Skip to content
Knowledge beta

Small Slices of Value

Principle Flow Lean

The fastest way to reduce cycle time without adding people or heroics is to make the work smaller. Thin vertical slices beat thick horizontal layers every time.

Break work into the smallest increment that delivers something a user or stakeholder can react to. A small slice is vertical; it cuts through all the layers (UI, logic, data) to produce a thin but complete piece of functionality. “Add the tax calculation for a single jurisdiction and deploy it” is a slice. “Build the database schema for all tax rules” is not; that’s a horizontal layer that delivers nothing on its own.

Small slices are the most reliable way to shorten cycle time without adding people or heroics. Little’s Law guarantees it: smaller items spend less time in the system, so WIP drops and throughput stays steady or improves. Combine small slices with WIP limits and you get a delivery rhythm where feedback arrives days after a decision, not months. That tight loop is what makes continuous delivery possible; you can’t ship continuously if each item takes three sprints to finish.

This principle becomes even more critical in AI-assisted development. When you’re generating code with an LLM, a small, well-defined slice gives the model a clear target and makes the output reviewable in minutes. A large, ambiguous task produces large, ambiguous code that is harder to verify and riskier to ship. Small slices keep humans in the loop where it matters: defining what to build, validating that it works, and deciding what to build next. The same feedback loop that makes slicing valuable for teams makes it essential when your co-developer is a model that doesn’t know your business context.

Resources