Read a sample: 1. Start with Assessment, Not Assumptions
A migration discussion often begins with a useful but deceptively simple question: “Which applications should we move to Azure?”
The first answer may be a list of servers, a collection of application names, or a set of systems believed to be nearing the end of their useful life. Each can help start the conversation. None, by itself, tells the team enough about how an application is used, what it depends on, what data it handles, or who can make decisions about it.
An application that appears straightforward from a server inventory may exchange files with a supplier every night. A database thought to support one service may also feed several reports. A system described as “not critical” may become essential during a short monthly processing window. These are not reasons to avoid migration. They are reasons to understand the estate before choosing an approach.
This workbook is about building that understanding. It gives you a practical way to collect evidence, record uncertainty, and prepare a handoff for the people who will evaluate migration options. Azure is the intended destination under consideration, but the assessment method begins with the applications and the organization that relies on them—not with a predetermined cloud service or architecture.
A pre-migration assessment creates a defensible picture of the current application estate and the questions that remain open. It does not need to capture every configuration setting before a team can make progress. It does need enough structure to show what has been investigated, how the findings were obtained, and where decisions would otherwise rest on untested assumptions.
At the end of this workbook, you should be able to assemble a set of connected deliverables:
- An application inventory with clear names, purposes, environments, and ownership information.
- Application profiles that connect business context with technical components and operating needs.
- Dependency records showing known relationships and identifying paths that need verification.
- Summaries of data sensitivity, security, availability, recovery, and operational requirements.
- An evidence and unknowns register that makes confidence and follow-up responsibilities visible.
- A prioritized backlog for further assessment, reviewed against stakeholder knowledge.
These deliverables are related, but they answer different questions. An inventory tells you that an application exists and helps distinguish it from similarly named systems. A dependency record shows what it communicates with or relies on. An unknowns register prevents missing information from disappearing into meeting notes. A prioritized backlog helps the team decide what to investigate next.
The output is not a migration plan in disguise. An assessment may uncover information that influences scheduling, design, or cost estimates, but this book stops short of selecting a migration approach. It does not replace architecture design, detailed discovery, security review, financial analysis, implementation planning, or testing. Those activities can use the assessment findings, challenge them, and add detail.
Think of the assessment as an organized starting point for those later decisions. Its value lies not only in what it confirms but also in how clearly it exposes what the team does not yet know.
“Assess all applications” sounds comprehensive until someone asks whether a spreadsheet with macros counts, whether a retired test system belongs in the inventory, or whether a supplier-hosted service is in scope. Without a boundary, teams can spend time debating individual entries or overlook systems because everyone assumes someone else is covering them.
A useful initial boundary states:
- Organizational scope: Which business units, teams, or services are included?
- Application scope: Which applications and environments are included or excluded?
- Technology scope: Are supplier-hosted services, integrations, shared platforms, and end-user tools included as applications, dependencies, or both?
- Assessment depth: What information will be collected now, and what requires a later investigation?
- Time frame and decision purpose: When is the assessment being conducted, and what decisions will it inform?
- Authority: Who can clarify the scope, accept changes, and receive the findings?
The boundary can change. The important practice is to record the change and its consequences. If a shared identity service is initially outside the application inventory but appears as a dependency of several in-scope systems, it may need its own investigation. That does not mean the original boundary failed. It means the assessment produced information that made the boundary more useful.
Scope also protects against confusing assessment with implementation. For example, the team may record that an application uses a database and that its recovery requirements are unconfirmed. It need not design the database’s future Azure deployment to document those facts. If the current database configuration is uncertain, the assessment should say so rather than fill the gap with a proposed design.
To make the method concrete, this book follows Northstar Outfitters, a fictional regional retailer of outdoor equipment. Northstar sells through stores and an online channel. Its technology team supports a mixture of internally managed applications, shared infrastructure, and services provided by outside companies.
Northstar’s leaders want to explore whether selected applications could eventually run in Azure. They have not approved a migration sequence or chosen target services. The immediate assignment is narrower: assemble an evidence-based view of a small group of applications that support sales and inventory operations.
The initial assessment includes production applications used for online ordering and inventory management, along with their known nonproduction environments. Shared services and external providers are included as dependencies where they affect those applications. The team will record them, but it will not attempt a…