Basic understanding of TOC, especially constraint logic and Five Focusing Steps, is present in the team or conveyed in the introductory briefing.
Theory of Constraints Thinking Process
Prerequisite
What needs to be finished first
A list of concrete Undesirable Effects (UDEs) with evidence exists (typically 5-15 items) that may follow a shared causal structure.
Preparation
What needs to be ready before start
Large whiteboard or digital diagram tool with symbols (boxes for statements, arrows for causality, ellipses for AND links); templates for CRT, Evaporating Cloud, FRT, Transition Tree; UDE list; Categories-of-Legitimate-Reservation checklist.
One TOC coach experienced in the Thinking Process; 4-8 people with domain knowledge; one scribe for assumptions and reservations; sponsor with implementation mandate.
UDE list with evidence per item; known causal hypotheses; stakeholder overview; previous improvement attempts and their results; Categories of Legitimate Reservation as poster.
Workshop series 2-5 days, followed by iterative refinement over weeks
Make the Thinking Process tools visible as phases: 1) Current Reality Tree, 2) Evaporating Cloud, 3) Future Reality Tree, 4) Prerequisite Tree, 5) Transition Tree. Display symbols as poster. Keep Categories of Legitimate Reservation (Clarity, Existence, Causality, Cause Insufficiency, Additional Cause, Predicted Effect, Tautology) ready for checking trees.
Core question
The one question this method answers
Which core conflicts and root causes generate the observed UDEs, which changes resolve them, and which prerequisites and transitions are needed to implement the changes?
Flow
Marker: Phase
| Step | Duration | Action | Hint |
|---|---|---|---|
1Phase 1: Current Reality Tree | 1-2 days | Place UDEs at the bottom. Build causal chains upward per UDE until root causes are reached. Merge multiple UDEs and identify shared root causes. Check Categories of Legitimate Reservation for each link. | CRT becomes large and hard to read. Required: check links logically, do not accept them intuitively. If a reservation applies, rephrase the link or add an intermediate element. |
2Phase 2: Evaporating Cloud | 0.5-1 day | Represent the core conflict from the CRT as a cloud: common objective, both necessary conditions, both demands, conflict. Make assumptions behind every connection explicit. Falsify at least one assumption to resolve the conflict. | Evaporating Cloud requires at least five assumptions per connection. Stopping at three assumptions means stopping too early. Assumption falsification is the lever. |
3Phase 3: Future Reality Tree | 1 day | Put the injection (the change from the cloud) at the bottom. Build logical consequences until Desired Effects (DEs) are reached. Identify Negative Branches (unintended negative effects) and trim them with additional injections. | FRT without Negative Branches is dangerous. Explicitly test at least 2-3 possible negative effects per injection. Trim through supplementary measures, not by ignoring them. |
4Phase 4: Prerequisite Tree | 0.5 day | For each injection, define which prerequisites must be fulfilled before implementation is possible. Identify obstacles and derive Intermediate Objectives. Establish logical sequence. | Prerequisite Tree shows implementation realism. If prerequisites are unrealistic, rethink the injection or clarify stakeholder buy-in first. |
5Phase 5: Transition Tree | 0.5-1 day | Detail implementation steps: need, action, effect per step. Add owner and date. Transition into ongoing steering. | Transition Tree is often skipped and the method stops at FRT. Without TT, change remains a concept without implementation path. |
Artifact
What comes out at the end
Current Reality Tree with UDEs, links, and root causes; Evaporating Cloud with assumptions; Future Reality Tree with injections, Desired Effects, and Negative Branches; Prerequisite Tree with IOs; Transition Tree with implementation steps; CLR review note per tree.
Own collection per project with date, sponsor, and coach. Version trees. On implementation iteration, create a new version with diff. Preserve assumption logic so falsification can be repeated later.
- Specialized tools such as Harmony for TOC Thinking Process
- draw.io or Lucidchart with custom TOC stencil
- Whiteboard with photographed trees
- Markdown with embedded diagrams under docs/toc-tp/<date>/
theory-of-constraints-thinking-process-working-template.md
Compact working template for Theory of Constraints Thinking Process with context, input, output artifacts, and next step.
Theory of Constraints Thinking Process Canvas
Context
What is this method used for?
Core question
Which question should be answered at the end?
Input
Which data, observations, or materials are available?
Working area
- Area 1:
- Area 2:
- Area 3:
- Relationships / patterns:
Output artifacts
- Current Reality Logic:
- Conflict Cloud:
- Future Reality Tree:
- Implementation Obstacles:
Open questions
- ...
Next step
Owner, date, and success signal.
Example output
Concrete filled scenario, fictional example
theory-of-constraints-thinking-process-beispiel.md
Concrete filled scenario, fictional example
TOC Thinking Process - Engineering Tribe Aurora, 2026-05-22
Coach: @marlene. Sponsor: VP Engineering.
UDEs (excerpt of 11)
- UDE-1: Releases take 18 days on average instead of planned 7.
- UDE-3: Engineers report 40% of time in waiting states.
- UDE-7: Bugfix response time for P1 above 24 h.
CRT (excerpt)
Root Cause: Shared branches with long-lived feature flags, shared database schemas between tribes. Linked with UDE-1, 3, 7 through conflict lines.
Evaporating Cloud
- Common Objective: High throughput with high quality.
- Need A: Early integration -> demand: shared main branch.
- Need B: Stable releases -> demand: long feature branches.
- Conflict between demands.
- Falsified assumption: "Stability requires long branches" is false if trunk-based development with feature flags and CI tests exists.
Injection
Introduce trunk-based development, short branches, automated tests blocking merge on failures.
FRT (excerpt)
Injection -> shorter cycle time -> faster releases (DE for UDE-1). Negative Branch: test flakiness could block trunk. Trimming: test stabilization as additional injection.
Prerequisite Tree
Prerequisites: CI pipeline reliability above 95%, feature-flag system, engineering training. IOs: pipeline health audit, flag-service migration, workshop series.
Transition Tree (excerpt)
- CI pipeline audit (Owner @ben, by 2026-05-30).
- Pipeline stabilization (Owner @ben, by 2026-06-30).
- Feature-flag-service migration (Owner @anna, by 2026-07-31).
- Training workshop (Owner @sabine, 2026-08-15).
- Pilot with two teams (from 2026-08-22).
- Rollout (from 2026-09-30).
Pitfalls
Recognize symptoms and steer against them
CRT without CLR check
Trees are accepted intuitively, logical gaps remain undiscovered.
Check Categories of Legitimate Reservation for every link. Be strict at least on Clarity, Existence, Causality. If a reservation applies, adjust the link.
Evaporating Cloud resolved too early
Assumptions are limited to three, the first plausible falsification is chosen, conflict seems solved.
At least five assumptions per connection. Falsification must be justified and evidenced in the concrete context, not just theoretically.
FRT without Negative Branches
Future Reality Tree only shows positive effects, risks remain unnamed.
Required: name 2-3 Negative Branches per injection and trim them with additional injections. Anyone finding no Negative Branches is not checking honestly.
Prerequisite and Transition Tree skipped
Method ends at FRT, implementation plan is missing, change fails.
Phases 4-5 are not optional. Implementation requires explicit prerequisites and step sequence. Otherwise change remains concept.
Trees too large and immobile
CRT fills three walls, nobody can use it anymore, method collapses under its own weight.
Aim for at most 30-40 nodes per tree. With more UDEs, split into clusters with a separate CRT each. Coach maintains discipline.
No external validation
Trees are developed and accepted by the same team, blind spots remain.
Use an external or at least cross-functional person as CLR reviewer. For strategic changes, involve a coach with method distance.
Stop criteria
Done signals checkable in under a minute
Finished the runsheet?
Go to the profile for purpose, similar methods, and sources or continue to the next method in the catalog.