Skip to content

Product Development as Learning

The foundation guide for the Product Learning Series

Series identity. Learn in cycles. Move in a line. The product keeps moving forward, while understanding improves through repeated PDSA learning loops.

Use this guide to: establish the operating philosophy before a team starts applying the individual guides.

PDSA emphasis: the whole loop: Plan, Do, Study, Act, then begin the next Plan.

Purpose

This guide frames product development as a learning system, not a handoff from requirements to design to build. The team still moves work forward in a line: decisions are made, slices are shipped, and evidence accumulates. But the learning happens in cycles because product work contains uncertainty, variation, competing perspectives, and changing context.

When To Use It

  • At the start of a product, program, or major capability.
  • When a team is treating requirements as fixed instructions even though the problem is still uncertain.
  • When leaders need a concise explanation for why discovery, experiments, delivery, and adaptation belong in the same operating model.

Core Questions

  • What uncertainty is shaping this product decision?
  • What must move forward even while we are still learning?
  • What evidence would change our mind?
  • Where are we confusing a decision with a fact?
  • Which parts of the system create the customer outcome?

Facilitation Pattern

  1. Start with the phrase: Learn in cycles. Move in a line.
  2. Draw one line for product movement and one PDSA loop that repeats along it.
  3. Ask the team to name one current decision, one uncertainty, and one observable outcome.
  4. Separate direct lineage from helpful adjacent ideas so the method has intellectual clarity without becoming academic.
  5. Close by selecting the first guide or worksheet the team will use next.

Working Example

A payments team is asked to add a new dispute workflow. The linear work is real: define the first thin slice, build it, release it, and support it. The learning cycle is also real: the team predicts where support agents will hesitate, observes actual cases, studies the difference, and adapts the next slice.

Common Traps

  • Calling the approach agile while still treating early requirements as settled truth.
  • Using PDSA as a status ritual instead of a way to make and test predictions.
  • Letting background theory crowd out practical decisions.
  • Treating product learning as only user research or only delivery metrics.

Outputs And Artifacts

  • A shared series map.
  • A working definition of the next learning cycle.
  • An explicit distinction between knowns, assumptions, predictions, and decisions.
  • Agreement that guides are lenses to revisit, not waterfall stages to complete once.

PDSA Linkage

PDSA is the connective loop for the whole series. Plan names the learning objective and prediction. Do creates a small contact with reality. Study compares what happened with what was expected. Act changes the product direction, the team's understanding, or the next question.

Working Worksheet

  • State the product movement line in one sentence.
  • State the current PDSA cycle in one sentence.
  • List the top three uncertainties.
  • Write one prediction before action begins.
  • Name the evidence source for Study.
  • Name the decision that Act will inform.

Lineage Notes

The direct lineage runs through Shewhart and Deming. Shewhart's work on statistical quality control and process variation shaped Deming's view that improvement depends on theory, prediction, observation, and learning. Deming's PDSA wording deliberately emphasizes Study rather than merely Check. Juran, Ishikawa, and Feigenbaum are close quality contemporaries whose ideas complement the series: fitness for use, cause-and-effect thinking, quality circles, and total quality across the whole organization. Ackoff adds systems thinking and the idea that many real problems are interdependent messes. Drucker contributes management clarity around purpose, objectives, and customers. Ries and Blank are later product-development descendants, useful for experiments and validated learning, but not part of the original PDSA lineage.

References