How to Find Themes Across Interview Transcripts

A theme is a recurring idea, subject, experience, or pattern that appears across the interview material.

It can be something obvious, such as:

• company growth
• a major challenge
• a product launch
• a difficult decision

But themes can also be more conceptual:

• fear of failure
• changing identity
• lack of trust
• learning from mistakes
• conflict between teams
• adapting to uncertainty
• unexpected consequences

A theme doesn't have to be explicitly stated by every interviewee.

It can emerge from the material as you compare what different people say.

For example, one person might talk about working longer hours, another about losing customers, and another about pressure from investors.

Individually, these are different answers.

Together, they might reveal a broader theme:

The pressure created by rapid growth.

That connection is part of editorial discovery.

When you're editing a single interview, you can often work chronologically.

With multiple interviews, chronology becomes much less useful.

Interview A might discuss the beginning of a story.

Interview B might explain what happened later.

Interview C might provide an entirely different perspective on the same event.

The material is distributed across different conversations.

Themes give you another way to organize it.

Instead of thinking:

Interview A → Interview B → Interview C

you can start thinking:

Problem → Response → Turning Point → Consequence

And then find the strongest material from each interview for every part.

This is especially useful when the story isn't fully defined before you begin editing.

One of the easiest ways to miss themes is to treat every interview as a completely separate unit.

You watch Interview A.

You finish it.

You move to Interview B.

You finish it.

Then Interview C.

By the end, you've reviewed everything—but you may not have noticed that all three people talked about the same idea.

Instead, once your transcripts are searchable, start looking across interviews.

For each major topic, ask:

• Who talks about this?
• What do they say?
• Do they agree?
• Do they disagree?
• Do they describe the same event?
• Does one person provide information the others don't?
• Does the topic appear repeatedly?
• Does the meaning change depending on who is speaking?

This turns transcript review into a comparative process.

Don't try to define the final themes immediately.

Start with broad topics.

For example, a project about a new product might contain:

• development
• customers
• competition
• failures
• funding
• launch
• team
• future

These aren't necessarily your final themes.

They're simply categories that help you organize the material.

As you collect quotes, you may discover that some topics are closely connected.

For example:

Funding + team + deadlines

could eventually become:

The pressure of scaling the company.

The theme becomes clearer as you see more of the material.

One of the simplest ways to identify themes is to look for concepts that appear repeatedly.

Search transcripts for:

• recurring topics
• important names
• events
• places
• problems
• decisions
• emotions
• outcomes

Don't only count exact words.

Different interviewees may use different language to describe the same thing.

One person might say:

“We were under enormous pressure.”

Another:

“There was no room for mistakes.”

Another:

“Everyone was working around the clock.”

The language is different.

The underlying idea may be the same.

This is where semantic thinking becomes important.

Imagine you're trying to understand why employees were leaving a company.

You might search for:

• leaving
• quit
• resigned
• turnover
• retention
• unhappy
• burnout
• workload
• management
• culture

You may discover that nobody repeatedly says:

“employee retention.”

But the transcript still contains extensive material about the subject.

The theme exists in the meaning of the answers, not necessarily in one repeated keyword. That same idea sits behind finding a specific quote — except here you are looking for a pattern, not a remembered sentence.

Once you've identified potentially related passages, group them together.

For example:

Theme: The First Major Failure

• Interview A — Describes what went wrong.
• Interview B — Explains why the team didn't see it coming.
• Interview C — Describes the immediate consequences.
• Interview D — Explains what changed afterward.

Now you have a collection of material that can be evaluated together.

This is much more useful than having four unrelated transcript highlights. Quote evaluation still happens inside each group — see how to find strong quotes in an interview transcript.

Sometimes the strongest themes emerge from answers to the same question.

Imagine asking five people:

“What changed after the launch?”

You might get:

• “Everything became much more complicated.”
• “We suddenly had customers everywhere.”
• “We realized the product wasn't ready.”
• “The team doubled in size.”
• “We had to completely change how we worked.”

These answers don't repeat the same sentence.

But together they describe a larger transition.

You can now investigate that transition as a potential theme.

Repeated agreement can indicate that an idea is important.

If several independent interviewees mention:

• the same challenge
• the same event
• the same turning point
• the same person
• the same consequence

that repetition is worth investigating.

It doesn't automatically mean the idea belongs in the final edit.

But it tells you that the material may have editorial significance.

Contradictions can be just as valuable as agreement.

Two interviewees might describe the same event completely differently.

One says:

“It was the best decision we ever made.”

Another says:

“We didn't really have a choice.”

That's not just repetition.

It's a relationship between two perspectives.

The contrast itself may become part of the story.

When comparing transcripts, therefore, don't only ask:

“What do people agree about?”

Also ask:

“Where do they disagree?”

A theme doesn't need to appear in every interview.

Sometimes one person provides the missing piece that makes another theme understandable.

For example:

• Interview A describes the problem.
• Interview B describes the consequences.
• Interview C reveals the decision that caused it.

The third interview may appear unrelated until you compare the material.

This is why thematic discovery should include both recurring patterns and unique information.

Themes rarely exist completely independently.

One can lead to another.

For example:

Rapid growth

↓

More customers

↓

Operational problems

↓

Team conflict

↓

Major restructuring

This starts to resemble a narrative structure.

The themes aren't just categories anymore.

They're becoming relationships between events and ideas.

That is often where story discovery begins. See what story discovery actually is.

This distinction is useful.

A topic is what people are talking about.

A theme is the larger idea that emerges from what they are saying.

For example:

Topic: Funding

Theme: The company had to sacrifice long-term plans to survive short-term financial pressure.

Or:

Topic: Leadership

Theme: The founder's leadership style changed as the company grew.

The topic helps you find the material.

The theme helps you understand what the material means.

The questions asked during the interviews can provide an initial structure.

If every interview includes questions about:

• origin
• biggest challenge
• turning point
• outcome
• lessons learned

you already have several categories to investigate.

But don't limit yourself to those categories.

Interviewees often reveal important information outside the question they were answering.

The transcript can help you discover those unexpected connections.

Once you've gathered enough material, create a simple theme map.

For example:

THEME 1 — Why the company started

Interview A, Interview C

THEME 2 — Early problems

Interview A, Interview B, Interview D

THEME 3 — The turning point

Interview B, Interview C

THEME 4 — Different views of the decision

Interview A, Interview D

THEME 5 — What changed afterward

Interview B, Interview C, Interview D

This gives you a much clearer view of the project than a collection of isolated transcript highlights.

Once themes are visible, you can start asking structural questions.

For example:

• What needs to come first?
• Which theme explains the central problem?
• Which theme contains the turning point?
• Where do different perspectives need to be contrasted?
• Which themes can be combined?
• Which themes don't contribute to the story?

At this point, you're moving from discovery toward story structure.

But don't rush this transition.

The purpose of thematic organization is to make the possibilities visible before committing to a structure. For turning selected material into a narrative, see how to turn interview transcripts into a story.

Here's a repeatable process for multi-interview projects.

1. Transcribe the interviews — Make the spoken material searchable.
2. Read across the transcripts — Don't treat every interview as an isolated document.
3. Identify broad topics — Mark recurring subjects, events, people, problems, and ideas.
4. Search for related concepts — Look for different language describing the same underlying idea.
5. Collect candidate passages — Save relevant quotes and surrounding context.
6. Group material — Organize passages by topic or emerging theme.
7. Compare interviewees — Look for agreement, disagreement, and unique perspectives.
8. Identify relationships — Ask how different themes connect to one another.
9. Create a theme map — Make the major patterns visible.
10. Start testing story structures — Only now begin thinking about how those themes might become a narrative.

AI can be particularly useful when the number of interviews becomes difficult to manage manually.

Instead of reading every transcript linearly, you can use AI-assisted discovery to explore questions such as:

• “What topics appear across these interviews?”
• “What do the interviewees repeatedly mention?”
• “What themes emerge around the company's biggest challenge?”
• “Where do interviewees disagree about the decision?”
• “Which interviews discuss the same event?”
• “Group these answers by theme.”
• “What different perspectives exist on the launch?”
• “Which topics appear across the largest number of interviews?”

This can help surface relationships that would otherwise require extensive manual searching.

But AI-generated themes should be treated as starting points for editorial exploration, not final conclusions.

The editor still needs to examine the underlying material.

A model might group two passages together because they discuss similar concepts, but that doesn't necessarily mean they belong together in the final story.

This distinction is particularly important in interview-driven editing.

AI can help answer:

“What material is related?”

The editor needs to answer:

“What does this mean for my story?”

Those are different questions.

An AI system can identify recurring language, concepts, and relationships.

But editorial decisions depend on context, intention, performance, structure, pacing, and the larger body of footage.

The goal is not to replace the editor's interpretation.

It's to reduce the time spent manually searching for the material that needs interpretation.

Most weak theme maps aren't weak because the interviews had no patterns. They are weak because the categories were decided before the material was compared.

If you decide what the story is before exploring the material, you can start ignoring information that doesn't fit your initial idea.

A word appearing frequently doesn't necessarily mean it represents an important theme.

People can express the same idea using completely different vocabulary.

Contradictions and different perspectives can be equally valuable.

Two passages may appear related until you listen to the original footage.

If every small subject becomes its own theme, the organization becomes difficult to use.

“Leadership” is a topic.

“How leadership changed as the company grew” is a potential theme.

There is no fixed number.

The goal is to create enough structure to understand the material without turning the transcript into an overwhelming spreadsheet of categories.

If you have dozens of small themes, look for relationships between them.

For example:

• deadlines
• budget
• workload
• hiring
• stress

might all belong to a larger idea:

The pressure of rapid growth.

The right level of abstraction depends on the project.

Your initial theme map isn't permanent.

Once you start building a rough structure, you'll often discover:

• two themes overlap
• one theme isn't actually important
• another theme needs more evidence
• a missing perspective exists
• a seemingly minor theme is actually central
• one interview contains an important connection you hadn't noticed

This is normal.

Story discovery is iterative.

The workflow can move back and forth:

Transcripts → Themes → Selects → Structure → Transcripts → More Discovery

The ability to return quickly to the source material is therefore extremely valuable.

The purpose of finding themes isn't to create a perfect database of everything everyone said.

It's to make the editorial possibilities easier to see.

You want to know:

• What ideas keep appearing?
• Which perspectives are different?
• Which events connect multiple interviews?
• Where are the strongest stories?
• What information is missing?
• Which themes could support the narrative?

Once those relationships become visible, you can start making stronger decisions about what belongs in the edit.

Multiple interviews create a unique editorial problem.

The story is rarely sitting inside one transcript.

It's distributed across conversations.

One person provides context.

Another provides emotion.

Another provides a contradiction.

Another provides the detail that connects everything.

Finding themes allows you to bring those pieces together.

You move from:

Interview → Quote

to:

Multiple interviews → Patterns → Themes → Relationships → Story possibilities

That is a fundamentally different way of working with interview footage.

The more interviews a project contains, the harder it becomes to remember who said what and where.

Supacut helps editors explore interview transcripts across a project, identify themes, compare answers, group related material, find potential soundbites, and navigate back to the original footage.

The goal isn't to have AI decide what your story should be.

It's to make the relationships inside your interview material easier to explore.

You decide what the story is. Supacut helps you see what's inside the interviews.

Your story. Your edit.

Start with how to review hours of interview footage efficiently and how to find strong quotes. For a related conceptual guide, see identifying narrative themes across interviews and what story discovery is.