This article explains how we conducted the pre-launch customer interviews for Betty UK, including our approach to asking questions and the tools we used.

Our interview approach

How our approach evolved

We initially started with a list of key topics we wanted to understand, with corresponding questions for each topic.

In practice, we were asking questions almost like a conveyor belt: moving from one prepared question to the next, regardless of how naturally the conversation was developing.

This created two problems:

• users often moved into unrelated topics before we had covered the areas we wanted to explore;
• even when users stayed on topic, they often found it difficult to describe their habits procedurally or deeply when asked direct questions on command.

To address this, we shifted to a waterfall/funnel questioning method. Instead of moving through a fixed list of questions, we used a more fluid discovery process. This allowed users to speak naturally, while still giving the interviewer a structure for guiding the conversation towards the key topics we needed to cover.

The waterfall/funnel method

The waterfall/funnel method is a fluid discovery process. This approach is built on four core principles:

1. The Vague Opener: We start with a single, highly vague question: "Walk me through your typical play session…" This allows the user to speak their mind freely without being led towards a specific answer.
2. The Gentle Pivot: When we hear a specific keyword related to one of our planned topics, we gently pivot into that area. For example, if the user naturally mentions bonuses, the interviewer can follow that thread with a targeted follow-up question. We repeat this process until all planned topics have been covered.
3. Systematic Progression: If the user naturally stops and has already answered a topic fully, we move directly to the next broad open-ended question (L1 level of inquiry) on our list.
4. Flexibility & Improvisation: Users often mention things we were not actively looking for up to that point, such as support or live chat expectations. If they hit on something useful, we let them speak on and explore the topic further. This allows us to capture unexpected insights that would not necessarily come up in a rigid interview script.

Four levels of inquiry

We designed our questions with different vagueness levels. This helped us move from broad discovery to deeper clarification without biasing the user.

Goal of the method

The purpose of this structure is not to make users give us the answers we want. It is to help them describe their own behaviour, preferences, and decision-making process as naturally as possible.