1. The Short Answer: The Rule of 20–30
For most early-stage B2B and consumer software concepts, you need between 20 and 30 completed customer interviews within a single, tightly defined buyer persona before you have sufficient qualitative confidence to build an MVP.
Fewer than 15 interviews leaves you vulnerable to idiosyncratic opinions and false positives. Conversely, conducting more than 40 interviews within a single persona usually produces diminishing returns and delays product momentum.
To learn how to source prospective candidates for these interviews, read How to Find People for Customer Discovery Interviews and How Do I Know If My Startup Idea Is Worth Pursuing?.
2. Understanding Theoretical Saturation
In qualitative research, theoretical saturation occurs when additional interviews cease to reveal new themes, workflows, vocabulary, or complaints. When the 22nd prospect describes the exact same Excel bottleneck, the same compliance headache, and the same friction with existing software that the 14th prospect mentioned, you have reached saturation.
If every conversation is still revealing completely different problems, your target persona is defined too broadly. You must narrow your beachhead customer definition.
3. Why Persona Homogeneity Dictates Sample Size
The number of interviews you need depends strictly on the homogeneity of your sample:
| Target Market Complexity | Target Persona Definition | Recommended Sample Size |
|---|---|---|
| Narrow Vertical B2B | Solo Dental Practice Owners in Tier-1 Cities | 15–20 Interviews |
| Two-Sided Marketplace | Freelance Video Editors (Supply) + Ad Agencies (Demand) | 20 Interviews per side (40 total) |
| Broad Prosumer / Creator | Substack writers with >5,000 subscribers | 25–30 Interviews |
4. The 5-Step Non-Leading Interview Structure
To extract unbiased data in 20 minutes, use this interview blueprint:
- Context Exploration (3 mins): "Tell me a bit about your daily workflow and what your primary responsibilities are."
- Workflow Friction (7 mins): "What is the most frustrating, time-consuming part of executing [Process X] today? When was the last time that happened?"
- Quantifying Impact (5 mins): "When that issue occurs, how much time or money does your organization lose trying to resolve it?"
- Current Workarounds (3 mins): "What tools, spreadsheets, or contractor workflows have you set up to manage this? What do you dislike about them?"
- Commitment Test (2 mins): "We are building a dedicated tool to automate this exact workflow. Can I show you our initial prototype next Tuesday?"
5. When to Stop Interviewing and Start Building
Stop interviewing and move to prototyping when:
- You can clearly articulate the top 3 core features required to deliver the primary outcome.
- At least 5 interviewees have asked to be notified immediately upon beta release.
- You have confirmed that target buyers have budget authority and willingness to pay. See How to Know If Customers Will Pay Before You Build.
6. Scaling Customer Discovery with ProdNet
Conducting 30 high-quality customer interviews requires weeks of cold sourcing and scheduling. ProdNet solves this by connecting founders with verified domain experts, buyers, and specialized research cohorts. We conduct, record, and synthesize structured customer discovery sprints so you get actionable insights in days.