Academy · Hands-on

Guided Lab: Build a Customer 360

Do it for real. Thirteen walkthrough steps condensed into ten hands-on exercises — from an empty model to golden records and a published report, each with a self-check so you know it worked. About 150 minutes in a real environment. You need access to an MDM Studio environment (Developer role or above). No environment? Request a hosted sandbox below.

MDM Studio Design Map Govern Trust
Sign in to track your progress

Ask the study tutor

Stuck on a concept? Ask in your own words — answers come from the course material with links to the source lessons. Signed-in learners only; daily limit applies.

Prepare a working model

  1. Sign in, pick your engine, and create (or clone) a model named LAB Customer 360 under Administration → Deployment Models.
  2. Activate it and check it out.
You know it worked when…
  • The header shows your model name with a checked-out indicator.
  • A colleague (or second browser) sees the model as read-only.
Walkthrough reference →

Create the Customer domain and entity

  1. Modelling Studio → Domains & Entities: create domain Customer with default sensitivity Internal.
  2. Add entity Customer inside it.
You know it worked when…
  • The domain lists one entity.
  • The ERD (Data Modelling) shows the entity.
Walkthrough reference →

Define the dictionary

  1. In the Data Dictionary add: customer_number (key, required), first_name, last_name, email, phone, country.
  2. Classify email and phone as Confidential.
You know it worked when…
  • Six attributes exist; customer_number is flagged as the business key.
  • email/phone show a Confidential badge.
Walkthrough reference →

Register two sources

  1. Integration Hub → Sources & Connections: create source systems CRM and Billing.
  2. Create a connection for each (CSV files are fine for the lab) and Test both.
You know it worked when…
  • Both connections test green.
  • Each source shows its connection attached.
Walkthrough reference →

Map both sources onto Customer

  1. Mapping Designer: map CRM columns → Customer attributes (e.g. CustNo → customer_number, EmailAddr → email).
  2. Repeat for Billing with its column names.
You know it worked when…
  • Two saved mappings target the same Customer entity.
  • Every required attribute has a mapped source column in each mapping.
Walkthrough reference →

Standardize on the way in

  1. Add standardization rules: trim + title-case names, lower-case emails, normalise phone format, upper-case country codes.
  2. Preview the cascade before saving.
You know it worked when…
  • The preview shows raw → standardized values changing as intended.
Walkthrough reference →

Ingest and verify counts

  1. Create an ingestion job per mapping and run both.
  2. Open each run's history.
You know it worked when…
  • Rows read = rows loaded (zero rejects) — or you can explain every reject.
  • Staged data is visible for the entity.
Walkthrough reference →

Quality: profile, rule, remediate

  1. Profile the Customer entity.
  2. Author two rules (email pattern; customer_number present & unique) in a ruleset and run it.
  3. Open Exceptions and resolve (or knowingly ignore) at least one failing row.
You know it worked when…
  • A quality score exists with per-dimension breakdown.
  • The exception you worked shows its new status.
Walkthrough reference →

Match, survive, publish

  1. Create a match ruleset: exact on customer_number, fuzzy on name + email; thresholds so obvious duplicates auto-merge and borderline pairs go to review.
  2. Set survivorship: email by most-trusted-source (trust CRM higher), everything else most-recent.
  3. Run survive, review any queued pairs, then Publish.
You know it worked when…
  • Golden records exist — fewer than the raw row count (duplicates collapsed).
  • A golden record's detail shows per-field winners and xref sources.
  • Email values are masked for a user without Confidential clearance.
Walkthrough reference →

Report it

  1. Analytics → Report Studio: build a report on Customer golden records — count by country, with a drill-through.
  2. Publish it and open it in the viewer.
  3. Check Pipeline Health end to end.
You know it worked when…
  • The published report renders with your golden data.
  • Pipeline Health shows the chain green.
  • You just went from empty model → trusted, reported golden records. That is the whole discipline.
Walkthrough reference →

Need an environment? Request a hosted sandbox

Our team sets up a private MDM Studio sandbox for you or your team to run this lab for real.