The components, interfaces, dependencies and decisions in scope are described. For an existing system, the documents show which details have been checked against the implementation.
ATAM (Architecture Tradeoff Analysis Method)
A practical method guide: clarify prerequisites and materials, understand the complete process, and look up outputs, pitfalls, and stop criteria.
Prerequisite
What needs to be finished first
Product and project representatives can explain business goals and constraints. People from affected areas are available to contribute scenarios and answer questions.
The evaluation team knows ATAM and has experience in architecture analysis. Specialists and relevant documents or measurements are available for the quality goals under review.
Preparation
What needs to be ready before start
Versioned architecture documents, business goals, scenario list, findings log and access to the sources used.
An experienced evaluation team with a facilitator and note-taker; architecture, product and project representatives; people from affected areas. Specialists contribute expertise for the quality goals under review. One person can prepare documents and partial analyses; the full evaluation brings these roles together.
Agree the evaluation scope, business goals, architecture version and participants. Assemble an experienced evaluation team and plan two evaluation rounds with time for analysis between them.
Plan preparation, a first evaluation round, analysis between meetings, a second round with further participants and the report. Agree the duration based on system scope and analysis needs. The time entered for one step applies to that selected work block.
Set up a shared document with stable scenario and finding IDs. Label facts, assumptions, technical evidence and decisions separately.
Core question
The one question this method answers
How does the architecture support our most important quality goals, and where do risks or trade-offs arise?
Flow
Phases with goal, actions, output and pitfall
ATAM examines an architecture through concrete situations, such as a load spike or an outage. The evaluation team first works with the architects to understand which decisions support the most important quality goals. Other stakeholders then contribute their requirements. The result is a reasoned account of risks, trade-offs and open questions.
ATAM evaluation report
Fictional worked example
Record scenarios, reasoning and next steps so others can follow the assessment.
Evaluation frame
Describe scope, architecture version, business goals, roles and dates. Also identify areas outside the assignment.
Reservation service v0.8; expansion; latency and reliable allocation.
Utility tree and scenarios
Organize quality goals, specific requirements and scenarios in the utility tree. Explain the business importance and technical difficulty of each scenario.
S1: 500 bookings/min, 95 percent confirmed within 2 s. S2: zone failure, service restored within 60 s. Both scenarios: high importance and difficulty; fictional requirements.
Scenario analysis
For each scenario, record the decision examined, reasoning, evidence and assumptions. Record risks, decisions judged acceptable with supporting reasons (non-risks), sensitivity points and trade-off points.
R1: Evidence for zone failure is pending. The required number of confirming systems affects response time and failure tolerance. Basis: fictional design.
Additional scenarios
Record new scenarios, represented perspectives, prioritization and changes to findings.
Operations adds S3: repeated requests should lead to the same booking.
Risk themes and handoff
Record risk themes, business impact, limitations and follow-up work with owners and review dates.
Reliable bookings: development and operations examine zone failure and handling of repeated requests. Evidence is pending.
All examples are invented. Target values are requirements; expected effects and open evidence needs are explicitly marked.
- 1
Scope
Clarify the assignment, business goals and architecture

Goal
Establish an agreed basis for evaluation.
Inputs for this step
- Evaluation assignment
- Business goals and architecture description
Actions
- Before the first meeting, agree which parts of the system to evaluate, who participates, who receives the report and how to handle confidential information.
- Introduce ATAM and what its findings can establish. Ask product representatives to explain the business goals and constraints, known in ATAM as business drivers. The architecture team explains the system structure and key decisions.
- For an existing system, mark differences between the documented and implemented architecture.
Output
Scope, goals and architecture version
Show which quality requirements follow from the business goals. Identify architecture views that still need to be prepared.
Quality check
Is the architecture baseline unambiguous? Are confirmed business goals and assumptions still to be checked clearly labelled?
Pitfall
The review starts with a list of preferred technologies. Link each technology to a concrete architectural decision and explain which business goal it supports.
Timing: Preparation and shared opening; schedule according to system scope.
- 2
Approaches
Identify architectural approaches

Goal
Identify which architectural decisions to examine.
Inputs for this step
- Agreed architecture baseline
- Business drivers
Actions
- Name important approaches such as replication, decoupling, caching or access control and their intended effects.
- Show where each approach is used and which components or interfaces it affects. Record settings and assumptions that still need clarification.
Output
Relevant architectural approaches
For each product used, describe its purpose, how it is configured and its expected effect.
Quality check
Is the intended effect of each approach described? Can affected structures and parameters be located?
Pitfall
An architecture pattern is treated as evidence that a quality goal is met. Describe how the pattern is implemented and examine its expected effect against the relevant scenario.
Timing: First evaluation meeting; arrange separate follow-ups for missing details.
- 3
Utility tree
Turn quality goals into concrete scenarios

Goal
Select important and technically difficult scenarios.
Inputs for this step
- Business drivers
- Quality requirements
- Architectural approaches
Actions
- Build a utility tree with architecture and project representatives. Start with quality attributes such as availability, refine them into requirements and describe concrete situations as scenarios.
- Specify stimulus, context, affected system elements, expected response and measurable target.
- Rate each scenario by business importance and expected implementation difficulty. Explain which scenarios you will examine first.
Output
Prioritized quality scenarios
Keep importance and technical difficulty as separate ratings.
Quality check
Can each prioritized scenario be understood in its business context? Are importance and difficulty justified?
Pitfall
The tree ends with general words such as fast or secure. Replace quality adjectives with concrete stimuli, context and observable responses.
Timing: First evaluation meeting; refine with the appropriate specialists as needed.
- 4
Analysis
Analyze approaches against prioritized scenarios

Goal
Identify risks and trade-offs with clear reasoning.
Inputs for this step
- Prioritized utility-tree scenarios
- Architectural approaches and available evidence
Actions
- Walk through how the architecture would respond to each priority scenario. Examine the relevant quality goal and involve performance, security or operations specialists as needed.
- Record risks. Also document decisions that support the quality goal under the examined conditions. ATAM calls these non-risks. State the scenario, reasoning and assumptions for each.
- Mark parameters that strongly influence a quality goal as sensitivity points. A parameter that affects several quality goals is a trade-off point. Describe conflicts in particular, such as faster responses at the cost of lower resilience.
- For each finding, state whether it is based on documents, expert judgment, a calculation or a measurement. Agree suitable checks for claims that remain unresolved.
Output
Risks, non-risks, sensitivities and trade-offs
For non-risks, record the conditions under which the assessment holds. For open claims, specify the check still required.
Quality check
Does each conclusion link a decision, scenario and evidence? Are opposing effects and knowledge gaps visible?
Pitfall
Agreement is treated as proof of runtime properties. State the basis of the assessment and plan suitable models or measurements for open questions about runtime behavior.
Timing: First evaluation meeting and agreed interim work; schedule deeper technical analysis separately.
- 5
Scenarios
Add further scenarios and review findings

Goal
Test the initial analysis against further relevant perspectives.
Inputs for this step
- Initial analysis
- Participants from other affected areas
- Utility tree
Actions
- Invite representatives of users, business, development, operations and other affected groups to the second evaluation meeting.
- Collect additional scenarios, combine duplicates and agree a reasoned order of priority with the group.
- Analyze the highest-priority scenarios using the same approach. Extend and correct the initial findings.
Output
Additional scenarios and findings
Record represented perspectives and scenarios left outside the analysis.
Quality check
Are important stakeholder perspectives represented? Were new scenarios actually analyzed?
Pitfall
New scenarios are accepted only when they confirm the initial analysis. Examine contradictory and missing scenarios explicitly and record changes to findings.
Timing: Second evaluation meeting after agreed interim work.
- 6
Report
Hand over findings and follow-up work

Goal
Hand over findings so the responsible people can decide what to do next.
Inputs for this step
- Reviewed scenario analyses
- Risks, non-risks and tradeoffs
Actions
- Present the evaluation scope, quality goals, examined scenarios and findings. Explain which conclusions the results support and which questions remain open.
- Group related risks into themes and connect them to business goals. Preserve differing judgments.
- Agree next investigations or decisions, owners and review dates with the responsible people. Keep decisions distinct from review findings.
Output
Report with risk themes and next actions
The report can support both changing and retaining a decision. Record the resulting decisions separately.
Quality check
Can another reader follow the chain of evidence? Does each priority open question have an owner?
Pitfall
The report recommends replacing the architecture wholesale. Tie each recommendation to affected scenarios, expected benefit, uncertainty and required evidence.
Timing: Schedule result presentation and report preparation separately.
Artifact
What comes out at the end
A versioned analysis report covering scope, architecture, scenarios, findings, evidence, open questions and next steps with owners.
Record the architecture version, date, participants and evidence used for each assessment. Mark superseded findings and retain earlier versions.
- Shared document and board
- Architecture diagrams and repository
- Linked measurements or models according to analysis needs
atam-markdown.md
Prepare the overall initiative or a selected analysis block with a concrete question, documents and participants.
ATAM (Architecture Tradeoff Analysis Method): Worksheet
Question
To be confirmed
Desired outcome
To be confirmed
Scope
Whole analysis initiative Time: Multiple meetings and interim work, to be agreed according to scope
Complete this worksheet on paper or in your own document during the work.
Preparation for this scope
Method setup
- Materials: Versioned architecture documents, business goals, scenario list, findings log and access to the sources used.
- Roles: An experienced evaluation team with a facilitator and note-taker; architecture, product and project representatives; people from affected areas. Specialists contribute expertise for the quality goals under review. One person can prepare documents and partial analyses; the full evaluation brings these roles together.
- Advance information: Agree the evaluation scope, business goals, architecture version and participants. Assemble an experienced evaluation team and plan two evaluation rounds with time for analysis between them.
- Overall time needed: Plan preparation, a first evaluation round, analysis between meetings, a second round with further participants and the report. Agree the duration based on system scope and analysis needs. The time entered for one step applies to that selected work block.
- Setup: Set up a shared document with stable scenario and finding IDs. Label facts, assumptions, technical evidence and decisions separately.
Prepare for the work steps
Scope: Clarify the assignment, business goals and architecture
- Evaluation assignment
- Business goals and architecture description
Approaches: Identify architectural approaches
- Agreed architecture baseline
- Business drivers
Utility tree: Turn quality goals into concrete scenarios
- Business drivers
- Quality requirements
- Architectural approaches
Analysis: Analyze approaches against prioritized scenarios
- Prioritized utility-tree scenarios
- Architectural approaches and available evidence
Scenarios: Add further scenarios and review findings
- Initial analysis
- Participants from other affected areas
- Utility tree
Report: Hand over findings and follow-up work
- Reviewed scenario analyses
- Risks, non-risks and tradeoffs
ATAM evaluation report
Evaluation frame
Describe scope, architecture version, business goals, roles and dates. Also identify areas outside the assignment.
...
Utility tree and scenarios
Organize quality goals, specific requirements and scenarios in the utility tree. Explain the business importance and technical difficulty of each scenario.
...
Scenario analysis
For each scenario, record the decision examined, reasoning, evidence and assumptions. Record risks, decisions judged acceptable with supporting reasons (non-risks), sensitivity points and trade-off points.
...
Additional scenarios
Record new scenarios, represented perspectives, prioritization and changes to findings.
...
Risk themes and handoff
Record risk themes, business impact, limitations and follow-up work with owners and review dates.
...
Enter your own information, sources and assumptions in the blank worksheet.
SEI: Architecture Tradeoff Analysis Method
Scope: Clarify the assignment, business goals and architecture
Expected artifact: Scope, goals and architecture version
Show which quality requirements follow from the business goals. Identify architecture views that still need to be prepared.
Entry:
...
- Is the architecture baseline unambiguous?
- Are confirmed business goals and assumptions still to be checked clearly labelled?
Approaches: Identify architectural approaches
Expected artifact: Relevant architectural approaches
For each product used, describe its purpose, how it is configured and its expected effect.
Entry:
...
- Is the intended effect of each approach described?
- Can affected structures and parameters be located?
Utility tree: Turn quality goals into concrete scenarios
Expected artifact: Prioritized quality scenarios
Keep importance and technical difficulty as separate ratings.
Entry:
...
- Can each prioritized scenario be understood in its business context?
- Are importance and difficulty justified?
Analysis: Analyze approaches against prioritized scenarios
Expected artifact: Risks, non-risks, sensitivities and trade-offs
For non-risks, record the conditions under which the assessment holds. For open claims, specify the check still required.
Entry:
...
- Does each conclusion link a decision, scenario and evidence?
- Are opposing effects and knowledge gaps visible?
Scenarios: Add further scenarios and review findings
Expected artifact: Additional scenarios and findings
Record represented perspectives and scenarios left outside the analysis.
Entry:
...
- Are important stakeholder perspectives represented?
- Were new scenarios actually analyzed?
Report: Hand over findings and follow-up work
Expected artifact: Report with risk themes and next actions
The report can support both changing and retaining a decision. Record the resulting decisions separately.
Entry:
...
- Can another reader follow the chain of evidence?
- Does each priority open question have an owner?
Open questions and next steps
...
Method guide: https://methodatlas.meierhoff-systems.de/en/methods/atam/run-sheet
Example output
Concrete filled scenario, fictional example
Fictional example for orientation. Counts and timings describe this case and are not universal requirements.
atam-beispiel.md
Concrete filled scenario, fictional example
Fictional worked example: booking platform
A reservation service plans to expand into more regions. Bookings should be confirmed quickly and preserved during failures.
S1, high importance and high difficulty: At 500 bookings per minute, at least 95 percent are confirmed within 2 seconds. S2, also high/high: Confirmed bookings survive the loss of an operating zone. Service returns within 60 seconds. These are fictional requirements; evidence that they can be met is still pending.
Architectural approach: Booking data is stored on several systems before confirmation. Sensitivity point: The time needed for this transfer affects response time. Trade-off point: The required number of confirming systems, known as the quorum, affects response time and failure tolerance. Risk R1: The design describes behavior during a zone failure; evidence for S2 is pending.
The second evaluation round adds S3: A request retried after a timeout should lead to the same booking. The team must define how the service recognizes and handles retries. This property is called idempotency.
Next step: Development and operations plan a test with zone failure and repeated requests. The report records design reasoning, assumptions and planned measurements. The responsible people then use the findings to decide which changes are needed.
Pitfalls
Recognize symptoms and steer against them
Quick overview, the full text is in the matching phase in the flow
- Clarify the assignment, business goals and architectureThe review starts with a list of preferred technologies.View in phase
- Identify architectural approachesAn architecture pattern is treated as evidence that a quality goal is met.View in phase
- Turn quality goals into concrete scenariosThe tree ends with general words such as fast or secure.View in phase
- Analyze approaches against prioritized scenariosAgreement is treated as proof of runtime properties.View in phase
- Add further scenarios and review findingsNew scenarios are accepted only when they confirm the initial analysis.View in phase
- Hand over findings and follow-up workThe report recommends replacing the architecture wholesale.View in phase
Clarify the assignment, business goals and architecture
The review starts with a list of preferred technologies.
Link each technology to a concrete architectural decision and explain which business goal it supports.
Identify architectural approaches
An architecture pattern is treated as evidence that a quality goal is met.
Describe how the pattern is implemented and examine its expected effect against the relevant scenario.
Turn quality goals into concrete scenarios
The tree ends with general words such as fast or secure.
Replace quality adjectives with concrete stimuli, context and observable responses.
Analyze approaches against prioritized scenarios
Agreement is treated as proof of runtime properties.
State the basis of the assessment and plan suitable models or measurements for open questions about runtime behavior.
Add further scenarios and review findings
New scenarios are accepted only when they confirm the initial analysis.
Examine contradictory and missing scenarios explicitly and record changes to findings.
Hand over findings and follow-up work
The report recommends replacing the architecture wholesale.
Tie each recommendation to affected scenarios, expected benefit, uncertainty and required evidence.
Stop criteria
When to pause, resolve prerequisites, or choose another approach
Finished the method guide?
Go to the profile for purpose, similar methods, and sources or continue to the next method in the catalog.

