The Study You Never Put in the Proposal: How Many Hours of Audio a Qualitative Project Really Is

This is Talk Type, from the team at Pepys, where we turn talk into text.

Here's the number that catches people off guard. You plan a study around interviews, a headcount, a recruitment target. You almost never plan it around hours of audio. And the hours are what actually land on your desk.

Start where the study starts. How many interviews is enough? A systematic review in twenty twenty-two put the typical band at nine to seventeen interviews before you reach saturation, the point where new interviews stop teaching you anything new. And saturation arrives faster than most people expect. One experiment on sixty in-depth interviews found it inside the first twelve, with the big themes showing up by interview six.

So your count sits somewhere between nine and seventeen. Now the second number. How long is each one? The workhorse of qualitative research, the semi-structured in-depth interview, runs anywhere from thirty minutes to several hours. Most teams plan around sixty minutes as a working middle. Not a rule, just a place to start the math.

Now multiply. Nine interviews at an hour each is about nine hours of audio. Seventeen is about seventeen. That's your bracket for a typical interview study. But watch what happens when the conversations run long. Stretch each one to ninety minutes, and the same study climbs to roughly fourteen to twenty-six hours. Same number of people. Nearly triple the pile.

Focus groups feel like they should be worse, because there are more voices in the room. They're actually lighter. Saturation usually lands within four to eight groups, each running sixty to ninety minutes. That's about four to twelve hours total. And here's the part people miss. A group of eight people is still one recording. Eight voices never mean eight hours. Audio tracks the number of sessions and their length, not the headcount.

So where does the surprise come from? It's the gap between the thing you counted and the thing you have to process. You recruited twelve people. What you're holding is twelve hours of recording, and every one of those hours has to become text before analysis even starts. That's the real study. Not the interviews you scheduled, the transcript pile they leave behind.

This is why knowing the number early matters so much. If you've got twelve hours of audio, you can plan the text step instead of getting buried by it. You're not stuck watching a tool quit halfway through a long session, because there's no reason to cap a recording at thirty minutes. Split the audio, transcribe the pieces, stitch the timestamps back together, and the length of any one file stops being the problem.

So before you write a single interview question, do the small multiplication. Count times length. Nine to seventeen sessions, an hour or so each. That's the study nobody puts in the proposal, and it's the one that decides how your next month actually goes.

That's this episode of Talk Type. The full write up, with the links and sources, is in the show notes. Pepys transcribes any file or link, any length, pay once, and we never train on your audio. Your first sixty minutes are free at pepys dot co. Thanks for listening, and we'll see you next time.