Design thinking for beginners: the phases and a first exercise

Step 1

Pick a small real problem and talk to people

The first phase in Nielsen Norman Group's model is empathize: researching what people do, say, think and feel. For a practice exercise, pick a small problem you can discuss with people around you: how residents of a building agree on shared tasks, how students in a group share materials, how parents find out about school events. Talk to a few people who have this experience.

• •ask about specific cases: "Tell me about the last time you agreed on repairing the stairwell";
• •look for workarounds: chats, spreadsheets, notices on doors, phone calls;
• •write down verbatim quotes and moments of frustration;
• •lay out what you heard on an empathy map: what the person says, thinks, does and feels.

💡 Don't suggest solutions during the conversation. As soon as you say "what if there were an app…", the person starts judging your idea instead of talking about themselves.

Step 2

Define the problem

The define phase is when observations turn into a clear problem. Group your notes and see what repeats across people. Then write a user need statement: who the user is, what they need and why. NN/g describes such a statement as a tool for agreeing on which problem exactly you're solving.

• •NN/g template: "[A user] needs [need] in order to [goal]";
• •example: "Residents of a building need to know who has already paid for the repair and when, so that agreements don't fall apart";
• •the statement contains no solution: not "they need an app" but "they need to know";
• •if there are several problems, pick one for this exercise.

Step 3

Generate lots of ideas

In the ideate phase, quantity matters more than quality: you'll judge later. To avoid settling on the first idea, reframe the problem as a "How might we…?" question. According to NN/g, this template was introduced by Procter & Gamble in the 1970s, later adopted by IDEO, and is now widespread in design thinking. A good question doesn't suggest a solution, is broad enough and focuses on the desired outcome.

• •example: "How might we let every resident see what stage an agreement is at?";
• •set a short time limit and write down as many ideas as you can, even strange ones;
• •bring someone else in: other people's ideas are good to build on and combine;
• •pick 1–2 ideas using IDEO's three conditions: do people need it, can it be sustained, can it be built.

💡 The solution doesn't have to be an app. Sometimes the best idea is a sign on a door, a message template or a change in the process.

Step 4

Make a simple prototype

A prototype is a quick way to check which parts of an idea work and which don't. Start with the cheapest: NN/g writes that a paper prototype lets you test early ideas at a very low cost and fix problems before money is spent on building something. Draw a few screens on paper or put together simple mockups in Figma, whichever is faster for you.

• •the prototype should show only the main scenario, not the whole product;
• •don't spend time on colors and fonts: logic matters now;
• •if the solution isn't digital, the prototype can be a mockup of a notice or a role-play;
• •prepare 1–2 tasks to check: "Find out who hasn't paid for the repair yet".

Step 5

Test, go back and see it through

In the test phase, show the prototype to people similar to those you talked to at the start and see whether the solution really helps. You'll almost certainly learn something new and go back to ideas or even to the problem statement, and that's normal: the process is iterative. NN/g's sixth phase is implement, and the article specifically notes that it's the one most often forgotten. For a practice exercise, implementation means a polished mockup and a written-up case study.

• •watch what people do, not just what they say about the idea;
• •write down what was confirmed, what wasn't and what you changed after the test;
• •in the case study, show the whole path: conversations → problem → ideas → prototype → test → fixes;
• •be honest about limits: how many people, how many iterations, what you didn't get to check.

💡 Show your need statement and prototype in the community. The most useful question: "Does this solution follow from the problem I defined?"