What is data saturation?

Also called: Saturation, Thematic saturation, Theoretical saturation

Data saturation is the point in qualitative research where new interviews or observations stop producing new themes or insights, which tells you that adding more sessions would mostly repeat what you already know.

Updated

Where the idea comes from

Barney Glaser and Anselm Strauss introduced theoretical saturation in The Discovery of Grounded Theory (1967): you keep sampling until a category stops gaining new properties. Applied research later loosened it into thematic or data saturation, which is how most product and UX teams use the term today.

How many interviews it usually takes

Guest, Bunce and Johnson (2006) coded 60 interviews and found 92% of their codes had appeared within the first 12, with the main themes present after 6. A 2022 systematic review by Hennink and Kaiser found most studies reached saturation between 9 and 17 interviews, when the audience was fairly similar and the research question narrow.

Those numbers move with your study. Expect to need more sessions when:

  • The audience is varied: more than one role, company size or market. Saturation happens per segment, not across the whole sample.
  • The question is broad, such as how a team runs research end to end, instead of how people use one feature.
  • Interviews are short or loosely structured, so each one covers less ground.

How to tell you've reached it

  • Track new codes or themes per interview as you go. When two or three interviews in a row add nothing new, you're close.
  • Check by segment. If every new theme is coming from one group, the others may not be saturated yet.
  • Write down the stopping rule before fieldwork starts, for example "stop when three consecutive interviews add no new themes", so the decision isn't made on tiredness or deadline.

What saturation doesn't tell you

Saturation says you've heard the range of views in your sample. It doesn't say how common each view is, and it can't fix a biased sample: if you only recruited power users, you'll saturate on power-user themes. Pair it with a survey or product data when you need to size a finding.

Go deeper

Questions

Is 5 interviews enough to reach saturation?

Sometimes, for a narrow question and a very similar group of people. Guest, Bunce and Johnson found the main themes appeared within 6 interviews. For most product questions with more than one user type, plan for 8 to 12 per segment and stop early if themes stop changing.

What's the difference between data saturation and theoretical saturation?

Theoretical saturation comes from grounded theory: you stop when the categories in your developing theory stop gaining new properties. Data or thematic saturation is the looser, more common use: you stop when new sessions stop producing new themes.

Does saturation apply to surveys?

Not in the same way. Surveys use statistical sample sizes based on confidence level and margin of error. Saturation is a judgement for qualitative work, where you're looking for the range of views, not their frequency.

Next step

Run the study, not just the method

Great Question handles recruiting from your own customers, scheduling, incentives, interviews, tests and analysis in one place.

Book a demo

No credit card required.

Related terms

Full glossary
  • Triangulation

    Triangulation is checking a research finding against more than one source, method or researcher, so you only act on it when different kinds of evidence point the same way.

  • Think-aloud protocol

    A think-aloud protocol is a usability research method where participants say what they are thinking, looking at and trying to do while they complete a task, so you hear the reasoning behind each action instead of guessing at it.