Solution Atlas
EverydayUser storyConsultative playbook

We have customer data in five systems and no unified view

Marketing, service, and sales teams work from different customer datasets and reach different conclusions about the same customer. The CDO needs a unified customer view that flows from a governed substrate, with segmentation and operational dashboards each team can trust.

Trigger
Conflicting customer numbers across teams; commercial pressure on cross-sell.
Good outcome
Customer 360 model on Fabric, Purview-catalogued, segmentation dashboards live for marketing and service teams.
Diagnostic discovery

Signals this story fits

Observable cues that confirm the conversation belongs here.

  • ·Marketing, service, and sales reach different conclusions about the same customer
  • ·Customer data in 5+ systems with no unified key
  • ·CDO has named "customer 360" as a priority initiative
  • ·Cross-sell or churn programmes blocked by data fragmentation
  • ·No catalogue — analysts cannot find or trust customer datasets

Questions to ask

Open-ended, SPIN-style — each one has a reason it matters.

  1. 1.How many "customer" datasets exist across the organisation, and how do you reconcile them?

    WhySizes the unification scope.

    Listen for: “CRM is one, billing is another, support has its own” · “we use Excel reconciliations” · “we know there are duplicates but cannot fix them”

  2. 2.When marketing and service argue about a customer count, who wins, and how is the dispute resolved?

    WhySurfaces the trust and ownership gap.

  3. 3.Is there a master data management capability today — MDM tool, golden record, customer ID strategy?

    WhyDetermines whether the engagement is greenfield or improvement on existing MDM.

  4. 4.Which teams are the first consumers — marketing, service, sales, product, or all?

    WhySequences the deliverable. Marketing and service usually first; product later.

  5. 5.What governance posture is needed — GDPR consent, marketing opt-out, do-not-contact lists?

    WhyDrives Purview catalogue and consent strategy. Consent often dictates which customers can be included in segments.

  6. 6.What is the operational use beyond reporting — segmentation activation, journey orchestration, real-time triggers?

    WhyStretches the value beyond analytics into Customer Insights / journey orchestration.

Baseline → target architecture

TOGAF-style gap framing — what we typically see today, and what the proposed end state looks like. The gap between them is the engagement.

Baseline architecture

Customer data lives in CRM, billing, support, marketing automation, and product systems with no unified key. Each team builds its own customer view. Excel reconciliations attempt to bridge the systems but drift quickly. No catalogue — analysts spend weeks finding and trusting customer datasets. Different teams give the board different customer counts.

Typical concerns

  • ·Different customer counts from different teams
  • ·Cross-sell programmes cannot identify the right targets
  • ·GDPR consent not unified — risk of contacting opted-out customers
  • ·No single ID — duplicates and matching failures
  • ·Catalogue absent — datasets undiscoverable

Capability gaps

  • ·Unified customer entity with stable ID
  • ·Match-and-merge logic for source duplicates
  • ·Fabric customer 360 semantic model
  • ·Purview catalogue with lineage to source systems
  • ·Consent and DNC integration
Target architecture

Microsoft Fabric ingests customer data from CRM, billing, support, marketing, and product systems. A match-and-merge layer produces a unified customer entity with stable ID and source lineage. A certified Power BI semantic model defines the customer 360 metrics and segments. Purview catalogues every customer dataset with lineage to source. Consent and do-not-contact information unified at the customer entity level. Marketing and service work from the same view — disputes go away.

Key capabilities

  • Unified customer entity with stable ID
  • Match-and-merge with source lineage
  • Certified customer 360 semantic model
  • Purview catalogue and lineage
  • Consent and DNC unification

Enabling SKUs

Resolved in the ‘Recommended cards’ section below.

Architecture decisions

Each decision is offered as explicit options with trade-offs — Hohpe's “selling options” principle. A safe default is noted where one exists.

  1. Decision 1.Match-and-merge approach — Fabric medallion vs Dynamics 365 Customer Insights vs third-party MDM

    Fabric medallion (bronze/silver/gold)

    When it fitsFabric strategic substrate; match-and-merge logic acceptable in code/notebooks.

    Trade-offsBuild effort; match logic needs analyst skill.

    Dynamics 365 Customer Insights — Data

    When it fitsMarketing-led customer 360; identity resolution + segmentation as a packaged capability.

    Trade-offsSeparate licence; tighter to marketing use cases.

    Third-party MDM (Informatica, Reltio, etc.)

    When it fitsEnterprise-wide MDM scope beyond customer; existing investment.

    Trade-offsIntegration to Fabric needed; longer implementation.

    Default recommendationCustomer Insights — Data for the trial when marketing is the lead consumer. Fabric medallion when the customer 360 sits inside a broader data substrate programme.

  2. Decision 2.Consent and DNC handling — unified at customer entity vs source-of-truth-per-channel vs hybrid

    Unified at customer entity

    When it fitsStrong governance posture; consent owned centrally.

    Trade-offsSource systems need to read the unified consent; integration overhead.

    Source-of-truth-per-channel

    When it fitsChannel teams own consent independently; CRM = sales consent, marketing automation = marketing consent.

    Trade-offsRisk of inconsistent decisions; harder to defend to regulators.

    Hybrid — unified read, source write

    When it fitsMost customers. Source systems remain the write surface; the unified entity is the consolidated read.

    Trade-offsLatency in propagation; requires clear rules on precedence.

    Default recommendationHybrid for the trial. Move toward unified-at-entity in phase two if regulator exposure justifies the integration effort.

  3. Decision 3.First consumer — marketing segmentation vs service operational dashboards vs both

    Marketing segmentation first

    When it fitsCross-sell or retention campaign waiting on customer 360; marketing has clear ROI.

    Trade-offsService capability deferred.

    Service operational dashboards first

    When it fitsService team is the customer-facing pain point; better experience is the value.

    Trade-offsMarketing value deferred.

    Both in parallel

    When it fitsBoth teams ready; capacity to deliver both.

    Trade-offsLarger trial scope; sequencing risk.

    Default recommendationMarketing segmentation first — clearer ROI for the trial — with service operational dashboards in phase two.

Low-risk trial — proof of value

90-day customer 360 trial on Fabric + Purview

12 weeks

Customer data ingested from three source systems (CRM, billing, one operational) into Fabric. Match-and-merge logic produces a unified customer entity for a subset of the customer base (typically one country or one segment, ~50–100k customers). Certified Power BI semantic model with top 15 customer 360 metrics. Purview catalogues the source datasets and the unified entity with lineage. One marketing segment built and activated to the marketing automation platform as proof.

Success criteria

  • Unified customer entity for the trial subset with stable ID
  • Match-and-merge accuracy validated by business sample (>95%)
  • Top 15 customer 360 metrics agreed and certified
  • One marketing segment activated end-to-end as proof

InvestmentFabric capacity F4/F8 + Power BI Pro + Purview consumption. Estimated ~€8–15k/month for the trial scope. Existing customer dashboards continue running in parallel as the control.

Proof metrics

  • ·Single certified customer count agreed across marketing, service, and sales
  • ·Match-and-merge accuracy >95% on business sample
  • ·Cross-sell pilot segment activated with measurable response
  • ·Catalogue adoption — analysts discover datasets through Purview in the trial period

Recommended cards

The SKUs and capabilities most likely to be part of the solution, with the editorial rationale for each in the context of this story. Add the ones that fit your situation.

Why for this story

The substrate for ingestion, match-and-merge, and the certified semantic model. OneLake removes the copy-and-load tax that fragments customer data in the first place.

Why for this story

The certified semantic model and segmentation dashboards live on Power BI. The surface marketing and service use to operate the unified view day-to-day.

Why for this story

Without a catalogue, customer 360 becomes another silo. Purview gives the analysts and the data owners the discovery and lineage layer that makes the unified entity defensible.

Back to Customer & operational analytics