Define solution strategies
Define solution strategies (about 25–30%) tests decisions made before and around the build: implementation approach, application lifecycle management, deployment, data migration, and servicing. Answers are judged on fitness for the customer's constraints, namely geography, compliance, data volume, and timeline.
Implementation approach
Match the approach to the constraint. A single-entity, single-country rollout often fits a waterfall or hybrid plan; a multi-country programme usually fits phased waves built on a golden configuration template so each wave inherits a proven core. Agile fits when the customer can absorb incremental releases and owns backlog prioritization.
ALM
Under One Version, environments run the same platform and you receive continuous service updates. Source lives in Azure DevOps, builds produce deployable packages, and Lifecycle Services (LCS) deploys those packages. Never change a sandbox without promoting the change through source control.
Deployment and environments
Cloud is the default and the strategic direction; on-premises deployments remain available but lag cloud capability and constrain scenarios such as dual-write. Plan a Tier 1 developer environment, a shared Tier 2+ sandbox for integrated testing, and production.
Data migration
Treat migration as a project: profile and cleanse legacy data, map to data entities, load to staging, validate, then run mock cutovers with reconciliation counts and financial tie-outs. Sequence by dependency: master data, openings, then transactions.
Worked example
A manufacturer with legal entities in six countries wants one global design. Reasoning: build and stabilize the core in a Tier 2 environment, freeze it as a template, then run per-country waves with local legal-entity configuration, migration, and UAT before each go-live. A single big-bang across six countries multiplies cutover risk and yields no reusable learning between waves.
Traps
Treating ALM as a post-go-live concern; designing a different core per country; migrating uncleansed legacy data; using production as a test bed; ignoring the service update cadence when committing to a go-live date.