How to Validate B2B SaaS Ideas Before Writing Any Code

The number one reason SaaS startups fail is not poor code or server crashes.

It is building something nobody wants to pay for.

Too many technical founders spend six months building a complex application in isolation, only to launch on Product Hunt to the sound of complete silence.

Here is a step-by-step validation playbook to prove commercial demand before spending a single dollar on software development.

One framing point before the steps: validation is not a single event, a survey, or a launch. It is a sequence of increasingly expensive commitments — your interviewee's time, then their email, then their calendar, then their money. Each stage exists to answer one question, and skipping a stage is how founders end up with a beautiful product and no buyers.

The 4-Stage Idea Validation Framework

Phase 1: Problem Discovery (20 Customer Interviews)
               ↓
Phase 2: Smoke Test Landing Page (Message Testing)
               ↓
Phase 3: Clickable Prototype Walkthroughs
               ↓
Phase 4: Pre-Sales & Letter of Intent (LOI)

Budget 2 to 4 weeks for the whole sequence and expect to spend under $500 excluding your own time. Compared with the cost of building the wrong thing — six months of engineering, or $30,000–$60,000 in agency work — it is the cheapest insurance available to a software business.

1. Conduct 20 Problem Discovery Interviews

Do not ask people: "Would you buy software that does X?"

People are polite and will say "Yes" to be encouraging.

Instead, follow The Mom Test methodology and ask about past behavior: * "How do you currently handle [Problem X] in your company?" * "What is the most frustrating part of that process?" * "How much time and money did you spend trying to fix this last quarter?" * "What tools have you evaluated or built internally to solve it?"

If they haven't actively tried to solve the problem with spreadsheets or existing tools, it is not a painful enough problem.

How to Actually Get 20 Interviews

This is the part most guides skip, and it is the hardest part. Twenty interviews is achievable in two weeks with these tactics:

• Warm intros first. Ten conversations from your network beats a hundred cold messages. Ask specifically: "Do you know anyone who manages [job title] at a company with [size]?"
• LinkedIn outreach at volume. Expect a 5–15% reply rate. Send 20 messages to book one call, so 20 conversations means roughly 150–250 messages. Personalise the first line or do not bother.
• Industry communities. Slack groups, subreddits and trade associations for the niche are goldmines because members self-select for the problem.
• Offer a small incentive. A $25–$50 gift card lifts response rates materially for junior and mid-level staff. Executives are usually moved more by flattery and curiosity than by money.
• Ask for 20 minutes, not an hour. Shorter asks get accepted far more often, and 20 focused minutes is enough.

What to Record, and When to Stop

Record each call (with permission), then log every answer into a simple table: who they are, the problem in their words, what they currently do about it, what it costs them, and whether they asked you anything. That last column is quietly one of the best signals — people who probe for a solution are showing genuine interest.

Stop interviewing when three consecutive conversations surface the same problem, described in the same language. That convergence is the signal. If 20 interviews produce 20 different problems, you do not have a niche yet, and no amount of engineering will create one.

Kill Criteria

Decide in advance what would make you walk away, so the decision is not made emotionally after six weeks of work:

• Fewer than half of interviewees describe the problem as urgent
• Nobody has spent money or significant internal effort trying to solve it
• Solutions are already considered "good enough" and switching would cost more than the problem
• The people who feel the pain do not control the budget

Any two of those together is a strong signal to change the idea rather than to build it.

2. Launch a Smoke Test Landing Page

Build a sharp, 1-page landing page highlighting: * A clear headline addressing the specific pain point. * 3 core benefit pillars with mock screenshots. * A transparent pricing tier. * A Call-to-Action button: "Join Private Beta" or "Schedule Demo".

Run $200 in targeted LinkedIn or Google Search ads to your exact buyer personas in the US, UK, Canada, or Australia. A conversion rate of >10% on email capture indicates strong message resonance.

The Honest Maths of Paid Smoke Tests

Here is where most advice gets vague, so let us be precise. B2B LinkedIn clicks typically cost $8–$15 for a well-targeted campaign, and Google Search clicks for commercial B2B keywords often run higher. $200 therefore buys roughly 15–25 visitors — far too few to draw conclusions from a 10% conversion rate, because two signups and five signups both look like noise at that volume.

Two practical fixes:

1. Spend $500–$1,000 on traffic, or drive free traffic first. A post in the communities you already interviewed in will produce more qualified visitors than $200 of ads, and the people arriving are closer to your actual buyer.
2. Point the ads at search intent where it exists. If people are already searching "how to [solve problem]", search ads tell you the problem exists today. If nobody is searching, that is itself a finding.

Measure qualified demo requests, not email addresses. Anything under about 3% intent (booking a call or requesting a demo) on a decent sample means the message is not landing — and the fix is usually the headline, not the product.

One important rule: never lie on a smoke test page. Say what the product will do, and be clear about its availability. Fabricated claims destroy credibility with exactly the early adopters you need most.

Phase 3 in Detail: Clickable Prototype Walkthroughs

The framework above includes a prototype phase, and it earns its place. Between "the message resonates" and "here is my credit card" there is a real question: does the workflow match how I actually work?

Build the clickable prototype in Figma in two or three days and walk five of your interviewees through it, screen by screen, asking them to describe what they expect each button to do. Do not guide them. Watch where they pause.

What this stage reliably reveals:

• Missing steps in your workflow that only a practitioner would notice ("where do I record the courier's reference number?")
• Vocabulary mismatches — you say "projects", your buyers say "matters" or "jobs" or "cases", and using their word measurably improves conversion
• Integration requirements that would otherwise appear as objections after launch
• Which features are actually load-bearing, so your MVP scope shrinks instead of growing

Do this before writing code. A workflow change discovered in Figma costs a day; the same change after launch costs a sprint.

3. Secure 3 Letters of Intent (LOI) or Paid Pre-Orders

In B2B software, the ultimate validation is money or a formal commitment.

Offer early access at a 50% lifetime discount in exchange for weekly feedback during the beta. If a business decision-maker refuses to sign an LOI or pay a small deposit, re-examine your value proposition.

Be realistic about what an LOI is worth, though. A signed non-binding letter is a goodwill signal, not revenue. Companies sign them to be helpful and then evaporate in procurement. Rank your evidence honestly:

Push for the deposit. A $500 hold credit is a small ask that filters out the enthusiasm-only crowd instantly. Ten design partners who each pay a deposit is a far more valuable position than fifty LOIs, and it funds your first quarter of infrastructure.

When you do sign design partners, define expectations in writing: the discount in exchange for feedback commitment, the timeline, what happens to their data if you pivot, and what they get if the product takes longer than planned. Being explicit up front is what keeps early customers with you when the roadmap slips.

Common Validation Mistakes

• Interviewing friends and family. They are not your buyer and they are not capable of saying no.
• Asking about the future ("would you use...") rather than the past ("how did you handle...").
• Treating signups as validation. An email address is a click, not a commitment.
• Building a free MVP "to see if people use it." Free usage tells you almost nothing about willingness to pay in B2B.
• Ignoring who holds the budget. If your champion cannot approve spending, you have a person with a problem and no purchase path.
• Never defining a stopping rule. Validation without kill criteria is just confirmation-seeking with extra steps.

Summary

• Validate the pain, not the solution.
• Measure behavior, not opinions.
• Secure commitments before building code.

Related reading: MVP vs prototype is the decision immediately after validation; the discovery phase is the paid version of this work; and which MVP features to build first picks up where this article ends.

What happens immediately after launch: Onboarding UX for B2B SaaS. Once your concept is validated, partner with KEHEM IT to build a lean, high-performance SaaS MVP in 6 to 8 weeks.

Have an idea and a hunch it is right? Tell us who pays for it and why and we will tell you what to test before building.