From a problem to a solution.

Key terms

Segment

A group of customers who share something that matters for your product, like what they’re trying to do.

Pain point

A moment where the customer gets stuck, confused, or frustrated.

Journey map

A visualization of the process a person goes through to reach a goal, stage by stage, with what they do, think, and feel.

Persona

A fictional but realistic description of a typical user in a segment, built from research, so the team keeps one specific person in mind.

User story

One sentence that states a need from the user’s side, such as: As a [type of user], I want [goal] so that [benefit].

Problem statement

A short description of who is stuck, on what, and why, with the evidence, written before any solution.

CIRCLES method

A seven-step checklist for open product questions, like “How would you improve…?” or “Design a…”, created by Lewis C. Lin: comprehend, identify, report, cut, list, evaluate, summarize.

The idea

Most work doesn’t arrive as a problem. It arrives as a solution someone has already picked. Add a chatbot. Add a filter. Add a banner. It’s tempting to start building right away. But if you don’t know who it’s for and where they get stuck, you can’t tell whether the request helps at all.

This lesson walks the whole path, from a vague request to a solution you can defend, as one connected flow. First you narrow down: the goal, one group of customers, their worst pain. Then you open up to many ideas, and choose one. One made-up example runs all the way through: the furniture store and its chatbot request.

The lesson, in seven parts

1. 1Understand the goal before the request.

   Write the request down exactly as it was given. Then, before anything else, find out what it’s for. You’re clarifying three things:

   • The goal. What should change, and for whom? More sales, fewer returns, fewer support calls? “Add a chatbot” isn’t a goal. “Fewer shoppers leave with an unanswered question” is closer.
   • The limits. How much time and how many people do you have? Is anything fixed, like a launch date or a platform?
   • The context. Which product, which part of it (the app or the website?), which customers, and what we see today that says something is wrong.

   Ask

   • “What would we see change if this worked?”
   • “Who asked for it, and what did they see?”
   • “What can’t change: the date, the budget, the platform?”

   Avoid

   • “When do you need the chatbot by?” It accepts the solution before you know the problem.
   • “Don’t you think customers would rather…?” It leads them to your answer.

   If nobody can point to something that shows customers are stuck today, finding that out is your first job. At the made-up furniture store, “add a chatbot” becomes a question: which shoppers have questions, about what, and at which moment? Sales tells you most of the questions come from people looking at sofas on the website, so you agree on a goal: more shoppers who are close to buying a sofa actually buy one, starting with the website.

2. 2Split customers into segments, and choose one.

   Different customers need different things, so a product for everyone usually helps no one much. List the main groups. A useful split is by what people are trying to do, not only by age or location. Try to cover everyone, so you don’t miss a group, and keep the groups from overlapping.

   Segment	What they’re trying to do	Why it might matter
   First-apartment movers	Furnish a small space on a budget, often alone	Many first-time buyers; small spaces raise fit questions
   Families replacing a sofa	Find something that survives kids and pets	Bigger orders; worried about durability
   Interior designers	Order for several clients at once	Repeat buyers with very different needs

   A made-up example. Real segments come from your own data and research.

   Then choose one segment to focus on first. Two good tests: how big the group is, and how badly its needs are met today. Once you choose, stay with it. Switching segments halfway through is how you end up with a feature that half-fits everyone. The store chooses first-apartment movers: a large group, and small rooms make their questions sharp.

3. 3Gather evidence, then talk to customers about real moments.

   Start with what your chosen customers already say and do: site searches (especially ones that return nothing), questions in chat and email, product reviews, the step where most people leave, and return reasons. Ask support for an export, or sit with them for an hour. Aim for 30 to 50 real items, copied word for word. Their words are evidence. Your summary of them isn’t.

   Data tells you where people get stuck. Conversations tell you why. Talk to about five people from your segment who shopped recently, and keep going until the stories repeat. Recent orders, support conversations, or people who left a sofa in their cart are good places to find them; a short, friendly note asking for twenty minutes often works.

   The key rule: ask about what actually happened, not opinions or predictions. People are poor at predicting what they’d do, and they’re kind, so they’ll say they’d use your idea. Rob Fitzpatrick’s book The Mom Test sums it up in three rules: talk about their life instead of your idea; ask about specifics in the past instead of generics or opinions about the future; talk less and listen more. Teresa Torres of Product Talk puts it this way: “Keep the interview grounded in specific instances of past behavior.”

   Ask

   • “Tell me about the last time you shopped for a sofa. Where did you start?”
   • “What were you trying to decide at that point?”
   • “What made you stop, or come back later?”
   • “What did you do to figure it out?”
   • “What happened next?”

   Avoid

   • “Would you use a chatbot?” It asks for a prediction.
   • “Do you like this design?” It asks for an opinion, not what they did.
   • “Wouldn’t it be easier if…?” It leads them to the answer you want.

   Listen hardest for workarounds: the tape measure, the call to the store, the photos sent to a friend. A workaround shows the need is real enough that people already spend effort on it. At the store, 19 of 45 customer questions turn out to be about size, and four of six people you talk to left mid-purchase to go and measure something. (Both numbers are made up.)

4. 4Map the journey, and write down who you’re helping.

   Now lay out what your customer goes through, start to finish. Nielsen Norman Group defines a journey map as “a visualization of the process that a person goes through in order to accomplish a goal.” Whatever they look like, journey maps share five key elements: whose journey it is (the actor); the scenario and what they expect; the stages; what they do, think, and feel at each stage; and the opportunities you spot.

   Fill it in from your evidence and interviews. Then look across the map for unmet needs, and check what competitors offer at the same stage.

   Stage	What they do	What they think or feel	Pain point
   Notice a need	Move into a new apartment	Excited, on a budget	None yet
   Browse	Search “small sofa”, open many pages	Overwhelmed by choice	Hard to filter by room size
   Decide	Compare three sofas	“Will it even fit?”	Only overall size is listed
   Check	Leave the site to measure the door	Unsure, a bit stressed	No way to check the doorway
   Buy	Return days later, or don’t	Still unsure	Fear of a costly return

   A made-up journey map, built from the evidence and interviews in part 3.

   Persona

   Maya, the first-apartment mover

   Behavior

   Shops on her phone in the evening; compares three or four options; reads reviews closely.

   Needs and goals

   Furnish a small apartment once, well, without overspending.

   Pain points

   Can’t tell if furniture will fit; worried about a costly return.

   A made-up persona. A real one is built from your research, not imagined.

   Notice that the chatbot never appears on the map. The pain sits at Decide and Check, and it’s about one thing: fit. The persona keeps that real for your team. Nielsen Norman Group defines a persona as “a fictional, yet realistic, description of a typical or target user of the product,” built from user research. That’s the line between a persona and a guess.

5. 5Rank the pain points, and state the problem.

   You’ll find more pain points than you can fix, so rank them. Rate each on two questions: how often does it happen, and how bad is it when it does (do people work around it, or give up)? A third is worth asking too: does it keep happening to the same people? Jakob Nielsen, a cofounder of Nielsen Norman Group, rates the severity of usability problems in a similar way, as a combination of three factors: frequency, impact, and persistence (a one-time problem, or one that bothers people again and again).

   Pick the top one to three. Then write the top one down in two short forms: a user story, in the customer’s terms, and a problem statement, with your evidence.

   Pain point	How often	How bad	Priority
   Can’t tell if it fits the door or room	Often: 19 of 45 questions	Bad: people leave to measure, and some don’t come back	High
   Fear of a costly return	Sometimes	Bad	Medium
   Hard to filter by room size	Often	Mild: they scroll more	Medium
   Unsure about fabric with pets	Rarely, for this segment	Mild	Low

   A made-up example. Counts come from part 3; how bad comes from the interviews.

   User story

   As a first-apartment mover, I want to know a sofa will fit through my door and in my room before I buy, so that I don’t pay to send it back.

   The template

   [Which customers], trying to [do or decide what], get stuck because [what’s in the way]. We see it in [your evidence]. We’ll know we’ve helped when [a change you can observe].

   Filled in, for the made-up furniture store

   First-apartment movers choosing a sofa get stuck deciding whether it will fit their door and room, because the product page only lists overall size. We see it in 19 of 45 customer questions, and in four of six interviews where people left to go and measure. We’ll know we’ve helped when fewer shoppers ask about fit before buying, and fewer sofas come back for size.

   The user story follows the template Mike Cohn of Mountain Goat Software describes: “As a [type of user], I [need/want/am required] to [do something], so that [reason or benefit].” It keeps three things in view: who it’s for, what they need, and why it matters. Notice that neither sentence names a feature. That’s on purpose. The solution comes next.

6. 6List solutions, weigh the trade-offs, and choose one.

   Now, and only now, open up. List several ways to solve the top pain point for this segment: at least three, including a cheap one. The original request goes on the list too. It’s one idea among several, not the answer.

   Then weigh them. For each idea, ask how directly it helps this segment with this pain, how much effort it takes, and what could go wrong. Lesson 3 teaches scoring methods in full. Here, a rough comparison is enough.

   Idea	Helps with fit?	Effort	Trade-off
   Chatbot (the request)	Partly, for shoppers who ask	Medium	Only as good as the size data behind it; silent shoppers get nothing
   Doorway-and-room check on the product page	Directly, at the moment of doubt	Low to medium	Needs accurate measurements for every sofa
   Filter search by room size	Partly, earlier in the journey	Low	Doesn’t answer the doorway question
   3D room planner	Directly, and more	High	Months of work; harder to use on a phone

   A made-up comparison for the furniture store.

   The store chooses the doorway-and-room check: it goes straight at the top pain, and it’s cheap to test. Before anyone builds it, two more things happen. You sketch the flow with a designer: product page, enter your doorway and room size, see a clear yes or no. (Lesson 7 covers wireframes and prototypes.) And you decide how you’ll know it worked: more first-apartment movers who reach a sofa page go on to buy, with fewer fit questions along the way and no rise in sofas returned for size. (Lesson 4 covers choosing metrics.) When you go back to sales, lead with what you found: “Most of the questions are about fit. I’d like to test a size check first. If fit questions don’t drop, the chatbot is next.” The chatbot isn’t rejected. It just isn’t first.

7. 7Say it out loud with CIRCLES.

   Sometimes you have to walk through this whole path in a few minutes: in a meeting, or in a product manager interview, when someone asks “How would you improve…?” or “Design a…”. That’s what the CIRCLES method is for.

   Lewis C. Lin created it, and teaches it in his book Decode and Conquer. In his own words, he invented it to fix “people freezing on open-ended product questions, whether in an interview or at their own desk.” Its seven steps spell the name, and they line up with the parts of this lesson:

   Step	What you do	At the furniture store
   Comprehend the situation	Clarify the goal, the constraints, and the context: what is it, who is it for, why?	Goal: more ready-to-buy sofa shoppers buy; website first
   Identify the customer	List segments, and choose one	First-apartment movers
   Report the customer’s needs	State their needs, for example as user stories	Check fit before buying; avoid a costly return
   Cut, through prioritization	Choose the one to three needs that matter most	Fit first
   List solutions	Several ideas for the top need	Chatbot, size check, room filter, 3D planner
   Evaluate trade-offs	Pros and cons of each idea	Size check: direct and cheap, but needs accurate measurements
   Summarize your recommendation	What you’d do, why, and why over the others	Test the size check first; watch sofa purchases from this segment, fit questions, and size returns

   Step names from Lewis C. Lin’s CIRCLES method, as listed by ProductPlan, with the goal, constraints, and context in the first step as Lin describes it. The furniture store column is made up.

   Treat CIRCLES as a checklist, not a script. In an interview, it stops you from skipping steps, and since you can’t do research there, say your assumptions out loud (“I’ll assume most shoppers are on their phones”) and move on. On the job, every step needs real evidence, not a confident guess.

With AI: AI can sort hundreds of customer questions into themes in minutes, draft a journey map from your notes, and suggest a long list of solutions. Use it for that. But check every theme against the raw quotes, and do the interviews yourself: the follow-up question you think of in the moment is often where the insight is. And because AI makes building faster, it also makes it faster to build the wrong thing. Knowing the problem first matters more, not less. Before you paste customer quotes into an AI tool, check your company’s rules, and remove names and contact details.

REQUEST, AS GIVEN:

1. COMPREHEND THE SITUATION
   Goal (what should change, for whom):
   Limits (time, people, fixed dates or platforms):
   Context (which product, which part, what we see today):

2. IDENTIFY THE CUSTOMER
   Segments:  a.            b.            c.
   Chosen segment, and why (size, how badly served today):

3. EVIDENCE (30 to 50 items, word for word)
   What they said | What it's about | Source
   Interviews (ask about a real, recent moment):
     Tell me about the last time you...
     What were you trying to decide?
     What made you stop, or come back later?
     What did you do to figure it out?
     What happened next?

4. JOURNEY MAP
   Stage | What they do | Think or feel | Pain point
   Persona: name, behavior, needs and goals, pain points

5. PAIN POINTS, RANKED
   Pain point | How often | How bad | Keeps happening? | Priority
   User story: As a ___, I want ___ so that ___.
   Problem statement: [Which customers], trying to [do what],
   get stuck because [what's in the way]. We see it in [evidence].
   We'll know we've helped when [observable change].

6. SOLUTIONS (at least three, including the request)
   Idea | How directly it helps | Effort | Trade-off
   Chosen idea, and why:
   Sketch of the flow:
   How we'll know it worked:

7. SUMMARY (two sentences)
   I recommend ___ for ___, because ___.
   We'll know it worked when ___.

Ask AI about this lesson

All lessons