A logic model is a visual representation of how your organization believes change happens. It maps the inputs you invest, the activities you conduct, the direct outputs you produce, the outcomes people experience, and the long-term impact you're working toward. Logic models aren't just theoretical exercises—they provide the foundation for everything else you do around impact measurement and strategy.
Many organizations have never articulated their logic model explicitly. Staff and leadership understand the work implicitly, but this understanding often differs between people and becomes clearer only when you try to write it down. A logic model forces this clarity. It also communicates your approach to funders, partners, and board members in a concise format.
The Five Elements
Inputs are the resources you invest to do your work. These include funding, staff time, volunteers, facilities, equipment, partnerships, and community relationships. Your inputs are what you bring to bear to address a problem. For example, a youth mentoring program's inputs include funding for staff, volunteer mentors, office space, curriculum, and relationships with schools.
Activities are what you do with your inputs. They're the programs, services, and interventions you implement. For a mentoring program, activities include recruiting and training mentors, matching mentors with youth, facilitating monthly meetings between mentors and youth, and providing support and coaching to mentors. Activities are things your organization does, not things that happen to participants.
Outputs are the direct products of your activities. They're the easily measured results of your work. For a mentoring program, outputs include the number of mentors recruited, number of youth served, number of mentoring sessions conducted, and number of mentors trained. Outputs answer the question "How much did we do?" They're necessary but not sufficient for demonstrating impact.
Outcomes are the changes that result from your activities. They're what participants gain from your intervention. For a mentoring program, short-term outcomes might include youth developing stronger academic skills, increased school attendance, or improved relationships with caring adults. Long-term outcomes might include higher graduation rates, college enrollment, or economic self-sufficiency.
Impact is the fundamental change in the broader systems or conditions you're trying to address. For a mentoring program, impact might be reduced dropout rates in targeted schools, reduced youth unemployment in the community, or strengthened systems supporting youth development. Impact is typically only visible at scale or over long time periods.
How Logic Models Work
The power of a logic model is showing how these elements connect. Your inputs enable your activities. Your activities generate outputs. Your outputs lead to outcomes. Your outcomes, at scale or sustained over time, contribute to impact.
Most importantly, a logic model includes your assumptions about how change happens. Why do you believe your activities will produce the outcomes you expect? What has to be true for your intervention to work? For a mentoring program, you might assume that caring relationships with adults support youth development, that mentors can effectively support academic improvement if trained properly, and that youth will engage with mentors if matched appropriately.
When your program doesn't achieve expected outcomes, your logic model helps you understand why. Did you lack sufficient inputs? Were your activities implemented as planned? Were your assumptions about how change happens incorrect? A logic model creates a framework for learning from outcomes that differ from expectations.
Creating Your Logic Model
Start by gathering your team. This shouldn't be a one-person exercise. Include staff implementing programs, leadership setting strategy, and ideally people with lived experience of the problem you're addressing. Diverse perspectives strengthen logic models.
Begin by describing the problem you're addressing. What evidence tells you this is a significant problem? Who is affected? What are root causes? Being specific about the problem shapes everything else in your logic model. A logic model for addressing drug addiction differs fundamentally from one addressing mental health generally.
Next, articulate your theory of change. How do you believe change happens? If you're addressing drug addiction, do you believe it stems from trauma, poverty, lack of treatment access, or something else? Your theory shapes your chosen intervention. If you believe addiction stems from lack of treatment access, your program might focus on expanding treatment capacity. If you believe it stems from trauma, your program might focus on trauma healing.
Map your inputs. List all resources you invest. Be comprehensive—include things you might take for granted like office space or experienced staff. Reviewing inputs helps you understand your actual cost of operations and where you might be under-resourced.
Describe your activities in detail. What specifically do you do? How often? For how long? Under what conditions? Don't just say "we provide mentoring." Describe how you recruit, train, match, and support mentors. Describe mentoring frequency and content.
Identify outputs. How will you know you're implementing your program? What metrics track output—number served, hours delivered, milestones completed? Most organizations already track outputs through program data. You're just being intentional about including them in your logic model.
Articulate outcomes. What changes should happen for people participating in your program? What will be different about their lives, their skills, their relationships, their opportunities? Distinguish between short-term outcomes (changes visible in weeks or months) and long-term outcomes (changes visible in years).
Connect the dots. Draw or describe how each element connects to the next. Your inputs enable activities. Activities produce outputs. Outputs lead to outcomes. Make sure each connection makes sense and flows logically.
Finally, articulate your assumptions explicitly. What has to be true for your logic model to work? Your assumptions might include that participants are motivated to engage, that your staff are skilled enough to implement effectively, that your theory about what causes change is correct, or that community context is conducive to change. Articulating assumptions helps you identify where your logic model might fail and what you need to monitor.
Logic Models and Measurement
Your logic model guides what you measure. You'll want to track some inputs and activities (to document what you do), outputs (to show how much you do), and outcomes (to demonstrate impact). You probably won't measure impact directly unless you're in a long-term initiative at scale, but you'll measure outcomes that contribute to your theory of impact.
Your logic model also helps you prioritize measurement. You have limited time and resources for data collection. Rather than trying to measure everything, focus on measuring outcomes that are most important and most directly connected to your intervention. Your logic model shows you which connections are most critical to test.
Finally, your logic model helps you interpret results. If outcomes differ from expectations, you can trace back through your logic model to understand why. Did you not invest sufficient inputs? Did activities not reach people? Did activities not lead to outputs? Did people engage with outputs but not experience outcomes? Were your assumptions about how change happens incorrect? Your logic model creates a diagnostic framework.
Different Types of Logic Models
The basic logic model (inputs—activities—outputs—outcomes—impact) is the most common, but variations exist. Some organizations create program-level logic models that describe a single program. Others create organizational-level logic models that describe how their overall organization works. Coalitions create coalition-level logic models that show how multiple organizations work together to achieve outcomes.
Some logic models are linear, with each element flowing in a single direction. Others are cyclical, recognizing that learning from outcomes feeds back into improving inputs and activities. Some include external factors (community conditions, policy changes, economic conditions) that affect whether your program achieves outcomes.
The format varies too. Some are visual diagrams. Some are written descriptions. Some include detailed information about measurements and data sources. Some are simple and concise. The format doesn't matter as much as the thinking process that goes into creating the logic model.
Frequently Asked Questions
Q: What if we can't articulate a clear logic model for our work?
A: This is actually common and valuable. If you're struggling to articulate your logic model, it often means you need to do more thinking about your theory of change. Why do you believe your work leads to outcomes? What's your evidence for that belief? Use the logic model development process to clarify your theory. If you still can't articulate a clear model, that's important information that you need to do more learning and research about your work.
Q: Should our logic model stay the same over time?
A: No. As you learn more about what works through evaluation, your logic model should evolve. If you discover that your assumptions about how change happens were wrong, update your logic model and adjust your program. If you discover that certain activities are more effective than others, adjust accordingly. Your logic model should grow and improve as your learning increases.
Q: Do we need a separate logic model for each program or one for the whole organization?
A: Ideally both. An organizational logic model shows how all your work connects to your mission. Program-level logic models show how specific programs contribute to organizational outcomes. The program models feed into the organization model. Having both creates clarity about alignment between individual programs and overall mission.
Q: Can communities have input into logic models?
A: Absolutely. In fact, community input improves logic models. People with lived experience of the problem you're addressing have insights about root causes and solutions that organizations might miss. Community members might challenge your assumptions or offer different theories about how change happens. Including community voice in logic model development produces models that better reflect community reality.