Designing for impact

How to build companies that change people’s lives

You’ve got the vision. You’re building something that matters. But how do you ensure that your product or service don’t just look good, but work to make the world a better place?

That’s what this section is all about: Designing for Impact. Not as a bolt-on or something you think about after launch, but as something that’s baked into the way you build.

This article covers three core practices:

1. Mapping a Theory of Change

2. Achieving Problem-Solution-Impact Fit

3. Embedding measurement into your product from the beginning

Let’s dive in.

1️. The theory of change for startups: a simple toolkit

A Theory of Change (ToC) is a tool that helps you clarify:

• What change you’re trying to create

• How you believe that change will happen

• What you’ll need to do to make it real

It’s a map, not of your tech stack or user journey, but of your impact logic.

The building blocks

A simple Theory of Change includes:

• Inputs – What you invest (time, money, expertise)

• Activities – What your startup actually does (e.g. coaching sessions, product features)

• Outputs – What is immediately produced (e.g. # of users trained, # of lessons completed)

• Outcomes – The real-world effects (e.g. better literacy, improved job placement)

• Impact – The long-term shift you’re aiming for (e.g. reduced unemployment, more equitable access to education)

Here’s a simplified example for a skills assessment startup:

Element	Example
Input	Team + tech platform
Activity	Deliver 20-minute skill assessments to jobseekers
Output	1,000 jobseekers complete assessments
Outcome	600 receive skill matches → 300 get interviews
Impact	Shorter time-to-hire and better job fit at scale

You can sketch this on a napkin. The point is to articulate your assumptions.

Once they're visible, you can test and refine them like any other startup hypothesis.

Tools to use:

• Miro or Whimsical for visual mapping

• Notion or Google Docs for a lightweight narrative version

• Sticky notes on your office wall (…this still works)

Top tip 1: keep it lean

You don’t need a 30-page report. A 1-pager can go a long way, especially early-stage. What matters is clarity, not perfection.

Top tip 2: keep it research-based

Many companies create a theory of change that is essentially a logic model. For bonus points (i.e. ensuring improving accuracy and robustness), we suggest working (informally or formally) with researchers to form a more reliable logic model and theory of change. They can help verify your assumptions, ensure your hypothesis is grounded in sound, existing research and identify potential impact connections you may have missed (further highlighting possible areas of impact!).  Check out this example of a research-based theory of change and judge for yourself: does your model look like this, with all the arrows? 

https://www.foreduimpact.org/thoughtleadership/theory-of-change

2️. Problem-solution-impact fit

We all know about product-market fit, but in mission-driven startups, that’s not enough. You also need problem-solution-impact fit.

This means three things line up:

1. You deeply understand the problem

2. Your solution is meaningfully connected to solving it

3. Solving it produces measurable, desirable impact

Step 1: validate the problem

Too many startups jump to building without understanding the real-life constraints and root causes their users face.

You can’t design for impact if you haven’t asked:

• “What’s blocking this person from succeeding right now?”

• “Who benefits if this problem goes away - and who might resist the change?”

• “Who is being left out when our solution is implemented?”

Great companies don’t just find problems. They find leverage points, the specific parts of a system where change is possible.

Step 2: map your solution to the outcome

Ask yourself:

“If someone uses this feature or service, what exactly is supposed to change in their life and how would we know if it did?”

If the connection isn’t clear or strong, that’s a red flag.

Step 3: test your impact hypotheses

Like any product hypothesis, your impact claims should be testable. There should be a clear, validated test for the hypothesis, one that can be implemented and tracked over time.For example:

• “If students complete 5 micro-lessons per week, they’ll improve reading levels by 15%”

• “If managers receive real-time coaching, team performance scores will rise within one quarter”

You can now build, test, and iterate on your impact, not just your interface.

3️. Embedding measurement into the product from day one

Too many startups treat measurement like a retroactive add-on:“We’ll worry about that after launch.” Or “We’ll look into impact after we have thousands of users”

But the smartest founders build their products so that evidence generation is a byproduct of use.

Not only does this help you learn faster, it also saves you from painful retrofitting later. Moreover, it protects you from a reputation damage: imagine you built a product that thousands use but researchers find has a negative impact on users!

Jenny Coogan, Brighteye mentor and co-founder at Newsela, commented:

"Absolutely. It's easy to fall into a trap of feature-creep where your product becomes so complex, it can be hard to isolate the data you need to measure impact. By understanding your impact measurement before you build, you'll think twice about any feature that might endanger your ability to track outcomes."

Step 1: decide what’s worth measuring

You don’t need to track everything. Focus on leading indicators of impact, early signals that someone is on the right track.

Begin with some outputs that you can easily measure. Examples:

• Time to first job application submitted

• % of learners who complete 3+ modules

• Number of coaching sessions attended within 2 weeks

Pick 1-3 key metrics linked to your intended outcome. Make them visible to your team.

Step 2: build the impact indicators into your UX/LXD

Make it part of the flow, not a survey tacked on at the end. For example:

• After completing a module, ask “How confident do you feel now compared to before?”

• Prompt jobseekers to log outcomes directly (“Did you get an interview?”)

• Use backend data to track time spent, actions taken, or follow-through

The goal is to make measurement frictionless, for you and your users.

Step 3: design with ethics and trust

If you're collecting impact data, you’re also collecting sensitive human data. Ethics is everything.

• Be transparent with your users about what you're collecting and why

• Allow opt-outs where reasonable and possibility to delete data when requested

• Think hard about how the data could be misused, or misunderstood

Your impact story is only as strong as the integrity of your data.

Bringing it all together

Designing for impact isn’t a separate phase. It’s a way of thinking that informs:

• How you define success

• What you choose to build

• How you measure and learn

When you map your theory of change, ensure problem-solution-impact fit, and embed lightweight measurement into your product, you’re doing more than chasing growth, you’re building something that works, matters, and lasts.