Discovery, Defined What design discovery is, how to frame a problem, and where your brief starts.
Workhorse Collective
NameAbraham GarciaEmailabe@wrkhrs.co
ART 418 · Fall 2026
MeetsMon + Wed, 1:30 to 2:50TodaySession 04 · 09.02.2026
meaninglayer.com/course/business-of-design
Session 04
Today: what discovery is, how to frame a problem, and then you start your own brief.
01 · The mindset
What Brown means by discovery, what it is not, the four activities it is made of, and the ways it goes wrong. Chapter 1, the reading for today.
02 · Frame the problem
Three kinds of assertion, what a problem statement is and is not, the framing canvas, five whys, four contexts, how to gather and how to process. Chapter 2. Your audience is due in a week and your problem statement in twelve days.
04 · Your discovery starts
Thirty minutes. You open the project brief from the template, write your first assertions, name the gaps, and pick the three questions you will answer first.
The Mindset
PRACTICAL DESIGN DISCOVERY, INTRODUCTION AND CHAPTER 1. THE READING FOR TODAY.
"Discovery is an attitude."
Practical Design Discovery, introduction. Italics in the original.
DBDan BrownAmerican designer. Co-founded the design firm EightShapes.
Not a process, not an artifact, not a phase, not a methodology, not a school of thought.
You are learning: about users, about the business, about the technology.
Brown opens with three claims about discovery
01
It frames the problem and the solution
"You can't truly understand a problem until you spend some time solving it." You try a few ideas to move forward, and that is allowed.
For youThe pitch on Sep 21 carries ideas. That is not cheating on the research. It is how the research gets checked.
02
It happens throughout
Nobody reaches a moment and says "I know everything, let the designing begin." Questions keep arriving. Assumptions get discarded.
For youYour brief keeps changing all semester. Each version you turn in on Blackboard is a step, not the final answer.
03
It is a mindset, not a phase
Time-boxing learning makes it predictable and cheap, and learning does not work on a timetable. Your brain makes the connection on its own schedule.
For youThe calendar has dates on it anyway. Brown's answer is to plan, then adjust. Chapter 4.
FROM THE INTRODUCTION TO PRACTICAL DESIGN DISCOVERY.
What discovery is not
Learning about the business, the technology, and the users. Not yet deciding about layout.
Not strategy
Strategy addresses the problem the business has. Discovery addresses the problem the design has to solve. They overlap, and they are not the same job.
Not execution
Execution implements what discovery established. The mindset shift is from learning to deciding.
Not a methodology
"A compromise, a way to squeeze what we know about creativity into what we need in business." Fits inside any process you already run.
Not requirements
"More than collecting requirements for all to ignore." Requirements read in context stop being a laundry list.
Discovery is four activities, in a loop
01
Gather
New information. Interviews, documents, the domain, the competition, watching people.
ProducesRaw data. Ninety percent of it qualitative, by Brown's count.
02
Process
What you learned, made meaningful: patterns, hunches, stories, pictures.
Different approaches. Patterns, relaxed constraints, sketching together, feedback.
ProducesLots of ideas. Chapter 3.
04
Focus
One approach. Revise, prioritize, imagine building it, consider consequences.
ProducesPrinciples, a concept, models, a plan. Chapter 3.
FIG 1.2 IN PRACTICAL DESIGN DISCOVERY. "NOT LINEAR, BUT INTERCONNECTED."
Six ways discovery goes wrong
01
Being dogmatic about process
Running the method because it is the method. The mindset fits any process; the process is not the point.
02
Regurgitating requirements
Writing down what you were told without the context that would make it mean something.
03
Researching without exploring
The tension between understanding the problem and trying to solve it is where the vision comes from. Skip the second half and you get a report.
04
Thinking of it as "done"
"Discovery isn't ever 'done,' but then again, neither is your website."
05
Jumping into execution
You know the users, it is a simple product, there are no stakeholders. Every team has said this. It is how discovery gets skipped.
06
Separating it from designers
Handing research to someone else and receiving a document. "No one is dispensable here."
FROM PRACTICAL DESIGN DISCOVERY, CHAPTER 1, PAGES 8 TO 10.
What discovery is for
01
Build a diverse team
"Healthy conflict is crucial to good design." More people, more ways of seeing the problem.
02
Focus on practical results
"If your work isn't leading to better tools, it's time to stop discovery." The guard against one more interview.
03
Build a shared language
The team finds out it means three things by "template." Discovery is where that gets settled.
04
Find new approaches
Confidence that you chose the right one, because you looked at others.
05
Establish a product definition
Start with a purpose, arrive at the product's "most basic definition, something the whole team buys into."
06
Encourage consideration
"Discovery gives you a direction, not a defined path." You can change your mind, and the more you learn the less criticism stings.
FROM PRACTICAL DESIGN DISCOVERY, CHAPTER 1. NUMBER FIVE IS WHAT THIS COURSE CALLS THE BRAND.
Frame the Problem
PRACTICAL DESIGN DISCOVERY, CHAPTER 2. THE AUDIENCE IS DUE SEP 9. THE PROBLEM STATEMENT IS DUE SEP 14.
"By not drawing a clear and compelling problem, you are cheating your team out of an incredible unifying and driving energy."
In an interview with Luke Wroblewski. Brown's epigraph for chapter 2.
TCTom ChiProduct designer. Worked on early projects at Google X.
Brown's own search-engine story: the demons already existed as complaints. "You can't control complaints, but you can use them to frame a problem."
Discovery produces assertions
01
Problem statement
Tells the team what needs to be fixed. "Association members were frustrated with the search results because they were out of date and irrelevant."
Its jobA test for every idea afterward: "Does my concept solve the problem?"
02
Project objective
Summarizes what you hope to accomplish. "A search experience that incorporates modern search features and sets the stage for personalized results."
Its jobWhat you will deliver, not what problem you will solve. Kept separate on purpose.
03
Contextual statement
Background about users, business, technology. "Our users are usually looking for a job search or an industry event. The site search uses Google Search Appliance."
Its jobEverything surrounding the product that will bear on the design.
"THEY AREN'T SET IN STONE: THEY REFLECT YOUR BEST UNDERSTANDING AT THE MOMENT. WRITE THEM DOWN IN A SHARED FILE AS SOON AS YOU HAVE A SENSE OF WHAT THEY MIGHT BE."
Four ways to get to a problem statement
01
Pretend to be a reporter
Summarize the situation for a broad audience. Introduce the project and the core conflict as quickly as possible. Focus on the users and what they are up against.
Gives you"Our project should fix this."
02
Ask why, five times
The chain in the book runs from "the search engine needs to be fixed" down to the organization's revenue models. The whys together are the statement.
Gives youA problem that may not be a single sentence. The book allows that.
03
State it as a hypothesis
An assumption about the product that you need to test. The crucial parts are what would prove it wrong.
Gives youA statement that already knows what research it needs.
04
Draw a problem picture
A flowchart of the broken process. A sitemap with the weak areas marked. A mental model showing the gap between needs and content.
Gives youSomething the team can react to. "Get something down on paper."
TWO TRAPS: TOO BROAD TO MEAN ANYTHING, OR TOO PRESCRIPTIVE, WHICH IS A SOLUTION IN DISGUISE. THE SIX SMALLER TRAPS ARE ON YOUR CHECK SHEET.
Framing
What a problem statement is not.
A solution in disguise
"People need an app that..." forecloses the whole first diamond before it opens.
A complaint
"Parking here is terrible." No one specific is in it, so nothing specific can be designed for.
A topic
"Sustainability in fashion." That names a subject. Nothing in it is in anyone's way.
The classic reframe
The elevator was never slow. The wait felt long. The answer was mirrors. Reframes come from knowing what they think.
The problem framing canvas
Who
Who is experiencing this problem?
How do we know that they are?
What
What is the nature of the problem?
What signs point to there being one?
Why
Why is it worth solving?
What does it cost the people who have it?
Where
In what circumstances does it occur?
In what circumstances does it not?
Four questions, arranged so two people can work on them at once. Fill all four, then write one statement that carries all four. That statement is your September 14 assignment.
ATLASSIAN TEAM PLAYBOOK, PROBLEM FRAMING. 30 MINUTES, 3 TO 10 PEOPLE.
Five whys
Problem
Students eat badly.
Why?
They eat whatever is fastest between classes.
Why?
They have fifteen minutes, and the dining hall is a ten minute walk.
Why?
Their schedule was built around which rooms were free, not around their day.
Why?
Registration optimizes for filling classrooms.
Why?
Nobody owns the student's day. Every office owns one piece of it.
A habit at the top, an org chart at the bottom. Every rung is a problem you could frame, and each one lets you design something else. Stop where you can still reach the people to ask.
BROWN, CH. 2. HIS OWN CHAIN STARTS AT "THE SEARCH ENGINE NEEDS TO BE FIXED" AND ENDS AT THE COMPANY'S REVENUE MODELS.
The same sentence, four dialects
You will meet this form again under other names.
Point of view and how might we
Stanford d.school. What most design teams open a project with.
Working backwards
Amazon. Write the press release and the customer FAQ before anyone builds anything. If the release is boring, the frame is wrong.
The job story
Jobs to be done. "When [situation], I want to [motivation], so I can [outcome]." Product teams live in this one.
The problem statement
Brown, chapter 2, and the version this course grades against: [users] must have [need] to achieve [objective].
Project objectives, five things to be specific about
01
Scope
Which aspects of the product you will focus on.
02
Outputs
What kinds of deliverables you need to create.
03
Fidelity
How closely the outputs will resemble a final product.
04
Time
How long you have for each aspect.
05
Priority
What you work on first.
"TREAT PROJECT OBJECTIVES AS HYPOTHESES TO BE TESTED RATHER THAN INSTRUCTIONS TO FOLLOW." THIS IS THE ONE-PAGE PROPOSAL BEHIND YOUR PITCH, SEP 21.
"Don't be the expert." Pick one or two aspects. Capture only what will affect the design.
Technology
Data, content structure, silos, integration with the rest of the suite.
"Don't be afraid to play dumb." The technologists on the team explain it; you write down the constraint.
Users
Expectations, cognitive constraints, the scope of everything they do around the activity, what they believe about the industry.
"You don't have to use personas." User data is a design tool, so keep it in a form you will reuse.
Content
Style, voice and tone; structure and relationships; scale and quality.
"Look beyond the quantitative." What the content is like, not just how much there is.
For a product with no market, the third box is your audience assignment and the first box is mostly empty. The empty box is not a failure. It is your research question.
FROM PRACTICAL DESIGN DISCOVERY, CHAPTER 2, PAGES 26 TO 35.
How to gather, without interviews
01
Domain research
Enough of the field's jargon and basic concepts to read the rest. What the product is, what category it would sit in, who else is in that category.
For your productThe paper, the patent, the grant abstract, the README, the press it got. Brown: "draw pictures" to get your head around it.
02
Existing assets
Everything the makers already wrote: documentation, a pitch deck, a website, a demo, a talk. Compare what is claimed to what is built.
For your productThese are your stakeholder interview, on paper. "Take nothing for granted": every noun they use is a question.
03
Competitors and comparables
Products that offer similar value set the tacit standard for what users expect. Look at what they prioritize, their navigation, their language.
For your productWhatever your audience uses today instead. Its reviews and forum threads are your users talking without being asked.
04
The current product
Use it, or watch the demo. Ask the purpose of every screen. Look for what was compromised for expediency.
For your productIf there is a demo, run it. If there is not, read the documentation as if it were the product.
THE BOOK ALSO COVERS STAKEHOLDER AND USER INTERVIEWS, AS DOES HALL, CHAPTER 5. THIS SEMESTER DOES NOT ALLOW FOR THEM. THE NEXT THREE SLIDES ARE HALL'S VERSION OF THIS, AHEAD OF THE SEP 9 READING.
Just enough: every assumption is a research question you have not asked yet
01
What research is
Hall: systematic inquiry. Asking a question you do not already know the answer to, in a way someone else could repeat. Not a focus group, not a survey you wrote to agree with you.
Hall, ch. 1 and 2"Just enough" is a budget: enough to make the next decision, and no more.
02
Four kinds
Generative: what problem exists? Descriptive: what is it like, for whom? Evaluative: does this design work? Causal: why did the number move?
September isGenerative and descriptive. Evaluative and causal research are not part of this course.
03
The plan, in five lines
The problem. The research questions. The approach, which for you is one of the four research activities. The sources, named. The date.
Hall, ch. 3A research question is not an interview question. "Who buys this?" is research. "Would you buy this?" is a question nobody answers honestly.
04
Assumptions first
Write down what you believe about the product and the audience. Every belief without a source is an assumption. Rank them by how bad it is if you are wrong.
In BRIEF.mdEvery question mark in section 2 is an assumption. Section 4 turns the three worst into questions, each with a source and a date. That is the workshop's step four.
HALL, JUST ENOUGH RESEARCH, CHAPTERS 1 TO 3. CHAPTERS 6 AND 9 ARE DUE SEP 9 WITH THE AUDIENCE, SO HERE IS THE PART YOU NEED FIRST.
The competitive audit, and the number with its context
01
Who the competition is
Whatever your audience uses today instead of your product. A spreadsheet. A YouTube channel. A senior colleague. Doing nothing. Hall's competition is the alternative, not the company in the same category.
Hall, ch. 6The workaround is the competitor. If you cannot name it, you have not found the audience yet.
02
What to audit
What they say about themselves: the positioning, the price, the promise. What users say about them: reviews, threads, the complaints that repeat. The gap between the two is your finding.
Not for copyingYou are looking for the standard the audience already expects, and the thing nobody is doing. Understand what they set, then differ from it on purpose.
03
A number tells you what
Never why. A count, a price, a share, a rank. Quantitative research describes the size of a thing. The reason lives in the reviews and the threads, which is qualitative.
Hall, ch. 9A count without its denominator is a decoration. "Thirty thousand" of what, out of how many, in what year?
04
How to cite one
The source, the year, the population it counted, and what it does not cover. The AIGA census counts people who answered the AIGA census. Say so on the slide.
On Sep 14 and Sep 21Every number in brackets points at a source in the list. Pitch slide 8, supporting details, is this slide done once for your product.
HALL, CHAPTERS 6 AND 9. THE AUDIENCE FORM ASKS "WHAT THEY USE TODAY INSTEAD" AND "WHAT THAT COSTS THEM." THAT IS A COMPETITIVE AUDIT AND A NUMBER.
What you are not doing this month, and why chapter 5 is still assigned
01
The guide has four parts
The study's goal in a line. The factual questions that place the answers. Two warm-ups. Then the questions that are the point. A checklist, not a script.
Hall, ch. 5Open questions that start a story. "Tell me about the groups you work with," not "do you talk to marketing often?"
02
The question that pays
"What kinds of work-arounds do you use?" A habit built around an obstacle is a problem nobody has framed yet, and the person describing it does not know they are describing a product.
Hall, ch. 4From her stakeholder set. It works on a user, a client and a maker. It is the one to memorize.
03
Why not now
Recruiting, scheduling and consent take longer than the twelve days you have. A bad interview done in a hurry produces a confident wrong answer, which is worse than a marked guess.
This monthDesk research only. The reviews and the forum threads are the users talking without being asked; read them with the guide's questions in mind.
04
Why read it anyway
A professional project would interview people, and this is the guide for it. Chapter 5 is read for Sep 21, beside Monteiro.
The templates sheetHer four parts, her three questions, and her museum question set are on the class page, in her words.
HALL, CHAPTERS 4 AND 5. NO INTERVIEWS THIS SEMESTER. THE GUIDE IS HOW A PROFESSIONAL WOULD RUN THEM.
How to process
01
Find the patterns
One observation per line in a spreadsheet, then move them into groups. Let go of the groups you walked in with. Four hundred index cards split six ways, in the book's healthcare example.
Watch forSegments you did not expect. Career confidence, not high school versus college.
02
Capture hunches
Write impressions down as they come, with the evidence, in a shared document the book calls the project whiteboard. Review them with the team.
Watch for"Learn to let go." A hunch the evidence stops supporting gets deleted.
03
Tell a story
A day-in-the-life with a dramatic moment, written from the user's side, with real quotes from the research. Or a scenario: a task with a beginning, middle and end.
Watch forWriting it from the product's point of view. He uses these instead of personas.
04
Visualize
Sticky notes, a collage, a timeline, a two-by-two, a bubble diagram. The search project in chapter 2 became a chemistry equation on a poster.
Watch forWorking everything in. "Too many concepts won't serve to move the project forward."
JON KOLKO, QUOTED IN CHAPTER 2: WITH NO TIME ALLOTTED FOR SYNTHESIS, THE MESSAGE TO THE DESIGNER IS "SYNTHESIZE ON YOUR OWN TIME, OR NOT AT ALL."
Chapter 2, on your calendar
Frame the problem: twelve days.
Sep 9 · The audience
Three contextual statements about users. Who they are, what they use today instead, what that costs them. Which one you lean to. Gathered, not guessed: every paragraph names what you read.
Sep 14 · The problem statement
One page, one audience, the four boxes filled, compressed to one statement from their side. Every claim names a source. Run it against the two traps and the six.
Read for Sep 9
The rest of Brown chapter 2: problem statements and project objectives. Hall chapters 6 and 9, competitive research and numbers. Davis chapter 10. Then for Sep 21: Monteiro chapters 1 and 2, and Hall chapter 5 on interviews.
Between now and then
Gather and process, in the loop, from published sources. The paper or the patent, the makers' own documents, the workaround and its reviews, hunches with evidence in the brief.
The book on the course calendar
Discovery is September.
Frame · Sep 2 to 14
Today the brief opens. Sep 9 the audience, three contextual statements about users. Sep 14 the problem statement, every claim sourced.
Direction · Sep 14 to 21
Two studio days, then the pitch, recorded: Brown's discovery document in five sections, with a concept and principles in the middle of it.
Document · Sep 23 to 30
Definition pressure-tests the strategy. The creative brief is written in class on the 30th and the brand answers to it for the rest of the semester.
Then · Oct to Dec
Explorations, the identity, the rules, the brand book. Brown's "execution." The mindset does not switch off; the deciding starts.
Your Discovery Starts
THIRTY MINUTES, LAPTOPS OPEN. YOU START THE SHARED FILE BROWN TELLS EVERY TEAM TO KEEP.
Three passes
01
Worlds
Underline every world your statement names: places you've worked, communities you're in, things you do that most people don't.
WhyYou know these places. You see problems an outsider can't, and you can reach people there to ask what they think.
02
Friction
In each world, list what's broken, slow, embarrassing, expensive, or worked around.
WhyWorkarounds are the best signal. A habit people built around an obstacle is a problem nobody has stated yet.
03
Match
Hold your friction list against the product list: real technologies with no market. Look for where the two meet.
WhyA technology with no audience, meeting a world where you already know the people, is a project you can carry for a semester.
AN AUDIENCE YOU CAN REACH AND ASK IS WORTH MORE THAN ONE YOU CAN ONLY IMAGINE.
In class, right now
01
Open the brief
A document called BRIEF.md, from the template on the class page, with the five headings on the next slide. Brown's shared cloud document. It is the project whiteboard and it never gets finished.
5 minutesSolo. Save it empty first, so it exists.
02
Assert, badly
Fill the four contexts with what you know right now about your product and your leaning audience. A problem statement, version zero. Mark every line you guessed with a question mark.
12 minutesSolo. Thin boxes are the output, not the failure.
03
Swap and press
Your partner reads your brief and presses on it the way Brown presses a stakeholder: picks any noun you wrote and asks "where did you read that?" Every line without an answer gets a question mark.
5 minutes each wayThe reader asks. The writer does not defend.
04
Three questions
From the question marks, pick the three you must answer to set a direction. For each: which of the four research activities, which document or product, by when. Anything not answerable by Sep 12 is not in the problem statement.
3 minutesWrite them into the brief. That is the research plan.
"WRITE THEM DOWN IN A SHARED FILE AS SOON AS YOU HAVE A SENSE OF WHAT THEY MIGHT BE, AND KEEP REVISING THEM OVER THE COURSE OF THE PROJECT."
BRIEF.md, five headings
1 · Assertions
Problem statement, version 0. Project objective, version 0.
Both dated. Both expected to be wrong. Later versions go above, not over.
2 · Context
Business, technology, users, content. A few lines each.
Question mark on every line you have not checked. The users section becomes Sep 9.
3 · Hunches
Impressions as they arrive, dated, each with the evidence beside it.
Deleted when the evidence stops supporting them. Learn to let go.
4 · Questions, and 5 · Log of what the machine did
The three questions, each with a research activity, a source, and a date.
Then one line per session: what you asked the model, what it gave you, what you kept.
Everything you turn in this month is a section of this file, revised. The audience is section 2. The problem statement is section 1. The pitch reads the whole thing aloud.
THE PROJECT BRIEF, FROM CHAPTER 4: "NOT A FORMAL DELIVERABLE. A SHARED CLOUD DOCUMENT" FOR THE PROBLEM STATEMENT, OBJECTIVES AND KEY INSIGHTS ABOUT CONTEXT. CH. 4.
Five people, one minute each
Read one of your three questions, and how you will answer it.
Say
The question as written. Which research activity answers it. What you will read or run. The date.
We listen for
A question that is really a solution.
We listen for
A source you could actually open tonight.
We listen for
The empty context box. Saying it out loud is the most useful thing that happens today.
Before you leave
The work between now and Wednesday.
Tonight
BRIEF.md saved with the five headings, your version-zero assertions, and three questions with dates. One line in section 5 if the machine did anything today.
By Friday
Question one answered from published sources, with the source named in the brief.
Wednesday · Sep 9
The audience: three contextual statements about users, and which one you lean to. Read the rest of Brown chapter 2, Hall chapters 6 and 9, and Davis chapter 10, the eleven questions.
Sep 14, then Sep 21
The problem statement, one page, every claim sourced. Then the pitch on Sep 21: Brown's five sections, recorded on video, about ten minutes.
Close
"There is no design without discovery."Dan Brown, the first line of the book's conclusion. See you Wednesday.
meaninglayer.com/course/business-of-design
