The first 3 interviews are free · Start free or see the prices

Thematic Analysis Codebook: Starter Codes for Interviews

Use this codebook as the first pass on interview transcripts about a task or a workflow. The 10 codes describe what people do and why, so themes come from behavior, not from opinions.

Analyze a transcript free

Use it for

  • Transcripts of 6 to 20 interviews about one task, workflow or decision.
  • Notes that record the participant ID, the segment and the date of each interview.
  • The research question of the study, written as one sentence.

Steps

  1. Read before you code

    Read 3 transcripts from start to end without codes. Write one line about each participant: the goal, the main problem and the tools. This gives you the context for each passage.

  2. Set the unit of coding

    Code one answer to one question as the unit. If an answer has two ideas, split it into two passages. Each passage can have more than one code, but most passages need only one.

  3. Pilot the codebook with two coders

    Two people code the same 2 transcripts alone. Then compare each passage. Where the codes differ, find the rule that caused the difference and make it clearer.

  4. Add codes only with a rule

    If a passage fits no code, put it in a holding list. Make a new code when 3 or more passages share an idea. Write its definition, include rule and exclude rule before you use it.

  5. Code all transcripts

    Code the rest of the transcripts with the final codebook. Record the participant ID for each passage. Count participants per code, not passages, because one talkative person can give many passages.

  6. Build themes from codes

    A theme is a pattern that explains something, not a code name. Look for codes that occur together, for example a Trigger before a Workaround. Write each theme as one sentence with a cause and an effect.

  7. Check each theme against the data

    For each theme, read all passages under it again. Look for participants who say the opposite. Keep the theme only if the passages support it, and report the opposite cases.

Codebook

Copy the codebook as a table. It pastes into a spreadsheet with one row for each code.

Goal

The result that the participant wants to get from the task, stated as an outcome.

Include
The participant says what they want to achieve or what success looks like for them.
Exclude
The participant asks for a specific feature or change. Use Request for that, because a request is a solution and a goal is an outcome.
Example
“At the end of the month I just want one number I can give my boss without checking it three times.”

Current process

The steps that the participant takes now to do the task, in the order they take them.

Include
The participant describes what they do, which tool they open, or who does which step.
Exclude
The step exists only because a tool or process fails. Use Workaround for that.
Example
“First I export the orders, then I paste them into the shared sheet, then Tom checks the totals.”

Pain point

A problem inside the current process that costs time, money or effort, or that causes errors.

Include
The participant names a step that is slow, hard, repeated or that goes wrong.
Exclude
The limit comes from outside the process and the participant cannot change it, such as budget or company policy. Use Constraint for that.
Example
“The export always cuts off the last day, so every month I find the missing orders by hand.”

Workaround

A step, file or habit that the participant made to get around a gap in a tool or process.

Include
The participant describes a custom fix: a side spreadsheet, a reminder, a copy and paste routine or a second tool.
Exclude
The step is part of the normal process and does not replace a missing function. Use Current process for that.
Example
“I keep a sticky note on my screen with the customer codes because the system will not let me search by name.”

Trigger

A specific event that made the participant start, stop or change a behavior.

Include
The participant names a moment in time: a new job, a failed deadline, a new manager or a price change.
Exclude
The participant describes a general, ongoing problem with no event. Use Pain point for that.
Example
“When we missed the tax deadline last spring, my manager told me to find a better way.”

Decision criterion

A factor that the participant uses to compare options or to choose one.

Include
The participant explains why they picked, rejected or would pick a tool, vendor or method.
Exclude
The factor is a fixed limit that removes options before any comparison. Use Constraint for that.
Example
“The one that won was the one my accountant could log into without a new account.”

Other people

The influence of another person or role on how the participant does the task or makes a decision.

Include
The participant mentions a manager, colleague, client or family member who approves, blocks, helps or sets an example.
Exclude
The person is mentioned only as a step in the process with no influence on choices. Use Current process for that.
Example
“Our finance director has the final say, and she does not trust anything that is not in Excel.”

Constraint

A limit outside the participant's control that shapes what they can do.

Include
The participant names budget, policy, law, available skills, time or hardware as the reason for a choice.
Exclude
The participant can change the limit or the limit is a step that goes wrong. Use Pain point for that.
Example
“IT only allows tools that are already on the approved list, so I could not even try it.”

Emotion

A feeling that the participant states about the task, a tool or an outcome.

Include
The participant uses a feeling word such as worried, relieved, annoyed or proud, or describes a clear emotional reaction.
Exclude
The passage only says that something is good or bad with no feeling. Code the cause instead, for example Pain point.
Example
“Honestly I dread the first Monday of every month. I sleep badly the night before.”

Request

A direct ask from the participant for a feature, a change or a new product.

Include
The participant says what they want built or changed, for example "I wish it had" or "you should add".
Exclude
The participant describes an outcome without a solution. Use Goal for that.
Example
“If you could just add a button that sends the report to my boss, that would be perfect.”

Report outline

  1. Research question and method: the number of interviews, the segments and the dates.
  2. Main themes, each as one sentence with a cause and an effect, ordered by how many participants support them.
  3. Evidence for each theme: 2 or 3 quotes with participant IDs and the number of participants who fit.
  4. Opposite cases: participants who do not fit a theme, and what is different about them.
  5. Requests, listed apart from the themes, with the goal behind each request.
  6. Open questions and the next study to answer them.

Mistakes to avoid

  • Do not use the codes as the themes. "Pain point" is a code. "Manual exports cause month-end errors" is a theme.
  • Do not count passages. Count participants, or one person who talks a lot will look like a pattern.
  • Do not add a new code each time a passage feels different. Hold it, and make a code only when 3 or more passages share the idea.
  • Do not report requests as needs. Find the Goal behind each Request before you include it in a finding.

Guides that use it

Discussion guideConcept Test Interview: Guide and Questions for a New Product IdeaUse this guide when you have a product idea on a page, an image or a short video. It checks if people understand the idea, if it fits their week and what stops them from trying it.Discussion guideMessaging Test Interview: Value Proposition Testing GuideUse this guide to test the words on your homepage, ad or sales deck. It captures the words customers use for the problem, checks what they remember and compares two versions.Discussion guideBrand Perception Interview: Discussion Guide and QuestionsUse this guide to learn what people think of your brand and why. It starts with unaided recall, then moves to associations, trust, past contact and comparison with other brands.Discussion guidePower User Interview: Guide and Questions on Why They StayUse this guide with your heaviest users. It finds the moment they became daily users, the workflows and workarounds they built, and the limits they hit.Discussion guideAI Feature Trust Interview: Guide and QuestionsUse this guide with people who use an AI feature in your product. It finds which tasks they give to the AI and how they check results. It also finds what a wrong result did to their trust.Discussion guideCart Abandonment Interview: Discussion Guide and QuestionsUse this guide with shoppers who added items to the cart in your online store and left without paying. It finds the step where they stopped, the doubt behind it and where they went next.Discussion guideLanding Page First Impression Test: Interview Guide and QuestionsUse this guide to test a landing page with people from your target audience who do not know your product. It checks the first glance, what people remember, and what they need before they act.Discussion guidePackaging Test Interview: Discussion Guide for Consumer ProductsUse this guide to test a new or changed pack design with shoppers in the category. It starts with a shelf image, then looks at the front, the back and the claims. It ends with value and fit.Discussion guideAd Concept Testing: Interview Guide and QuestionsUse this guide to test 1 to 3 ad concepts, from storyboards to finished cuts, with your target audience. It checks recall, brand link and main message before it asks about feelings and fit.Discussion guideEmployee Experience Interview: Questions and Discussion GuideUse this guide to understand how employees experience their work, beyond what an engagement survey shows. It asks about a recent week, high and low moments, support, growth and the reasons people stay.Discussion guidePatient Experience Interview: Non-Clinical Discussion GuideUse this guide to learn how people experience a care service: booking, arrival, waiting, communication and follow-up. It stays away from clinical details and gives the moderator clear steps for difficult moments.

Questions

A codebook is a list of codes. Each code has a name, a definition, an include rule and an exclude rule. It lets two people code the same data in the same way.

Use both. Start with these 10 codes as a first pass. Then add codes from the data when 3 or more passages share an idea that no code covers.

Most interview studies need 8 to 20 codes. Fewer codes make the coding faster and more consistent. If you have more than 25 codes, merge the ones that coders often confuse.

Yes. The transcripts are text like any other interview. In User Evaluation, each claim in a study report links to a clip. You can code those clips to check the themes.

Ask real people your question

The AI moderator interviews screened people by voice or text. Each claim in the study report links to its clip. The first 3 interviews are free.

Start free