Chapter 3: Ask what happened, not what people want, lesson 5 of 7
Write a plan someone else could run
Write a research plan someone else can follow, and get the same kind of answers. The question, who to talk to, consent, what to ask, and what to note.
Warm-up
From last time, watching Sara order. Is each note something you could see, or a guess?
-
Right, it's something you could see. Anyone watching would agree. You could count it.
Not quite. It's something you could see. Anyone watching would agree. You could count it.
-
Right, it's a guess. It's a guess about how she feels.
Not quite. It's a guess. It's a guess about how she feels.
-
Right, it's something you could see. A count, at a time.
Not quite. It's something you could see. A count, at a time.
Last time, you watched Sara order lunch, and wrote down what you could see. The team now has six interviews and one morning of watching.
Next week, Mia is away. Nora Kim, the growth manager, will run two more interviews. Peter, the head of sales, has an idea for what they should find out.
Think first
Peter, head of sales, to the team: Nora’s question should be simple. Do offices want group ordering?
How would you write Nora’s research question?
You
You
What this means
A research question asks what you need to learn, not what you want to build. Peter’s question can only get a yes. Mia’s can find a problem nobody expected.
What goes in the plan
Mia’s plan has to work without her in the room. So it says everything she would have done in her head.
- The research question. What you need to learn.
How do offices get lunch for a group today, and where does it go wrong?
- Who to talk to, and how to find them. Written as what people do, not their job title.
- Consent, word for word. Who you are, what you’ll record, that there are no right answers, and that they can stop.
- The questions, in order. Five about the past, and one
show me
. With a note on what to follow up. - What to write down. Things you can see, each with a number. Quotes in the person’s own words.
- The debrief, straight after. Three things that surprised you, before you forget them.
Find the right people
Practice
Nora needs people who organise lunch for a group. Which of these questions would find them?
-
Right, it's a question that finds the right people. It asks what they did, recently. Only people who really organise lunch say yes.
Not quite. It's a question that finds the right people. It asks what they did, recently. Only people who really organise lunch say yes.
-
Right, it's a question that could let the wrong people in. A job title. Some office managers never order lunch, and Ken isn't one, but he does.
Not quite. It's a question that could let the wrong people in. A job title. Some office managers never order lunch, and Ken isn't one, but he does.
-
Right, it's a question that could let the wrong people in. Age says nothing about who orders lunch.
Not quite. It's a question that could let the wrong people in. Age says nothing about who orders lunch.
-
Right, it's a question that finds the right people. It finds the payers, and tells you which kind of office it is.
Not quite. It's a question that finds the right people. It finds the payers, and tells you which kind of office it is.
-
Right, it's a question that could let the wrong people in. An opinion. People who never do it can still answer.
Not quite. It's a question that could let the wrong people in. An opinion. People who never do it can still answer.
Questions like these, asked before anyone is invited, are a screener. A good screener asks what people did, not who they are.
How many interviews are enough?
The six interviews so far show a pattern. Three of the five managers, Anna, Dana and Ken, struggle to get paid back. But Ray’s company pays for lunch, so his trouble is paperwork. And Sara’s money is already solved.
So the pattern is real, but new kinds of office are still turning up. That’s why two more interviews are worth it. When new interviews stop showing you anything new, you’ve reached saturation, and you can stop.
The test: could a stranger run it?
Here’s how to check a plan. Hand it to someone who wasn’t there. Could they run the same interviews, and get the same kind of answers? Check each part:
- The research question is written down, in one sentence.
- It says who to talk to, and the screener questions that find them.
- Consent is written word for word.
- The questions are in order, with what to follow up.
- It says what to write down, with numbers.
- It says how long, and where.
If any part lives only in your head, the next person will fill it in differently. Then their answers can’t be compared with yours.
Think first
Nora’s first interview comes back full of views on group ordering. There are no stories about real days.
Which part of the plan most likely went missing?
You
You
What this means
Opinions come from questions about the future. A plan that lists the questions, starting with Tell me about the last time
, stops that. Mia adds them, and Nora’s second interview is full of stories.
A plan with every part written down has a name: a protocol. It’s what makes research something others can repeat, not just something you did once.
Words from this lesson
- Screener
- Questions that decide who you invite, written about what people do. In this lessonIn the last two weeks, did you order lunch for five or more colleagues?
- Protocol
- Everything someone else needs to run the same study. In this lessonMia's plan: the question, who, consent, the questions, what to note.
- Saturation
- The point where more interviews stop teaching you anything new. In this lessonNew interviews stop showing new kinds of office.
Back to your first answer
Your first answer
The question
Your answer
Your answer will show here after you write one at the top of this page.
Your bet: what should Mia do with her notes first?
Your bet
The interviews are done. Mia has the notes from every interview, a morning of watching, and 40 help requests.
Mia, research lead, to you: Peter wants an answer by Friday. Where do we start?
What should Mia do with her notes first? Make your bet, and say how sure you are.
You'll find the answer in Chapter 4: Turn notes into decisions.
Try this during the week: write a three-line screener for people who use an app you know. Ask what they did, not who they are.
Sources
- Steve Portigal, Interviewing Users, second edition, 2023. On planning interviews and writing screeners.
- Greg Guest, Arwen Bunce and Laura Johnson, How Many Interviews Are Enough?, Field Methods, 2006. On saturation.
- Tiffin, Mia, Nora, Peter, Anna, Dana, Ken, Ray and Sara are made up for this course.