Do’s and Don’ts of Creating Your First Logic Model

Logic models. Theories of Change. Conceptual frameworks. Conceptual models. What’s the difference?? Does it matter? Which one should I use?

I’ll be the first to admit, I occasionally use the words interchangeably. Ugh, I know. The queen of specificity over here, being imprecise with language. See? Nobody is perfect ;)

My take (and note that people legit argue about this) is that these are all more or less the same thing. Generally speaking, theories of change tend to be broader and messier, and logic models more actionable and precise. Conceptual lands somewhere in the middle.

If you’re working with me, I’m going to come in and push you to be uber-precise and really hone in on defining concrete mechanisms for creating change and aligning them to outcomes. I care about the components, not the name. In the words of the great Taylor Swift: “So call it what you want, yeah, call it what you want to.”

If you’d like to borrow my terminology, I go with “Logic model (aka testable theory of change)” because … who doesn’t love a good compromise? I don’t want to fight. I want to create order & meaning.

A Logic model (aka testable theory of change) is where I ALWAYS start with clients. For me, the purpose of a logic model is to map out, in a cohesive framework, (connecting both theoretically and practically) all key areas of your organization/program/idea/etc. That is — your inputs (what you put in), your outputs (what you produce) and outcomes (what happens as a result of what you put in and produce). Sometimes we find that you have logic models within logic models or that there are tiers of outcomes (leading and lagging; proximal and distal), but we can save that complexity for another day.

A LOGIC MODEL IS CRITICAL TO ANY MEASUREMENT AND EVALUATION WORK. IF EVALUATION HAD A GUIDING STAR, THIS WOULD BE IT.

This foundational work is to ensure that there is a clear hypothesis about how you intend to create change, including precise definitions of all aspects of your organization/idea/program/etc. that are aligned to the changes you hope to see. It sets the stage for identifying precise, measurable, meaningful goals. All researchable questions and instruments/indicators should directly map to your logic model, ensuring relevance to your mission and fidelity to your approach.

In short, you use a logic model to Define + Align. You lay out all the key aspects of your organization/idea/program/etc.; specify what they are, how they work, and why they matter; and align them to your goals (which are also very clear and specific. For more on this see: Why I’m Not Buying What You’re Selling: On trading big ideals for precise, measurable outcomes).

Okay, so, you’re sold. You NEED a logic model. Where the heck do you start? What common missteps do I see?

Check out my list below for my top “Do’s” and “Don’ts” for creating your first logic model.

DO:

• Set aside three times as much time as you think this will take. I have seen this process take anywhere from 3 hours to 9 months. The more complex your organization/idea/program/etc. is, the more time you need to set aside. Even if you think you know your organization like the back of your hand — don’t underestimate the amount of thought, negotiation, and buy-in required to get a team to drill down, get really tactical about how change is being created, create a shared and common language for these changes, and commit to all of this as a guiding document. I find we’re all a little imprecise and commitment-phobic in this space, so plan for it.

• Get super specific — what do you provide? What you provide = your inputs and outputs! It’s those things you show up at work to do. It’s your secret sauce. And, in this case, it doesn’t fly to say “we use a secret sauce”- you gotta let the logic model in on the secret. For instance —you don’t improve education, you create and offer a virtual learning experience for math teachers to use during the school day to teach students how to x,y,z math skills while developing a,b,c social-emotional learning skills. And, really, even all that could be WAY more specific.

• Connect the dots — make sure you can explain the flow of the model from input to outcome. Read your logic model out loud as a sentence (or series of sentences). Does it make sense? Does “A” lead to “B” lead to “C”? Or, another way does your INPUT lead to an OUTPUT which leads to an OUTCOME. Or, are you missing something? Are there activities that don’t connect to an outcome? Or, are you missing an output? Interrogate why that may be — is it an activity you shouldn’t be focused on? Or, is there a goal you overlooked? If you can’t verbally describe the model see #1 below…

DON’T:

• Beat yourself up if you have to go back to the drawing board once. Or, twice. Or, 20 times. This is an iterative, engaging process. Focus on getting “it right” not perfect. “Right” in that it accurately describes what you do. Yes, you want the logic model locked down to guide your measurement and evaluation efforts, but, don’t let perfection be the enemy of the good. Your logic model isn’t carved in stone… as your idea/program/organization evolves, so will your logic model!

• Get bogged down in minute details — focus on the essentials! Ha! I know, this sounds contradictory — be specific but don’t get bogged down in minute details. Huh? Trust me — it’s good advice. Sometimes in creating these models, people are thinking SO MUCH about programming that they start to get down in the weeds — breaking down key programming aspects into 10–15 subtasks. If that is happening… ZOOM OUT! You can be specific and not in the weeds. If one of your key inputs is time or coaching or money — that’s great. We don’t need a punch card, list of coaches, or budget line-item… yet.

• Stress about how you will measure all these things - that’s the next step! Another common sticking point in creating logic models is that they get stuck on HOW to measure all of these pieces and feel that if they can’t think of how to measure it, then they can’t have it on the logic model. Here’s the thing — the logic model is explaining how you create change. How you measure all those pieces is a problem for a different day. Let the logic model drive measurement development, not the other way around.