved support case older than 5 days” so the service team can follow up. The IT pro then sets up ‘export destinations’ — for example, sending the London segment to Dynamics 365 Marketing for an email campaign, and the unresolved-case segment to a Power Automate flow that triggers a reminder call.

Finally, the IT pro monitors the system. They check the ‘data quality’ dashboard to see how many records failed to match. They notice that 5% of records have no email address, so they ask the sales team to start capturing emails during sign-up. They also schedule weekly data refreshes so the unified profiles stay up to date. The end result: the marketing campaign drives a 20% higher open rate because messages are relevant to recent purchasers, and service calls drop because the team proactively reaches out to stuck customers.