Best practices
Frameworks
Connect your resources, activities, and intended results.
16 minute read · Reviewed September 2, 2026
What a useful nonprofit framework does
A framework is a bounded representation that helps a group ask consistent questions and make a decision. A systems map explores context and relationships. A theory of change explains how and why actions are expected to contribute to change. A logic model organizes resources, activities, direct outputs, and sequenced outcomes. A responsibility map clarifies ownership and participation. A learning cycle turns uncertainty into questions, evidence, interpretation, and adaptation. These tools can connect, but they are not interchangeable and none proves that a program works.
Why it matters
- A shared model lets staff, participants, board members, partners, funders, and evaluators identify where they agree, where definitions differ, and which links remain assumptions.
- Separating resources, activities, outputs, outcomes, and longer-term contribution prevents delivery volume from being presented as community change.
- Naming context and assumptions makes a program easier to pilot, measure, challenge, and adapt without pretending the organization controls every condition around it.
- Choosing a framework by decision need reduces template theater: the team can use the smallest useful tool, keep a record, and stop when the decision is clear enough to act.
Stage-specific guidance
Use frameworks differently as the organization develops
Start with the need, people affected, history, existing efforts, system conditions, and limits of the organization’s perspective. A systems map and early theory of change can keep a promising idea from becoming an untested program assumption.
- Invite people closest to the issue to identify conditions, actors, relationships, assets, exclusions, prior efforts, and unanswered questions.
- Draft several possible change pathways instead of treating the first intervention idea as inevitable.
- Mark what comes from community input, lived experience, research, administrative data, organizational belief, or an unresolved question.
Ready to move on when: You can describe the need and context without collapsing them into your proposed service, and you know whose perspective is still missing.
Interactive framework workspace
Logic model builder
Connect your resources, activities, and intended results.
Fictional example
Illustrative example: Willow Street Family Resource Network
A fictional nonprofit is considering neighborhood legal-navigation appointments. Staff know residents face housing and public-benefits questions, but they have not yet tested whether unclear information, limited trusted referrals, appointment supply, language access, or another condition is the most important barrier.
Template-first model
“If we hold workshops and appointments, then families will become stable and the community will thrive.”
Testable working pathway
“If trained navigators and trusted partners provide accessible eligibility information, referrals, and appointments, then participating residents may better understand options and complete appropriate next steps. Over time, this may contribute to more timely problem resolution if appointment capacity, language access, referral quality, and available legal or public-benefit remedies are sufficient.”
The working pathway names participants, activities, near-term change, longer-term contribution, and conditions that may break the sequence. It avoids promising legal results and creates questions the pilot can examine.
A connected five-framework practice
- 01
Map the system
Describe the need in context: people, history, assets, actors, relationships, policies, incentives, constraints, feedback, prior efforts, and forces outside program control.
Prompt: What shapes this issue, whose perspective is represented, and where could our map be wrong?
- 02
State the theory of change
Explain the proposed causal pathway in plain language. Use If, Then, So That, add a Because statement, and identify conditions required for the pathway to hold.
Prompt: If we do what with whom, then what may change, so that what becomes possible—and why?
- 03
Build the logic model
Connect available inputs to activities, direct outputs, sequenced outcomes, longer-term contribution, contextual factors, and evidence questions.
Prompt: What do we invest, what do people experience, what is produced, and who or what should change over time?
- 04
Map responsibility
For each critical deliverable and decision, identify one accountable owner, the people completing work, required contributors, approval authority, and people needing information.
Prompt: Who owns the result, who decides, who contributes, who does the work, and who must know?
- 05
Run the learning cycle
Select uncertain links, define questions and feasible evidence, interpret findings with people affected, decide what changes, and version the model.
Prompt: What evidence would strengthen, weaken, or change this decision, and who will use it?
Framework quality checklist
- The framework has one named decision, audience, scope, owner, version, and review date.
- People affected by the work have a meaningful way to shape, question, correct, or decline participation in the model-building process.
- The need is described separately from the organization’s proposed program or preferred solution.
- Inputs, activities, direct outputs, near-term outcomes, intermediate outcomes, and long-term contribution use distinct definitions.
- Every arrow or causal link can be explained in words and has assumptions, alternative explanations, and relevant contextual factors.
- Outcome statements identify who or what changes, the direction of change, and an appropriate timeframe without promising attribution the program cannot establish.
- Evidence sources, definitions, ownership, timing, access, privacy, limitations, and intended decisions are proportionate to the question.
- Responsibilities identify a clear owner while preserving participant voice, appropriate governance, and subject-matter review.
- The model records uncertainty, disagreement, missing perspectives, risks, and possible unintended or unequal effects.
- The team has defined when the framework will be reviewed, revised, archived, or retired.
Common framework failures
- Choosing a framework because a funder or template uses the name.
- First name the decision or question. Then use the smallest framework that helps the intended users make that decision, while meeting any required format separately.
- Building the model alone and presenting it as shared truth.
- Document authorship and limits, involve people closest to the work, preserve disagreement, and create a correction and revision process.
- Treating outputs as outcomes.
- Outputs are direct products of activities, such as sessions or referrals. Outcomes are changes for people, organizations, relationships, practices, or conditions.
- Using arrows without explaining what makes the link plausible.
- For every important link, state the mechanism, assumptions, evidence, dependencies, competing explanations, and conditions outside program control.
- Making a one-way model for complex or adaptive work.
- Use a simple model for communication, then record feedback, context, iteration, and uncertainty in the narrative and learning process.
- Polishing the diagram instead of using it.
- Review the model with delivery, budget, responsibility, and evidence records. Version it when decisions, activities, definitions, or context change.
Evidence that a framework is helping
Judge a framework by whether it improves shared understanding and decisions, not by visual complexity or completion. Keep framework use distinct from evidence of program outcomes.
- 01Shared language: intended users can explain key terms, distinctions, causal links, assumptions, and uncertainties consistently enough to work together.
- 02Decision use: the framework is referenced in a documented program, budget, responsibility, measurement, partnership, or adaptation decision.
- 03Perspective quality: affected people and relevant partners can see how their input, disagreement, and corrections changed the model or why it did not.
- 04Pathway integrity: activities, outputs, outcomes, timing, resources, and responsibilities remain aligned in actual operations rather than only in the diagram.
- 05Learning use: evidence is connected to a focused question and results in a recorded confirmation, revision, pause, test, or stop decision.
- 06Model stewardship: owners, versions, review dates, definitions, evidence states, and changes remain current and accessible to intended users.
Primary references
Sources and review
- Coach House Accelerator
Coach House
The internal sequence behind this guide: Need Statement alignment, If-Then-So That Theory of Change, Systems Thinking reflection, explicit assumptions, and program piloting.
- Step 2 — Describe the Program
Centers for Disease Control and Prevention
Current guidance on logic models, program narratives, inputs, activities, sequenced outcomes, context, and collaborative development.
- CDC Program Evaluation Framework
Centers for Disease Control and Prevention
A practical, nonprescriptive evaluation framework organized around context, program description, questions, evidence, conclusions, and action.
- Step 4 — Gather Credible Evidence
Centers for Disease Control and Prevention
Guidance for connecting logic-model activities and outcomes to defined indicators and feasible qualitative or quantitative measures.
- Evidence Readiness Resources
AmeriCorps
Federal training and templates for logic models, evaluation readiness, research questions, data culture, capacity, sustainability, and scaling.
- Developing a Logic Model
AmeriCorps
A training resource defining inputs, activities, outputs, and short-, medium-, and long-term outcomes.
- Developing a Project Logic Model and Its Associated Theory of Change
U.S. Agency for International Development
Guidance connecting context, causal outcomes, interventions, assumptions, indicators, a theory of change, and its visual logic model.
- Tools and Training — Logic Model
Minnesota Department of Health
State public-health guidance and examples describing logic models as maps of what a program does, why, what it expects, and how it will examine progress.
This educational guide and planning workspace do not determine program effectiveness, causal attribution, evaluation readiness, grant competitiveness, legal compliance, community agreement, or impact. Requirements vary by funder, program, population, evidence question, jurisdiction, and context. Review consequential models and claims with affected people and qualified program, evaluation, legal, financial, accessibility, privacy, and subject-matter professionals as appropriate.