How to Master the Practice

of Defining Problems
Problem Defintion Tool
by Stefaan Verhulst, Andrew Young, and Andrew J. Zahuranec

1
INTRODUCTION

Problem Definition Tool

The first step in addressing any problem should be first understanding what it is. A well-
defined problem leads to targeted solutions. It prevents practitioners from wandering from
the research path and into thickets of extraneous and unnecessary work. It frames
hypotheses in ways capable of yielding strong and robust conclusions, limits the problem
space to what is feasible and relevant, optimizes efforts to engage with the communities most
affected by the problem at hand, and supports organizations in prioritizing problem-solving
efforts and resources.

As we believe many initiatives fail due to lack of refined problem definition, this tool serves to
support public entrepreneurs in better defining the problem their project is trying to solve.
The tool described here leads users through a step-by-step self-reflective journey that
includes core questions, brainstorming tips, and examples to help the reader develop a
sound problem statement.

1
 Problem Definition Tool

BRAINSTORMING TIPS:

● Shorter is better. Try to stay under 280

characters;
● Avoid jargon or details that an outsider
would have difficulty understanding.
● Avoid passive voice to frame the problem
in relation to the problem owner and
group you want to help.

USEFUL RESOURCES:

● Innovation Skills Accelerator: Problem

Definition
● Solving Public Problems, Chapter 4:
Defining a Public Problem
● Periodic Table of Open Data Elements

Photo by Patrick Perkins on Unsplash

2
2


PHASE TWO:

Problem Definition Tool

CLARIFICATION CYCLE
BRAINSTORMING TIPS:

After drafting your problem statement, ● The Five Whys: Using the initial starting
expand it to a full paragraph describing the point, ask yourself “why is this problem
context, stakeholders, and process that taking place?” Once you have an
would be required to solve the problem. As answer, ask yourself the question four
the title suggests, the goal of this phase is more times to arrive at a root cause or
to arrive at a more granular understanding more actionable piece of the problem.

of the problem area, disentangling ● The Seven “So What’s:” Using the
individual elements until you arrive at an initial starting point, ask yourself “why
issue that can be feasibly addressed. The does this problem matter?” As with
brainstorming tips to the right might be before, repeat the question six more
useful in this. For example:
times with increasing specificity.

● Reframing: After establishing the

● Why are people overdosing in the legitimacy and value of the project,
State? Because they are addicted to bring in educated outsiders who can
opioids.
speak freely and provide input. Ask
● Why are they addicted to opioids? them what is missing and what the
Because of overprescribing by doctors objective should be.

and pharmacists.
● Divergent Thinking: Ask yourself what
● Why are prescribers overprescribing? you would have if the problem were
Because of an overemphasis on pain solved. Then, ask what actions might
reduction and a lack of awareness of need to be taken to achieve that vision
addiction risks.
and what obstacles might emerge.

● Why do prescribers not know addiction ● Challenging Assumptions: Ask yourself

risks? Because they do not know what problems might underlie your
prescribing norms or patient history.
understanding of the problem. What
● Why do they not have this information? evidence supports your understanding
Because there is no accessible of the situation?

database for tracking this information.

USEFUL RESOURCES:

Revised Problem Definition: “Large numbers of state

residents are overdosing on opioids. This problem is ● People-Led Innovation: Toward A
exacerbated by misconceptions about the safety and Methodology for Solving Urban
efficacy of painkiller drugs among prescribers,
leading to overprescribing, as well as a lack of
Problems in the 21st Century 

awareness as to the sheer number of painkillers 

prescribed. The state does not have a resource to 

allow doctors to see data from drug companies, other
prescribers, or law enforcement.”


3
PHASE THREE: MAPPING BRAINSTORMING TIPS:

Problem Definition Tool

CYCLE
● Topic Mapping: By reviewing existing
literature in the field, produce a scan of
With a high-level perspective of the core the issues related to the topic that can
problem developed during the clarification provide a basic overview of the situation.
phase, can further expand your This might reveal gaps in research or
understanding by looking at the context in trends and commonalities.

which the problem exists. The goal of this ● Systems Mapping: Workshop and
stage is to ensure that the proposed present the aspects of the problem with
solution will have the desired effect by an emphasis on contributing and knock-
examining potential positive and negative on effects. Use this exercise to better
effects. It is also to uncover potentially understand causes and effects and
overlooked factors that might prevent the ensure risks do not outweigh rewards.

approach from succeeding. The ● Process Mapping: Map out the real-world
brainstorming tips to the right can help timeline of how the problem has
officials through this cycle. For example:
progressed and the likely impact on that

timeline if a solution were introduced.

● Disrupting Mental Models: Write down


 and interrogate key assumptions. Ask if

 your beliefs are the truth, if the truth is

 obvious, if it is based on real data, and if

 that data is representative.


 ● Self-Evaluation: Interrogate your


 capabilities and position to see if you can

 realistically offer a solution to the

 problem.


 ● Prioritizing: Ask how different strands of


 the problem should be prioritized

 according to greatest potential impact,

 feasibility, and relative importance. If

 multiple strands of the problem area are

 to be addressed, consider the sequence

 in which it should occur.

USEFUL RESOURCES:

● Adolescent Mental Health: Using A

Participatory Mapping Methodology to
Jointly Identify Key Topics, Questions,
and Priorities for Future Work and Data
Collaboration

● Mapping how data can help address

COVID-19

● PeaceTech Topic Map: A Research Base

For An Emerging Field


4
PHASE FOUR: QUESTION BRAINSTORMING TIPS:

Problem Definition Tool

CYCLE
● Insight Type: Research questions can
yield many types of insights, but your
After you’ve mapped the problem goal might be targeted at one type of
statement, you might ask yourself how this “insight” over another. Ask yourself if
information can be synthesized into a you seek situational awareness(also
research question. In doing so, you can known as descriptive insights), an
reorganize your thinking toward the aspect understanding of cause and effect,
of the problem you hope to study and what (also known as diagnostic insights) a
a “success” for you might look like. What is prediction about the future (also known
the target of your work? What are the as predictive insights), or an
metrics you want to track or goal you hope assessment of impact (also known as
to achieve? Who might be involved and has prescriptive insights).

similar work been conducted in the past that ● Public Value: Research questions will
you could build on?
also, ideally, allude to the ways in which
In these ways, question-driven thinking, the information can help a problem
more than solution-driven thinking, can be stakeholder. Ask yourself if your
useful in exposing how you specifically question can improve governance,
relate to the challenge at hand. For empower people, create opportunity or
example:
solve some other public problem.

● Audience: It can often be helpful to

Revised Problem Definition: “Large numbers of state frame your question in a way that
residents are overdosing on opioids. This problem is recognizes who will be affected by the
exacerbated by misconceptions about the safety and
efficacy of painkiller drugs among prescribers,
answer. Often, those charged with
leading to overprescribing, as well as a lack of solving the problem are not most
awareness as to the sheer number of painkillers affected by them and so individuals
prescribed. The state does not have a resource to must think critically about what the
allow doctors to see data from drug companies, other larger, societal consequences of their
prescribers, or law enforcement. The questions, then,
intervention might be.

are: How will overdose rates in the State be affected

by enabling data collaboration between doctors, ● Using Expert Crowdsourcing: To help
other prescribers, drug companies, and others? identify questions that are useful and
Understanding the sensitivities of this data, what is relevant, especially on issues where
the minimum asset that doctors need to make you might lack knowledge, it might be
informed decisions and how can it be meaningfully beneficial to solicit the input from an
protected?

outside group of expert stakeholders.

USEFUL RESOURCES:

● Smarter Crowdsourcing

● The 100 Questions Initiative

● The Data Assembly: Responsible Data

Re-Use Framework


5
 overdose numbers, the state can examine the number
PHASE FIVE: CHANGE

Problem Definition Tool

of emergency room visits and opioid-related death
rates. While this data is available in standardized
THEORY CYCLE
formats in major cities, additional resources are likely
necessary to collect and standardize data from rural
Finally, the problem statement can be communities. An additional possible obstacle exists in
supplemented with a concrete theory of the decentralized nature of the state’s government.
change articulating the underlying Officials might need to liase local bodies to gain real-
time access to data.

assumptions of what will happen as a result

of your project and a theory of why that is
the case. With detail, develop a cohesive BRAINSTORMING TIPS:
story starting with a problem you want to
solve, and then describe the activities you ● Developing Issue Trees: Write out how
will perform, the outcomes that those events might proceed in different
activities will produce and how those circumstances, ranging from most to
outcomes will end up having the expected least likely.
impact. Using the tips to the right, you might ● Identifying Barriers: List the obstacles
produce something like:
most likely to arise if different
hypotheses are acted upon to verify
Revised Problem Definition: State residents are that long-term and short-term
overdosing due to opioid-related causes. This
consequences are not opposed, and
problem, borne from misconceptions about the safety
and efficacy of prescription painkillers, can be best the solution does not merely shift the
mitigated at its source: Prescribers. With a burden.
prescription drug monitoring database supported by ● Quantifying and Measuring: Discuss
state agencies, drug companies, and healthcare what types of impact are important for
providers, the state can give doctors and pharmacies
the project and whether the data points
data to limit use and misuse of addictive drugs. We
seek to investigate whether this information will
for understanding it are available. In
meaningfully impact overdose rates and what kind of clear, realistic terms, articulate how
information prescribers need, understanding the data long it will take to measure success or
sensitivities involved.
failure and how often data must be
This project has several possible outcomes. The most collected.
optimistic scenario entails complete adoption of the
USEFUL RESOURCES:

service by all prescribers in the state, leading to a

significantly reduced overdoses from prescription Open Data Impact: When Demand and
●
drugs. The most pessimistic scenario envisions the
database being ignored or obstructed by most
Supply Meet

prescribers. Both these scenarios are unlikely. More ● Open Data in Developing Economies:
plausibly, the platform will be adopted by many Toward Building an Evidence Base on
prescribers across the state, though gaps will exist What Works and How
where IT infrastructure is lacking. Obstacles might ● Leveraging Private Data for Public
arise if drug companies decline participation in the
Good
system or if the state is unable to devote the
resources needed to keep the database up-to-date.
● The Third Wave of Open Data Toolkit
● Open Data in Action: Initiatives During
Several metrics emerge, the most obvious being page the Initial Stage of the COVID-19
views and the number of searches, which should be
Pandemic
available to the web host. To observe the effect on

6
 Problem Definition Tool

PHASE TWO: CLARIFICATION CYCLE

Using the problem definition above, fill out the following tables:

Why is this problem taking place?

Why is that aspect of the problem

taking place?

Why is that aspect of the problem

taking place?

Why is that aspect of the problem

taking place?

Why is that aspect of the problem

taking place?

7
 Problem Definition Tool
Why does this problem
matter?

Why does that matter?

Why does that matter?

Why does that matter?

Why does that matter?

Why does that matter?

Why does that matter?

How do I view the problem and its various How does an outsider view the problem?
elements?

8
 Problem Definition Tool
What would I have if the problem were
solved?

What steps would need to be taken to

achieve that?

What evidence supports my conception of

the problem?

9
PHASE THREE: MAPPING CYCLE

Problem Definition Tool

In the space below, sketch out a diagram that explains the aspects of the problem, timeline,
or alternative hypotheses that might explain it. Use the diagram to inform a discussion about
your capabilities and the priorities that you should have going into the project.

10
PHASE FOUR: QUESTION CYCLE

Problem Definition Tool

What research question am I trying to
address?

What is my metric of success?

What kind of insight will this research

generate?

What kind of public good will the insights

yield?








11
PHASE FOUR: CHANGE THEORY CYCLE

Problem Definition Tool

Using your previous work to fill out the following tables:

What are possible outcomes from this work?

Highly Likely Scenario:

Somewhat Likely Scenario:

Somewhat Unlikely Scenario:

Highly Unlikely Scenario:

What obstacles might emerge in each of these scenarios?

Highly Likely Scenario:

Somewhat Likely Scenario:

Somewhat Unlikely Scenario:

12
 Problem Definition Tool
Highly Unlikely Scenario:

What will I measure? How will I gain access to the metric?

13