liveBook · Manning

Technical decision

As explained in Decision Engineering

A choice among technical options made to serve a specific business outcome while accounting for risks and uncertainty.

A technical decision is a choice among options made to serve a specific business outcome. A good decision considers the available options and known risks, rather than treating technical preference as sufficient. Its quality rests on the reasoning used at the time: uncertainty remains, and even a well-reasoned choice may not produce the hoped-for result. section 1.5

Match the effort spent deciding to the cost of being wrong:

  • For a two-way-door choice that is relatively easy to reverse, try an option, observe the result, and adjust. A frontend component’s layout or a logging library in one microservice may warrant speed rather than extended debate. section 1.1.2
  • For a one-way-door choice that is expensive or disruptive to reverse, use more careful analysis and involve stakeholders. Data storage architecture, a multi-year SaaS contract, or a company-wide architecture standard can affect an organization for years. section 1.1.2 section 1.5

Test before committing

Before adopting a library across an entire codebase, use it in a small isolated module. The trial can expose constraints while the choice is still easy to revisit. Broader adoption then calls for a decision at the appropriate scale, rather than treating a successful trial as the whole decision. section 1.1.2

Figure 1.1 Decisions that require structured analysis sit in the upper quadrants: high cost of change or high organisational impact.

Use cost of change and organizational impact to judge when structured analysis is warranted; decisions high on either dimension sit in the figure’s upper quadrants. section 1.1.2

A repeatable decision lifecycle frames the choice, evaluates and selects options, then makes the choice stick in execution. It is not necessarily linear: a failed prototype or newly discovered constraint can prompt reframing, while difficulty putting a choice into practice can prompt renewed evaluation. section 1.5

How Decision Engineering covers it

How deep it goes
Distinguishes reversible from hard-to-reverse choices and outlines an iterative three-phase lifecycle without prescribing a technology.
Examples
Contrasts component layout and logging-library choices with data storage architecture, SaaS contracts, and company-wide architecture standards.
What it assumes
Assumes competence in building software while developing decision-making as a discipline distinct from implementation skill.
What you can do after
Practice/perform decisions with analysis proportionate to their stakes, then revisit the framing or evaluation when new evidence warrants it.
How it connects
Links architectural decision-making to business outcomes, risk, prototyping, stakeholder involvement, and the decision lifecycle.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Decision Engineering. The section links point to the parts of the book it draws on.