Intended audience: DataKind Volunteers
“This is THE MOST IMPORTANT part of the entire DK process… converting an ill-formed business problem into a data science-ready project is absolutely the key to success.” - Jake Porway, DataKind’s founder and Executive Director of nine years
At this point, you have already looked at organizational data readiness, availability, and maturity. You are in the process of scoping a couple project ideas, and are ready to write the impact statement:
“I want to (analysis) using (data) so that (behavior change) so that (impact).”
Example: Medic Mobile wants to predict the number of lives saved for new deployments using the Lives Saved Tool data so that they can more effectively select new public health programs so that more lives are saved. (See additional examples here)
Here is a version of the project statement template ready to share with the partner organization. Additionally, this Playbook article introducing project statements is a great resource to share with a partner organization that wants more details on what to expect.
Inputs
- Your notes from your discovery calls with the partner
- A copy of the impact map template for your project
- A copy of your partner’s theory of change
- A call scheduled - consider working together on this with the partner organization!
1 Start with your organization’s goals
From your discovery calls, you already have answers to the following questions:
- What’s your mission?
- What’s your vision?
- What change do you want to see in the world?
- What keeps you up at night?
- What does the world look like when your team has fulfilled its mission?
The “so that (impact)” will be that ultimate goal that the partner is trying to achieve. Your project should ultimately map to this goal. Pick a couple possible answers to that question, and put them in the associated column in your impact map.
2 Identify pain points
Review your partner organization’s theory of change. Going back to your discovery call notes, review your partner’s answers to the following questions:
- What is the process for achieving your vision?
- How do you map your activities to your mission?
- What are the outcome levels?
- What is the user experience for your beneficiaries?
- Review your logic model or theory of change, and…
- Identify bottlenecks
-
Imagine areas that could be improved
- What are some of the pain points or challenges your team faces?
Look at these answers in combination with your partner’s theory of change, and add possible (behavior change)’s to your impact map, identifying areas that you could address to truly move the needle on a goal. As you scan the theory of change, you’ll likely see several options of things that could go here. Keep thinking creatively, and come up with several options. What change do you want to make that will end up completing the vision?
3 Suggest some data science solutions!
Now, it’s time to start brainstorming some data science solutions to address the challenges you uncovered! Going back to your discovery call notes, review your partner’s answers to the following questions:
- What data science project do you think would help you address these issues or pain points?
- Do you want to understand the past, improve the present, or predict the future?
- Who will be the users of the data solution? How will they use it?
- If 100% of people adopted the solution, what would it solve?
Look at these answers together with your partner’s theory of change, and add possible “I want to (analysis) using (data)”s to your impact map.
Make sure all your ideas clearly map to the goal, and add any intermediate “so that” steps needed to clearly demonstrate how the data science project will lead to the change. Create as many project statements as you can think of, this is the time for creativity! Later, the data audit will narrow the scope of the project to something realistic.
4 Craft the external version
That impact statement can be hard to read externally, so we can also frame the same information like so:
Organization X exists to (mission). They do this through (theory of change). They have (a problem statement / hurdle to doing their work). We’re proposing to use (data) to do (analysis) that will ultimately (behavior change), leading to (impact).
For example: MedicMobile exists to provide life-saving healthcare to those who live prohibitively far from hospitals. They do this through deployments of cellphone-based healthcare programs. They have difficulty knowing which programs to deploy in which regions to maximize the number of people helped. We’re proposing to build a software interface to the “Lives Saved Tool”, a statistical model for public health officials to estimate impact, that will predict the number of lives saved for new MM deployments. This solution will ultimately help MM more effectively plan their public health programs, leading to more lives saved.
Every project should have a paragraph like this that all team members can understand.
Contributer(s): Mitali Ayyangar, Cassy L Cox, Shanna Lee, Daniel Nissani, Jake Porway, Rachel Wells
Contact us
If you would like to learn more about us, partner with us, or get in touch, visit our website or email community@datakind.org.
Subscribe