Module 08 Lesson 2 · 7 lessons in this module

Implementation Phases

In brief: Domain sequencing — deciding which data domain to tackle in which phase — is one of the highest-leverage decisions in MDM implementation planning. The pilot domain selected for Phase 1 sets the template for every subsequent domain: it establishes the governance process, tests the technology configuration, and produces the results that justify Phase 2…

Watch: How to sequence domains across implementation phases for maximum business and AI impact, why phase durations consistently run longer than planned, and how to build realistic timelines that maintain executive confidence when complexity emerges.

Module support notes

Sequencing Domains Across Phases

Domain sequencing — deciding which data domain to tackle in which phase — is one of the highest-leverage decisions in MDM implementation planning. The pilot domain selected for Phase 1 sets the template for every subsequent domain: it establishes the governance process, tests the technology configuration, and produces the results that justify Phase 2 investment.

Choosing a domain that is genuinely high-impact and AI-critical, while remaining achievable within the Phase 1 timeframe, requires honest assessment of data complexity, stakeholder readiness, and AI dependency urgency. Organizations that sequence domains based on technical convenience rather than business and AI priority consistently find that Phase 2 stakeholder engagement is harder to sustain because Phase 1 results weren't compelling enough.

Tip

Score candidate domains against four criteria before assigning them to phases: business impact, AI dependency urgency, data complexity, and stakeholder readiness. A domain that scores high on impact and AI urgency but low on complexity and readiness may be a better Phase 1 choice than one that is technically straightforward but carries little business consequence — because Phase 1 results need to be compelling enough to sustain the program through Phase 2.

A domain sequencing decision matrix with candidate domains — Customer, Product, Supplier, Employee, Location — as rows and sequencing criteria — business impact score, AI dependency urgency, data complexity rating, stakeholder readiness score, and composite priority — as columns. Domains are ranked by composite priority with Phase 1, 2, and 3 assignments shown.

Domain sequencing decisions made at the start of Phase 1 shape the entire program trajectory — invest time in getting them right.

Duration Realism in Phase Planning

Phase duration underestimation is the most consistent implementation planning failure in MDM programs — and one of the most damaging to stakeholder confidence when the timeline slips. Data quality is almost always worse than the profiling conducted before implementation begins, because profiling samples typically underrepresent the tail of the quality distribution where the most problematic records live.

Governance role recruitment and onboarding consistently takes longer than planned. AI pipeline integration complexity is particularly prone to underestimation because it often reveals integration gaps that weren't visible during architecture design. Building explicit contingency into every phase timeline — and communicating it transparently to executive sponsors rather than presenting best-case estimates — is a discipline that distinguishes experienced MDM implementation teams from those running their first program.

Warning

Build contingency into every phase timeline — and communicate it honestly to executive sponsors from the start. The five most common causes of Phase 1 delay are: data quality worse than profiling suggested (adds 4–8 weeks), governance role recruitment and onboarding (adds 2–6 weeks), IT resource contention (adds 2–8 weeks), stakeholder decision delays (adds 2–4 weeks), and AI pipeline integration complexity underestimated (adds 4–8 weeks). Budget for all five before presenting the timeline.

A comparison between planned and actual Phase 1 durations across five common causes of delay: data quality worse than profiling suggested adding 4–8 weeks; governance role recruitment and onboarding adding 2–6 weeks; IT resource contention adding 2–8 weeks; stakeholder decision delays adding 2–4 weeks; and AI pipeline integration complexity underestimated adding 4–8 weeks.

Build contingency into every phase timeline — and communicate it honestly to executive sponsors from the start.

Each implementation phase has a clear purpose, a defined set of deliverables, and explicit AI readiness checkpoints — all three need to be planned before implementation begins.

A three-phase implementation timeline showing deliverable lists, duration ranges, and AI readiness milestones for each phase, illustrating how the phases build on each other from foundation through expansion to optimization.

Each phase has a clear purpose, a defined set of deliverables, and explicit AI readiness checkpoints — plan all three before implementation begins.

Lesson progress

0% watched

← Previous Next lesson →