methodatlas
RunsheetTeam Design

Team Topologies Interaction Modes

ComplexityMedium
TimeHalf day
ParticipantsTeam-Leads plus Architecture
FormatWorkshop
MaturityEmerging
01

Prerequisite

What needs to be finished first

Complete firstTeam API

Teams have clarity about own responsibilities and services, ideally documented in Team API form.

Without: Without self-understanding per team, interaction modes are chosen for unclear interfaces and agreements stay vague.
02

Preparation

What needs to be ready before start

Materials

Whiteboard or Miro board with heatmap template (team list vertically and horizontally, cells for interaction modes); definitions of three modes with examples; list of current interface problems; pens with three colors for modes.

People / roles

One facilitator with Team Topologies experience; team leads of all involved teams; architecture lead or engineering director as sponsor; scribe for agreements.

Pre-read

Team list with main responsibilities (Stream-Aligned, Platform, Enabling, Complicated-Subsystem); current interface relationships; known frictions or escalations; planned platform initiatives.

Time needed

Half day (4 h)

Setup

Heatmap with teams in rows and columns. Colors per mode: blue=Collaboration (intensive, temporary), green=X-as-a-Service (clearly defined service, durable), yellow=Facilitating (coaching, enabling). Definitions visible.

03

Core question

The one question this method answers

Which interaction mode fits per team pair, and how do we agree modes bindingly to reduce interface friction?

04

Flow

Marker: Phase

StepDurationActionHint
1Phase 1: Introduce modes with examples
20 minExplain three modes. Collaboration: two teams work closely together temporarily (for example new feature, architecture change). X-as-a-Service: one team consumes clearly defined service of another. Facilitating: enabling team supports stream-aligned team through coaching, durably bounded.If modes are unclear, participants choose wrong. Bring concrete own examples, not abstract theory.
2Phase 2: Map current interaction
60-90 minMark current state per team pair. Which mode runs today. If unclear, mark. Mark friction points with red note.Honestly describe what runs, not what ideally should run. Mismatch between target and actual is common source of friction.
3Phase 3: Discuss target modes
60-90 minDefine target mode per team pair. Discuss rationale. For change between actual and target, think transition plan (for example Collaboration -> X-as-a-Service through service buildup).Avoid mixed forms. Exactly one mode per pair. Anyone saying "sometimes Collaboration, sometimes Service" has not understood mode. Conscious choice is core.
4Phase 4: Agreements and heatmap
30-45 minComplete heatmap with target modes. Document agreement per pair (what mode means concretely, which communication channels, SLA). Transition plan for changes.Heatmap without agreements is diagram. Concrete agreement per pair with example workflow. Agreement should be 1-2 sentences, not abstract definition.
05

Artifact

What comes out at the end

Form

Heatmap with all team pairs and modes (colored); table list with pairs, mode, agreement, transition plan, owner. Linked with Team API pages of involved teams. Visible engineering-org overview.

Versioning / ownership

Re-run semiannually or on org change. Mode change with date and rationale. Make transition phases visible (for example "Q3: Collaboration, Q4: X-as-a-Service targeted"). Archive previous version.

Tool alternatives
  • Miro or FigJam with Interaction Modes template
  • Lucidchart with heatmap template
  • Confluence page with embedded matrix
  • Notion database with mode/agreement properties
  • Specialized tool such as Plandek with topology features

team-topologies-interaction-modes-working-template.md

Compact working template for Team Topologies Interaction Modes with context, input, output artifacts, and next step.

Team Topologies Interaction Modes 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

  • Interaction heat map:
  • Agreements by pair:

Open questions

  • ...

Next step

Owner, date, success signal.

06

Example output

Concrete filled scenario, fictional example

team-topologies-interaction-modes-beispiel.md

Concrete filled scenario, fictional example

Team Topologies Interaction Modes - Engineering org, 2026-05-18

Sponsor: @julia (CTO). Participants: 8 team leads, Architecture Lead.

Teams

  • Stream-Aligned: Product A, Product B, Product C
  • Platform: Identity, Data Platform, Infrastructure
  • Enabling: DevEx, Security Coaches
  • Complicated-Subsystem: AI Platform

Heatmap excerpt

Team pairCurrentTargetAgreement
Product A <-> IdentityCollaboration (chaotic)X-as-a-ServiceIdentity provides Auth API with SLA, Product A consumes without direct contact.
Product A <-> DevExFacilitatingFacilitatingDevEx coach supports Product A with tooling migration, durable 0.5 FTE for 2 quarters.
Product B <-> Data PlatformCollaborationCollaboration (Q3) -> X-as-a-Service (Q4)Close collaboration for data pipeline buildup until Q3, self-service from Q4.
Product B <-> AI PlatformCollaborationX-as-a-ServiceAI Platform offers inference API, Product B integrates without direct model discussion.
Identity <-> InfrastructureX-as-a-ServiceX-as-a-ServiceInfrastructure provides Kubernetes cluster with clear SLA.
Security Coaches <-> allFacilitatingFacilitatingSecurity reviews on demand, max 2 days per team per quarter.

Previous friction points

  • Product A <-> Identity: every new feature required Slack DM discussions with Identity engineers (Collaboration without SLA). Solution: Auth API docs improved, SLA fixed, no direct engineer contact needed.
  • Product B <-> Data Platform: expectation of service availability, but platform still in build-up. Solution: transition plan with clear Q4 date for service mode.

Transition agreements

  • Q3 -> Q4: Product B <-> Data Platform switches to X-as-a-Service. Preconditions: self-service docs, API stability, SLA definition by Data Platform by end of Q3.

Next review: 2026-11-18 (semiannual).

07

Pitfalls

Recognize symptoms and steer against them

Trap

Mixed forms per pair

Symptom

Team pair marked with "Collaboration and X-as-a-Service depending on topic", mode unclear.

What to do

Exactly one mode per pair. If mixture desired, split pair into two sub-areas (for example "feature build: Collaboration, maintenance: Service"). Force clarity.

Trap

Wishful thinking instead of actual state

Symptom

Heatmap shows target as actual, frictions not visible.

What to do

Phase 2 is actual mapping. Honestly describe what runs. Target comes in phase 3. Mismatch between actual and target is starting point for actions.

Trap

Transitions without plan

Symptom

Change from Collaboration to X-as-a-Service agreed, but prerequisites (service buildup, docs, SLA) missing.

What to do

Transition plan with prerequisites and dates. "X becomes service from Q4" without preparation is wish. Preparation must be explicit.

Trap

Enabling teams become operations

Symptom

Enabling team (for example DevEx) embedded permanently in product teams, loses coaching character.

What to do

Facilitating is time-bounded. Enabling engagement with start and end date, max 2 quarters. For extension, explicit re-agreement.

Trap

Heatmap without follow-up actions

Symptom

Diagram built, sits in wiki, nobody implements mode changes.

What to do

Owner and deadline per pair with mode change. Semiannual review cadence. Workshop ends with concrete action plan, not diagram alone.

08

Stop criteria

Done signals checkable in under a minute

Fewer than 4 teams, heatmap would be overhead.
Teams have no clear responsibilities, mode choice would be speculation.
Sponsor or engineering lead not present, agreements would be non-binding.
Org in transition (restructuring), heatmap would be outdated in 4 weeks.
Three modes not understood, workshop becomes method training instead of topology design.
No existing interface problems, workshop is academic without lever.

Finished the runsheet?

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