Weeks 2-3: Customer Development & Talking to People
Part of the "Building My Second Startup from Scratch" series
In my last post, I covered how I research startup ideas through frameworks and competitive analysis. Now comes the part everyone talks about but few do well: actually talking to people.
I’m currently in week 6 (ish) of building my new startup, but I’m a couple weeks behind on these posts. Today’s piece covers weeks 2-3: customer development and the mistakes I made while validating my initial ideas through conversations.
Let me be upfront: these early interviews were messy. I made a lot of mistakes. But that’s exactly why I’m writing this. To share what didn’t work and what I’m doing differently now.
The Discovery Approach That Failed
When I started talking to potential customers, I did what most frameworks tell you to do: I picked a niche (salespeople) and asked general discovery questions.
The idea was start broad. Ask open-ended questions about their workflow, pain points, and day-to-day operations. Then narrow down to a specific problem (hopefully)
It didn’t work.
I conducted 10-15 interviews across a few different niches, primarily focusing on salespeople as my target audience.
Here’s what went wrong:
Problem 1: It’s Too Vague
When you ask someone “What do you do on a daily basis?” or “Tell me more about your workflow,” they feel interrogated. They think you’re interviewing them for a job. People don’t enjoy that feeling.
More importantly, they weren’t opening up. Without showing you understand their space or have expertise in it, they resented giving you their time. Why would they spend 30 minutes with a stranger asking vague questions?
Problem 2: You’re Not Showing Value
If you show up as someone knowledgeable in their space, they’re more receptive. If you show up as someone just fishing for ideas, they close off.
The people I was talking to (particularly in sales and marketing) are tech-savvy professionals saturated with solutions. One person told me directly: “We pay for too many tools and don’t use half of them.”
These people weren’t desperate to solve a problem. They were drowning in tools.
The Results Were Useless
After 10-15 interviews with this approach, I had no meaningful insights. No clear problems. No real validation signals. Just a pile of vague conversations that went nowhere.
The people I was targeting were overloaded with tools and solutions. They weren’t actively searching for a new one. And without me positioning myself as someone worth listening to, they had no reason to invest time.
The Approach That I’m Testing Now
Instead of starting with a general audience and narrowing down, I decided to:
Come up with a specific hypothesis for a problem
Come up with a specific hypothesis for a solution
Find the audience and ICP based on that hypothesis
This is a bit counterintuitive if you follow most lean startup advice. Most experts will tell you: talk to people first, figure out the problem second, then build the solution.
I think starting with a hypothesis is more productive.
Here’s why: when you have a specific idea, you can position yourself as someone knowledgeable. You can share your experience. You can relate to their challenges. You become someone worth talking to. And be more specific.
2 Types of Interviews
I split my customer development into two buckets:
1. Expert Interviews
An expert is someone with status and knowledge in the space. Someone who’s been there, done that.
My goal: Get a shortcut to answers. Understand the space at a high level.
Why it matters: If I’m researching B2B SaaS sales, I don’t need to interview 20 salespeople. I can talk to someone who’s run a sales-led SaaS product for 5+ years and who understands the landscape.
Example: If you talk to me about marketing automation (my space as someone who sold Encharge), I can give you an objective view of the entire market because I’ve run a product in that space and exited. I understand the players, the problems, the trends.
2. Buyer Interviews
A buyer is the actual ICP—the person who would potentially buy your product.
My goal: Understand their workflow, their challenges, their buying behavior.
Why it matters: This is where you validate if the problem is real and if your solution could actually work.
How I Structure Expert Interviews
With experts, here’s my approach:
Step 1: Share Your Experience & Context
I open with something like:
“I’ve run a B2B SaaS product with a sales motion similar to what you’re doing, self-serve plus sales-led. I’m trying to help companies like yours solve specific problems.”
This immediately positions me as someone who isn’t wasting their time. I have skin in the game. I understand their world.
Step 2: Share Your Problem
I then share the specific problem I’m trying to solve:
“I struggled with lead qualification, getting trials to convert into demo calls, and closing those demos into paying customers.”
Why? So they can relate to me. So they know I’m not a random person asking random questions.
Step 3: Share Your Value
I mention my experience working with similar companies. The processes I’ve helped them implement. The results I’ve seen. This further reinforces that I’m a credible person having this conversation.
Step 4: Share Your Agenda
“My main goal is to help companies like yours solve these specific problems. I want to understand what tactics and strategies have worked best for high-performing salespeople.”
This frames the conversation as collaborative, not extractive.
Step 5: Ask Informed Questions
Now you can ask questions like:
“What would be the tactics and strategies that have worked best for high-performing sales reps in your space?”
You’re asking the expert about their ICP. You’re trying to understand your buyer through their lens.
Result: Experts are relieved when they hear you have experience. They feel like you’re worth talking to. The conversation flows naturally.
How I Structure Buyer Interviews
With actual potential customers, I take a different approach:
Focus on the Past and Present, Not the Future
People are terrible at predicting the future. They’re great at describing what they’ve already done.
So I focus on real experiences:
What are the KPIs you track? (How do they measure success?)
How do you budget for software? (What’s their buying process?)
Who’s the economic buyer in your company? (Who approves the budget?)
Have you tried tools like this before? If not, why?
Understand Their Role Deeply
How do you run a demo? How do you close?
What’s your typical deal size?
What’s your sales stack right now?
What’s your customer profile?
Understand the Market
How much time does follow-up take your team?
Have you paid for tools to speed up sales before?
What’s the urgency around solving this?
My Honest Note: I rarely ask all these questions during a call. Conversations naturally flow. These questions are more like compass points, I use them to guide where the conversation goes.
Implicit Feedback (a.k.a Reading Between the Lines)
Here’s the critical insight I learned: Most people will not give you explicit answers.
They won’t say: “This is the problem I have and I desperately need a solution.”
Instead, you have to read between the lines. Listen for implicit feedback.
Example: The Digital Sales Room Idea That I Invalidated
Let me give you a concrete example.
I had an idea for an AI-powered Digital Sales Room (DSR) tool. Here’s my hypothesis:
The Problem: Sales reps manually create digital sales rooms (personalized micro-sites they send to leads after calls). They have to drag and drop templates, PDFs, playbooks, task lists — all manually.
My Solution: What if we used AI to automate this? One-click DSR generation?
It sounded solid. But when I talked to potential customers, I realized it wouldn’t work.
How I Found Out (Without Explicit Feedback)
I asked salespeople about their DSR workflow. Here’s what I heard implicitly:
“We only do DSRs for very high-value, enterprise deals.”
“We don’t do that many of them.”
“I generally hate doing DSRs.”
“The potential buyers don’t even look at the DSRs we send.”
“We have templates for everything. We just drag and drop. It takes 15 minutes, max.”
What This Meant:
Digital sales rooms aren’t used at scale. They’re used for a handful of enterprise deals. The process doesn’t take hours. There’s no pain point that needs automation.
An AI-powered solution to automatically generate DSRs? There’s no market for it because the problem. Manual work isn’t actually a problem. It’s a non-issue.
I learned this not because someone said “This won’t work.” I learned it by listening to what they did and didn’t say.
TLDR;
The shift from general discovery to hypothesis-driven customer development changed everything:
I positioned myself as credible. Not as a random person asking questions, but as someone with experience.
I focused on real experiences. Past and present. Not hypothetical futures.
I learned to listen between the lines. Implicit feedback is more valuable than explicit answers.
I invalidated ideas faster. With signal from just 3-5 conversations, I could confidently move on.
I wasted less time. Instead of 10-15 vague discovery calls, I could validate or invalidate an idea in 4-5 focused calls.
The Results
I’ve used this approach to invalidate most of my initial ideas. The DSR idea is gone. My next idea—a customer intelligence platform—didn’t pass muster either.
But I also found one idea that’s generating real traction: AI agents for Slack.
This was a pivot from the customer intelligence idea, and early conversations are showing more promise. I’m building a landing page and am currently reaching out to people to validate.
The Key Takeaways
If you’re doing customer development, avoid what I did wrong:
❌ Don’t use vague discovery questions. Position yourself as credible.
❌ Don’t ask about the future. Ask about the past and present.
❌ Don’t wait for explicit feedback. Learn to read between the lines.
❌ Don’t do 15 vague calls. Do 4-5 focused calls instead.
✅ Do come with a hypothesis. It makes you more credible.
✅ Do share your experience. Show you understand their world.
✅ Do focus on real experiences. Past and present, not hypothetical futures.
✅ Do listen for what they don’t say. That’s often more valuable.
What’s Next
In my next post, I’m planning to dive deeper into my validation sprints and pivots. I’ll walk through:
The pivot to AI agents for Slack
Whether I actually landed that pilot customer
Thanks for reading. See you in the next one.
If you’re finding this series helpful, consider subscribing. I’m documenting everything as I build this startup in real-time