The Complete Guide to Validating Market Demand

Market demand validation is one of those topics where everyone has an opinion and most opinions are wrong. The reason: validation isn't one technique, it's a taxonomy of techniques. Different ideas at different stages need different signals.

This guide breaks down every validation method we've seen produce a real signal, when each one applies, and what signal-to-noise ratio you should expect. Skip the ones that don't fit your situation. Master the ones that do.

The three categories of validation

Validation methods fall into three buckets. You typically run techniques from all three — sequentially, not in parallel.

• Qualitative validation: learning what users believe, want, and fear. High signal, low scale. Run early.
• Quantitative validation: measuring whether the demand exists at scale. Lower per-data-point signal but objective. Run middle.
• Pre-launch commitment validation: testing whether anyone will pay. Highest signal of all. Run last.

Part 1: Qualitative validation

Method 1: Customer discovery interviews

The original technique, still the most underused. 10–15 unstructured conversations with people in your target audience. Goal: hear them describe the problem in their own words, understand current workarounds, quantify the cost.

When it works: You have access to your target audience (you're part of it, or you can DM cold). The problem is something people talk about openly.

When it fails: You lead with your solution (poisoning the answer), you interview your friends (politeness bias), or your sample is too small (5 is not enough, 15 is the floor).

Method 2: Mom Test framing

Rob Fitzpatrick's book "The Mom Test" deserves its reputation. The principle: only ask questions even your mom couldn't fudge the answer to. Instead of "would you use this?", ask "walk me through the last time you ran into this problem." Past behavior is real data. Hypothetical opinions are noise.

Method 3: Concierge MVP

For ideas where the product would be a software workflow, manually do the workflow for 3 early customers using whatever tools you have (spreadsheets, email, Zapier). Charge them. See if they pay. See if they keep using it.

This is the highest-signal qualitative method because you're generating real revenue with a real customer doing the real workflow. The cost: it doesn't scale and it's exhausting. The reward: by the time you've done it for 3 customers, you know exactly what the v1 needs to automate.

Part 2: Quantitative validation

Method 4: Landing page conversion test

Build a one-page site, drive 100+ targeted visitors, measure signup rate. The benchmarks:

• <1% conversion = weak signal, rework the headline
• 1–3% = decent, you have something
• 3–7% = strong, especially with a price anchor
• 7%+ = very strong, often too good to be true (audit your traffic source)

The trap: conversion rate depends heavily on traffic source. A 5% conversion from highly targeted Twitter visitors is much less impressive than 5% from broad Reddit traffic. Track source attribution.

Method 5: Keyword volume + competition analysis

Open Ubersuggest or Ahrefs (free tiers work). Search for your target keywords. Look at:

• Monthly search volume — is there latent demand at scale?
• Keyword difficulty — how competitive is SEO ranking?
• Top-ranked results — do the existing pages look weak (good wedge) or strong (hard)?
• Suggested questions — what specific subqueries does the keyword have?

Healthy SaaS categories show 1,000–10,000 monthly searches on the head term and a long tail of related questions. Categories with 100 monthly searches on the head term aren't big enough yet to support a venture-scale business — though they can still support a bootstrapped one.

Method 6: Paid ad smoke test

Spend $100 on targeted Meta or Reddit ads pointing at your landing page. Look at click-through rate and signup conversion. This is a real demand test because you're paying for traffic — if the unit economics don't work at $100, they won't work at $10K.

Two numbers matter: cost per click ($0.50–$2 is healthy depending on niche) and cost per signup ($5–$20 is healthy for early-stage SaaS at consumer/SMB price points).

Method 7: Job posting analysis

Search LinkedIn or Indeed for job postings that mention the problem you're solving. Companies hire to solve problems. If 500 companies posted jobs for "data engineer to manage X workflow" in the last 90 days, that's 500 companies who might have bought your product instead.

This method works especially well for B2B SaaS in operational categories — sales ops, data, analytics, finance. It works less well for individual or creator-facing tools.

Method 8: Adjacent product growth as proxy

If your idea fits inside a larger ecosystem (Shopify apps, Notion add-ons, AI tools), look at the growth of the ecosystem itself. A 50% YoY growing host platform creates a constantly-expanding addressable market for your add-on.

Look at: app store rankings of adjacent apps, growth in funded companies tagged with the ecosystem, conference attendance, ecosystem-specific funds (Notion Capital, Shopify Ventures).

Part 3: Pre-launch commitment validation

Method 9: Waitlist with credible promise

Get email signups for a launch within 30 days. The credibility of the promise matters: "we'll send you an email someday" gets noise. "Get early access on June 15, $39/mo regular, $19/mo for waitlist" gets signal.

Healthy waitlists convert to actual signups at 20–40% on launch day. If your waitlist converted 5%, the demand was soft.

Method 10: Pre-orders with real money

The single strongest validation signal in startups. People who paid $1 are 100x more committed than people who gave you their email. Stripe Payment Links + a landing page is all you need to set this up — no product required.

Even 5 pre-orders at $1 is strong signal. Even 5 pre-orders at $99 with a 30-day money-back guarantee is "quit your job" signal. ShipFastvalidated his entire product before writing the code, with strangers paying $200+ for access.

Method 11: Letter of intent for B2B

In B2B, you can't always take credit cards from procurement upfront — but you can get a signed Letter of Intent saying "if you build this for these requirements at this price, we'll buy it." A signed LOI from 3 enterprise customers is worth $50K in seed funding.

Method 12: Manual concierge revenue (highest signal)

Already covered in qualitative as a discovery technique, but it's also the strongest possible validation. By the time you have 3 paying customers receiving manual delivery, you're no longer validating — you're a startup with a service, ready to productize.

The validation funnel

Run methods in this sequence. Don't skip steps unless you've already done them elsewhere.

1. Week 1: Methods 1, 2 — interviews and Mom Test framing.
2. Week 2: Methods 4, 5 — landing page + keyword research.
3. Week 3: Method 6 — paid smoke test if the landing page conversion was decent.
4. Week 4: Methods 9, 10 — waitlist + pre-orders. Decision point.

If after four weeks you have: 15 interviews showing clear pain + 3–5% landing page conversion + 5+ pre-orders or paid waitlist signups, you have validated demand. Start building.

If after four weeks you have less than that, the demand isn't there — at least not for this framing. Pivot the framing or pick a different idea.

Three signals to ignore

1. "Everyone I know loves this idea"

People love ideas. They don't pay for ideas. Compliments from your network are not data.

2. Sign-ups that don't engage

Easy email signups from broad sources (cold traffic, viral posts) are heavily diluted by people who'll never use the product. The metric to watch is reply-to-email rate or post-signup behavior, not raw signup count.

3. Competitor activity alone

"Someone else is building this so the market is real" is partially true and partially a trap. Markets where 10 well-funded competitors are fighting are validated but already crowded. Markets where 1 well-funded competitor exists are often the right timing. Markets where zero competitors exist are either too early, too small, or misframed — assume the worst case until proven otherwise.