How to Decide Your UX Interview Sample Size
Stop guessing your UX interview sample size. Saturation is a measurable design parameter, not a feeling. Here's how to plan for it.
Imagine walking into a massive department store looking for something fresh. For the first few minutes, everything feels new: different colors, cuts, and styles everywhere you look. However, after wandering through a few aisles, you start to notice a pattern. The clothes begin to repeat, and no matter which direction you turn, the variety ceases to expand. At this moment, you stop searching, not because the store has physically ended or because you are tired, but because you realize the environment has nothing new to offer you. This everyday phenomenon perfectly illustrates the most critical, yet often misunderstood, concept in qualitative research: saturation. In UX Research, determining interview sample size should never be a guessing game or blind adherence to the famous “rule of five users,” but rather a rigorous methodological pursuit of this point of analytic redundancy.
The central problem facing our industry today is treating saturation as a “feeling” rather than a measurable outcome. It is all too common to hear teams say they will stop interviewing when they “feel” like nothing new is coming up. However, qualitative literature is explicit that this intuition-based logic severely compromises data validity. True data saturation occurs when additional data collection stops producing new information relevant to the research question, ensuring the study is replicable. It is not about exhausting time, budget, or the researcher’s patience, but about exhausting the phenomenon being studied. If we stop too early based on a superficial sense of repetition, we risk building products on incomplete views; if we stop too late without criteria, we waste valuable resources.
To improve our practice, we must distinguish between “code saturation” and “meaning saturation.” Code saturation is what usually happens quickly, often within that classic range of nine to seventeen interviews for homogeneous groups. It is the point where you have heard all the high-level issues, and no new major topics are appearing. Many UX teams stop here, mistakenly believing they have fully grasped the user experience. The danger lies in ignoring meaning saturation, which is far deeper and takes longer to achieve. Reaching meaning saturation requires continuing the investigation until one understands all the nuances, contradictions, and contextual variations of those themes. Knowing what the issues are is vastly different from deeply understanding why and how they manifest differently across various user profiles.
Therefore, the answer to “how many users do I need to interview?” should never be a fixed number, but rather a research design strategy. Studies focused on simple workflows with homogeneous users may indeed saturate quickly. Conversely, research exploring complex mental models, cultural behaviors, or decision-making processes requires significantly larger samples to capture pattern stability. Rigor in UX Research demands that we abandon the convenience of universal “golden rules” and start planning saturation as an explicit design parameter. We must monitor the intake of new information with every interview and possess the intellectual honesty to admit whether we have merely listed topics or if we truly understand the depth of the human experience we are attempting to design for.
Originally published on Medium.