Read a sample: 1. The Release Evidence Problem
The deployment is scheduled for Friday evening. The build is green, the test team reports that the important scenarios passed, and the service desk has been told to expect the change. The delivery lead is ready to ask for the release decision.
Then a reviewer asks a straightforward question: “Which approved requirement does this test result demonstrate, and does the result apply to the version we are about to deploy?”
The team has the pieces of an answer. The requirement is in Azure DevOps. A tester has posted a screenshot in a chat. The build number appears in an email, and the release manager remembers discussing an exception at a meeting. What the team lacks is a dependable path between those pieces. Reconstructing it will take time, and some of the context may no longer be available.
This is the release evidence problem. A team can complete substantial delivery work without being able to demonstrate, clearly and consistently, what was requested, what changed, how it was checked, who made the relevant decisions, and what was deployed. In a regulated setting, that gap can delay a release or leave the organization unable to explain it later.
Evidence is not an administrative substitute for good engineering. A record of a test cannot make a defective change work, just as a successful deployment cannot establish that the right change was approved. Reliable delivery requires both the work and a credible account of the work.
A release review must answer two related but distinct questions.
The first is operational readiness: Is the change sufficiently prepared to enter its intended environment? The answer may depend on the software’s behavior, deployment procedures, support arrangements, dependencies, unresolved defects, and recovery plans. A team might reasonably examine whether the deployment has been rehearsed, whether support staff understand the change, and whether known issues are acceptable to the people who own the risk.
The second is evidence readiness: Can the organization substantiate the claims on which that operational decision depends? A reviewer needs to find the relevant records, understand their context, and determine whether they apply to the proposed release. “Testing passed” becomes more useful when the reviewer can see what was tested, the outcome, the version or environment involved, and how failures were handled.
Neither question replaces the other. A well-organized evidence package may describe a release that is not operationally ready. Conversely, a change may appear to work and still lack a necessary approval or a reliable connection between its test results and the artifact selected for deployment.
Evidence readiness is also relative to a defined set of expectations. Different organizations and changes can require different reviews, records, retention periods, and levels of independence. The delivery lead should not assume that an Azure DevOps configuration, a checklist in this book, or a familiar practice from another project automatically satisfies the obligations of the current organization. Those expectations must be identified and confirmed with the appropriate owners.
Throughout this book, we will follow a fictional organization called Harbor Health Services. Harbor operates a patient-services application used by staff to maintain patient contact information. Its delivery team works in a regulated environment and must follow internal procedures for changing software that handles sensitive information.
Harbor’s proposed change concerns the address-change requirement. Authorized staff need to update a patient’s mailing address and receive clear feedback when the submitted address does not meet the application’s validation rules. The release must preserve an understandable record of the change and its verification without exposing unnecessary patient information in delivery evidence.
The team has named its planned release Decision Review Release 1. Its immediate objective is to deliver the agreed address-change capability to the intended production environment after the applicable release criteria have been reviewed. The name identifies the release being assessed; it does not mean a decision has already been made or that approval is automatic.
This is one running project, not a role-change workflow. Harbor will need to consider who can perform the address update and who is authorized to make release decisions, but the software change being traced through the book is the address-change requirement. Access to the application and approval of the release are separate matters. A person may be permitted to use a feature without being the person who approves its deployment.
At the start of the story, Harbor has a request and an intended outcome, not a finished evidence trail. The detailed requirement, acceptance criteria, scope, tests, technical changes, and release decisions still need to be recorded and connected. Later chapters will build those records step by step. For now, the scenario gives us a way to examine the questions a delivery lead must ask before treating a release as ready.
Azure DevOps can store and connect delivery information, but it does not decide what Harbor’s evidence means. People establish the expectations, perform the work, and make decisions. The delivery lead’s first task is to understand whose contribution each claim depends on.
At Harbor, the business owner explains why staff need the address-change capability and confirms whether the proposed outcome meets the business need. A business analyst helps express that need as a requirement with acceptance criteria that the team can review and test. Developers implement the change and can explain the relationship between the work, code changes, and resulting build.
The test lead or designated testers plan and execute checks against the requirement. They record outcomes, identify…