Most founders jump straight to building. They spend weeks — sometimes months — writing code, designing interfaces, and setting up infrastructure before a single real user has confirmed the problem is worth solving.

Eric Ries, author of The Lean Startup, called the alternative Validated Learning — the process of proving that people actually want a solution, not just say they do.

At SOTS Network, we work with founders before any code is written. Here are the two methods we use.

The #1 reason startups fail:"No market need" — according to CB Insights analysis of startup post-mortems. Not bad code. Not lack of funding. Building something nobody wanted.
01
The 5-Step Founder Validation System
No code. No landing page. Just real signals from real people.

Most founders overcomplicate validation. They think they need a landing page, a polished MVP, and a thousand users. They don't. You need signals, not scale.

❌ What founders think they need
A landing page
A built MVP
1,000 users
Months of work
✓ What you actually need
Proof of painful problem
15–30 real conversations
Pre-sales or email signups
1–2 weeks of research
Step 01

Nail the Problem, Not the Idea

Your idea might be interesting, but the only question that matters is: is the problem painful enough?

"Would my target audience pay for any solution here?"

If the answer is no, you're building a vitamin. Investors and customers consistently choose painkillers. Look for a problem people are already spending money, time, or energy to solve imperfectly.

Step 02

Create a Simple Concept Offer

Forget prototypes, wireframes, and logos. Write a single paragraph that captures your value proposition:

  • Who it's for
  • What problem it solves
  • How life is different after using it
Example:"Busy freelancers can automate invoice reminders with 1 click — no setup, no spreadsheets."

No UI. No logo. Just clarity. If you can't explain it in one paragraph, the idea needs more work.

Step 03

Find 15–30 Real Target Users

Not friends. Not Twitter mutuals. People who are actively experiencing the problem right now.

  • Reddit communities where your audience vents
  • X (Twitter) keyword searches around the problem
  • Slack groups, Indie Hackers, Facebook communities
  • Cold DMs with genuine context — not pitches

You are not selling. You are learning. The goal is patterns, not permission.

Step 04

Run the Pain Interview

Three questions. That's all you need in each conversation:

  • "What's your biggest frustration with [X]?"
  • "What are you doing today to solve it?"
  • "If a tool did [Y] for you, would you try it?"

After 15–20 conversations, you'll start hearing the same frustrations repeated. Those patterns are your product roadmap. The exact language people use becomes your copy.

Pro tip: People are polite. They will say your idea sounds great to avoid conflict. Only commitments count — pre-orders, email signups, introductions. Compliments are not validation.
Step 05

Offer Pre-Sales or Email Signups

When someone says "I'd use that," push one level further:

"Would you pay for it today if I shipped it in 2 weeks?"

  • "Want early access?" — collect emails
  • "Drop your email to join the beta."
  • Ask for a pre-payment, even $1 — it separates interest from intent
No buy-in = no validation. Interest without commitment is noise.
Bonus: Build demand before you build product. People buy solutions to painful problems — not software features. Your job in validation is to confirm the pain exists and people are ready to pay to end it.
02
AI-Powered Market Research with Claude
Real story: 12 ideas → validated in 10 minutes → $2.3k MRR
📈 Real Founder Story · B2B SaaS
Now at $2.3k MRR

How I Used Claude to Validate My Idea in 10 Minutes

A few months back, a founder had 12 different SaaS ideas scattered across Notion docs — and no clear signal about which one people actually cared about.

The standard advice — "talk to your users," "validate first" — felt abstract. Where exactly are these users? What do you ask them without sounding like you're running a survey?

Like many developers, they ignored the advice and started building. First project: 3 signups. Second project: never made it past localhost.

The classic mistake — building solutions to problems they had, not problems other people were willing to pay to solve.

The AI research approach that changed everything:

01
Activate Claude Research Mode
Open Claude and activate the deep research option for real-time web synthesis.
02
Prompt for Pain Point Research
Prompt Claude to scrape Reddit threads, Quora answers, and G2 reviews focused on a specific problem area — in this case: "cold email personalization problems."
03
Receive Real-User Synthesis
Claude returned a 3-page breakdown — real quotes from salespeople describing frustrations: templates that don't convert, manual personalization that takes hours, reply rates that don't move.
04
Score the Opportunity
Ask Claude to rate the opportunity 1–10 based on demand vs competition. The result: 8.5, with clear reasoning about why the market gap existed — using the research data, not general knowledge.
05
Commit and Build
With enough signal to act, the founder built Introwarm — a tool that generates personalised email openers by analysing what prospects are posting and sharing online.
The result: First paid customer ($29) in week 2 after launch. Now at $2.3k MRR and growing — primarily through cold outreach using the tool itself, and community posts.

Why This Method Works

  • People vent online constantly. Reddit, G2, Quora, and app store reviews are free, real-time market research — if you know how to read the patterns.
  • AI synthesises faster than manual research. Reading through hundreds of complaints manually takes weeks. Claude surfaces the patterns in minutes.
  • You don't need perfect validation.You need enough signal to know you're not solving a problem nobody has.
The key insight: Find where your target users are already complaining — and let them tell you what to build. The best product ideas are hiding in plain sight in online communities.

Vitamins vs Painkillers: The Most Important Filter

Every startup idea falls into one of two categories. Knowing which one yours is determines whether validation will even work.

💊 Vitamins (nice-to-have)
People appreciate it
They might use it occasionally
They won't pay much for it
Hard to build urgency around
🩹 Painkillers (must-have)
People urgently need it
They're already paying for workarounds
Price sensitivity is low
Word of mouth is natural

The validation question to ask at every stage: "Is this a vitamin or a painkiller?" If users say "I'd love that" rather than "I need that right now" — keep digging for the real pain.

What Real Validation Signals Look Like

Not all signals are equal. Here's how to tell the difference between real validation and false positives:

Strong signals ✓

These count as real validation

  • Someone pays you, even a small amount, before the product exists
  • Someone gives you their email and confirms they'll use it at launch
  • Multiple interviews surface the same frustration in the same words
  • Someone introduces you to another person who has the same problem
  • A user asks "when can I start using this?"
Weak signals ✗

These do not count as validation

  • "That sounds really interesting!" with no follow-through
  • Friends or family saying they'd use it
  • Social media likes or reposts
  • People agreeing the problem exists without confirming they'd pay
  • Any feedback from people who don't have the problem

Frequently Asked Questions

How do you validate a startup idea before building?
Two approaches work consistently: (1) The 5-step founder system — nail the problem, write a concept pitch, find 15–30 real target users, run pain interviews, and push for pre-sales or email signups. (2) AI-powered market research — use Claude to synthesise real user complaints from Reddit, G2, and Quora into a structured opportunity assessment.
What is validated learning in the lean startup?
Validated learning, coined by Eric Ries in The Lean Startup, is the practice of proving — with real data — that people want a solution, not just saying they do. It means collecting hard signals (pre-sales, signups, repeated pain patterns in interviews) rather than building based on assumptions.
How many user interviews do I need to validate my startup idea?
Aim for 15–30 conversationswith real target users — not friends, not followers. After 15–20 interviews, you'll start hearing the same frustrations repeated. That's your signal. If you can't find 15 people experiencing the problem, the market may be too small or the problem not painful enough.
What is the difference between a vitamin and a painkiller startup?
A vitamin is a nice-to-have — people appreciate it but don't urgently need it. A painkiller solves an acute, ongoing problem people are already spending money or effort to fix badly. Painkillers have natural word-of-mouth, lower price sensitivity, and are far easier to validate. The test: would your target user pay for any solution to this problem?
Can Claude be used to validate a startup idea?
Yes. By using Claude's research mode and prompting it to analyse real user content from Reddit, Quora, G2, and app reviews, founders can identify genuine pain points and market gaps in minutes. One founder used this exact method to validate a B2B cold email SaaS concept and reached $2.3k MRR within weeks of launching.
How long does startup idea validation take?
With focus, the full 5-step validation process takes 1–2 weeks. AI-powered research with Claude can produce an initial opportunity assessment in under an hour. The time investment before building is always worth it — it saves months of building the wrong thing.