methodatlas
RunsheetArchitecture

Architecture Inception Canvas

ComplexityLow
Time1-2 h
Participants3-8
FormatWorkshop
MaturityEstablished
01

Prerequisite

What needs to be finished first

Complete firstArchitecture contextnot in catalog

There is a concrete reason for architecture work, such as a new system, a major redesign, or a platform launch, ideally with a business sponsor.

Without: Without a reason, the canvas stays academic and will not be used as a decision basis.
Complete firstStakeholder Mapping

The main stakeholders for the system are identified so that goals, constraints, and quality attributes come from the right side.

Without: Without stakeholder understanding, canvas fields are filled one-sidedly and important requirements are missing.
02

Preparation

What needs to be ready before start

Materials

Whiteboard or Miro board with canvas fields (goals, stakeholders, constraints, risks, quality attributes, system boundary, external interfaces, assumptions, next steps); stickies; template for ADR stub.

People / roles

One architect as owner; 3-6 participants from Engineering, Product, Operations, Security; one facilitator, ideally external but can be the architect; scribe.

Pre-read

Business goal of the system; relevant business requirements; existing technical landscape; known constraints (budget, time, technology, compliance); similar systems as reference.

Time needed

3-4 h

Setup

Put the canvas on the wall. Add a short explanation to each field. Rule: fill fields in sequence, not in parallel. Focus on a 10-page architecture, not a 100-slide deck.

03

Core question

The one question this method answers

Which minimal architecture decision covers the goals, constraints, and stakeholder needs without prematurely closing options?

04

Flow

Marker: Phase

StepDurationActionHint
1Phase 1: Goals and system purpose
20 minDescribe the system purpose in 2-3 sentences. Define the top 3 business goals and their metrics. Clarify what will change when the system is live.If the goals are only about better architecture, they come from Engineering. Demand business outcomes.
2Phase 2: Stakeholders and quality attributes
30 minList stakeholders with their main requirements. Derive the most important quality attributes from these requirements, such as performance, availability, security, and modifiability.If all quality attributes are high, the prioritization was not serious. Force ranking by the decider.
3Phase 3: System boundary and interfaces
30 minDefine what is inside the system and what belongs to the environment. List external interfaces, such as consumers, suppliers, and data sources.Clear boundaries prevent scope creep. If the boundary is fuzzy, requirements are often fuzzy too.
4Phase 4: Constraints, risks, assumptions
30 minList constraints (technology, budget, time, compliance), risks (technical, organizational, business), and assumptions explicitly. Mark assumptions that are critical for the architecture.Assumptions are often where architectures fail. Keep them explicit with date and owner so they can be revisited.
5Phase 5: Next steps and ADR stubs
30 minIdentify the top 3 architecture decisions that should become ADRs. Schedule the next workshops, for example QAW or Software Architecture Canvas. Assign an owner per stub.The canvas is inception, not the end product. Next steps must be scheduled explicitly, otherwise the artifact disappears.
05

Artifact

What comes out at the end

Form

Architecture Inception Canvas as an image plus Markdown with all fields filled with bullet points, tables, and short explanations. Supplemented by ADR stubs and a list of next workshops.

Versioning / ownership

Canvas with date and version. Changes need a rationale. For larger changes, create a new entry and archive the previous one. Link ADR stubs.

Tool alternatives
  • Miro or FigJam with canvas template
  • Confluence page with canvas structure
  • Markdown in the docs/architecture folder
  • Notion with database fields for recurring canvases
  • Lucidchart or Whimsical for formal diagrams

architecture-inception-canvas-working-template.md

Compact working template for Architecture Inception Canvas with context, input, output artifacts, and next step.

Architecture Inception Canvas 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

  • Inception Canvas:
  • Risks:
  • Goal List:
  • Open Questions:

Open questions

  • ...

Next step

Owner, date, success signal.

06

Example output

Concrete filled scenario, fictional example

architecture-inception-canvas-beispiel.md

Concrete filled scenario, fictional example

Architecture Inception Canvas - Workspot Booking API v2, May 2026

System purpose: Booking API that supplies Workspot and partner platforms with consistent availability and booking data.

Goals

  • Scale to 5x booking volume within 12 months.
  • Partner integration in under 5 person days per partner.
  • p95 below 300 ms for search, below 600 ms for booking.

Stakeholders and top requirements

  • Internal web/mobile: low latency, GraphQL preferred.
  • Partner: REST with OAuth2, sandbox environment.
  • Compliance: GDPR data residency EU.
  • SRE: multi-region HA, RPO 5 min.

Quality attributes (prioritized)

  1. Availability (99.95% SLA)
  2. Performance (see goals)
  3. Modifiability (partner onboarding)
  4. Security (OAuth2, rate limiting)
  5. Observability

System boundary: Booking API plus datastores. Auth service external. Notification service external.

External interfaces

  • Identity service (REST)
  • Payment provider (webhook)
  • Workspace owner app (GraphQL)
  • Partner API clients (REST with OAuth2)

Constraints

  • Cloud provider AWS, region eu-central-1.
  • Existing Postgres 14 database, migration in phase 2.
  • Phase 1 budget: EUR 350k.

Risks

  • Partner volume spike at launch is hard to predict.
  • Existing coupling to the Inventory Service DB prevents multi-region.

Assumptions

  • Partners respect the rate limit. Revisit by 30.07.
  • Auth service can scale horizontally. Owner @marcus, validate by 15.06.

Next steps

  • Quality Attribute Workshop on 25.05.
  • Software Architecture Canvas on 02.06.
  • ADR-014 multi-region strategy, owner @lisa.
  • ADR-015 GraphQL vs. REST internal, owner @ben.
07

Pitfalls

Recognize symptoms and steer against them

Trap

Canvas turns into a 100-slide architecture

Symptom

Fields become mini-documents with detailed diagrams.

What to do

The canvas is inception, every entry should stay within 1-3 bullets. Move details to follow-up documents, such as ADR or SAC. Keep the facilitator disciplined.

Trap

Assumptions stay implicit

Symptom

The discussion relies on unstated assumptions that later blow up.

What to do

Force the assumptions field to be filled. Give each assumption a validation owner and a date. Plan a spike for critical uncertainty.

Trap

Gold solution upfront

Symptom

The architect arrives with a finished picture and the workshop is just a show.

What to do

Keep options open and discuss at least two alternatives for each top decision. The canvas is for option openness.

Trap

No follow-up ADRs

Symptom

The canvas is filled, but nothing is decided and architecture work stalls.

What to do

Identify the top 3 decisions, create ADR stubs, and assign owner plus due date. Without this step, the canvas is not complete.

Trap

Stakeholders are missing

Symptom

Only Engineering is in the room and the quality attributes become technical only.

What to do

At least Product, Operations, and Security need to be represented. If that is not possible, run separate interviews first and bring the results to the workshop.

08

Stop criteria

Done signals checkable in under a minute

No business sponsor is present, so goals and constraints are not stable.
Stakeholder breadth is not covered, so critical requirements would be missing.
Architecture decisions are already made, so the canvas would only be theater.
The workshop is compressed to under 90 min, so the canvas remains incomplete.
There is no willingness to create ADRs as follow-up artifacts, so the canvas will decay.
The system will be replaced within 6 months anyway, so inception is not worth it.

Finished the runsheet?

Go to the profile for purpose, similar methods, and sources or continue to the next method in the catalog.