How to validate a SaaS idea in 2026: low-cost playbook

Most SaaS ideas die the same way. You spend three months building, ship it, and get crickets. Not because the idea was bad — but because you never checked whether anyone would pay for it before you wrote the first line of code.

This is the “I built it, nobody came” failure mode. It’s embarrassing, expensive, and entirely preventable. Flexport ran an experimental marketing page — no working product yet — to test whether shippers would sign up before writing a single line of logistics software. Vanta knew they had product-market fit when customers started emailing them through word of mouth — no ads, no launch, a barebones homepage with a contact address. Neither of them needed a working product first.

This playbook runs in a weekend to a few weeks. The full tool stack costs between $0 and $31/year. Here’s what to do in order.

Who this is for

Indie hackers and early-stage founders with a dev background who want a battle-tested, low-cost validation process. If you’ve already launched and have paying users, this isn’t your article. If you’re still deciding whether to build, read on.

SaaS validation toolkit

Free tiers of everything unless noted. Full stack:

Tool	Purpose	Free tier	Paid
Perplexity	Desk research	Free (limited Pro searches)	Pro $20/month
Cal.com	Interview scheduling	Unlimited bookings	Teams $12/user/month (yearly)
Tally	Pre-launch survey	Unlimited forms + responses	$24/month
Carrd	Smoke-test landing page	Free (carrd.co subdomain)	Pro Standard $19/year
Lemon Squeezy	Pre-order / payment-intent	$0/month	5% + $0.50/transaction
Claude / GPT	Interview analysis	Free tier	Pro $20/month

Total for this playbook: $0. Transaction fees on Lemon Squeezy only fire when someone actually pays you — which is the entire point.

Step 1: Mine 1-star reviews (30 minutes)

Before you open a browser tab with your idea’s landing page draft, spend 30 minutes reading angry customers.

Go to the App Store, G2, Capterra, and Reddit. Search for the incumbent tool in your category. Filter for 1-star and 2-star reviews. Read 50 of them. You’re looking for recurring pain that the current solution doesn’t solve — not feature requests, but frustrations people can’t work around.

For example: if you’re building a time-tracking tool, you might find that every third 1-star review of Toggl mentions “I forget to start the timer.” That’s a problem. Your value proposition isn’t “a better timer” — it’s “automatic time tracking from your calendar, zero manual entry.”

Use Perplexity for the desk research layer. Run searches like:

"[competitor] alternatives" site:reddit.com
"[problem] tool" reviews 2025 OR 2026

Perplexity cites sources in real time and surfaces recent community sentiment faster than a Google dive. Three or four targeted searches will expose patterns.

Outcome: a 3-sentence value proposition. Not a feature list — one specific problem, one specific person, one measurable relief.

Time: 30 minutes.

Step 2: Run 5 problem interviews (2–3 days)

Before writing code, talk to five people who have the problem you think you’re solving. Not to pitch. To understand.

Cal.com’s free tier gives you unlimited scheduling links and automations. Post a short Tally form in the relevant subreddit, Slack group, or Discord. Ask: “Are you struggling with [problem]? I’m doing research — 20 minutes on Zoom, no pitch.” Cal.com handles the booking.

The five interviews you want are not “people who might use your product.” They’re people who have actively tried to solve the problem and failed. Paul Graham’s test: if someone has tried three different tools and is still frustrated, that’s the person who will pay you. If they’ve learned to live with the problem, you probably don’t have a hair-on-fire opportunity.

Run each interview with three questions:

Tell me about the last time [problem] cost you money or time.
What have you tried? Why didn’t it stick?
If this problem were solved tomorrow, what would change for you?

Don’t ask “would you use/pay for X?” That question is useless — people are polite. You’re not pitching. You’re listening.

After the five interviews, paste all transcripts into Claude. 200K context handles a full interview set in one prompt. Ask it to identify: the three most common frustrations, the three most common workarounds, and the exact language people used when describing the problem. Use their words verbatim in your landing page copy — not your words.

Outcome: validated problem, real customer language, signal on whether the pain is a hair-on-fire problem or a chronic-but-tolerable one.

Time: 2–3 days to schedule and run. 1 hour to analyze with Claude.

Step 3: Smoke-test a landing page (2 hours)

A smoke test is a landing page for a product that doesn’t exist yet. The point is to measure intent — do people click “get early access” or do they bounce?

Build it on Carrd. The free tier covers a single-page layout on a carrd.co subdomain. If you want a custom domain, Carrd Pro Standard is $19/year — not $19/month, $19/year. There is no reason to spend more at this stage.

Your landing page needs exactly four things:

Headline: the problem in the customer’s words (from your interview analysis).
Sub-headline: the specific relief, in concrete terms (“Stop losing time tracking manually — automatically logged from your calendar, every day”).
One CTA: “Join the waitlist” or “Get early access.”
One proof point: a real quote from an interview if you have one, or a number from your 1-star research (“43% of [tool] reviews mention manual entry as the top complaint”).

No feature list. No pricing section. No fake testimonials. They add noise, and manufactured social proof is dishonest.

Drive 200–500 visitors through Reddit posts, niche Slack groups, or a focused Twitter/X thread. Don’t spend money on ads yet — you don’t have enough signal to know what you’d optimize for.

If your conversion rate on the CTA is under 5%, the message isn’t working. That’s not a verdict on the idea — it’s a verdict on how you described it. Change the headline. Run it again.

Outcome: a CTA conversion rate. Under 5% from a warm-ish audience means the message needs work.

Time: 2 hours to build. Traffic within 48 hours from community posts.

Step 4: Charge before it exists (1 day)

The only real validation signal is someone handing over money. Waitlist signups are interest. A payment is a commitment.

Set up a payment-intent waitlist with Lemon Squeezy. The free tier costs $0/month to start — you pay 5% + $0.50 per transaction, but only when a payment processes. Create a product, set a real price (not $1 — that tests nothing), and put the checkout link on your landing page.

Your offer should be one of three things:

Pre-order at a discount: “Lock in $49/year (regular price $99) — only available before launch.”
Founding member: “First 50 users get lifetime access for $X.”
Refundable deposit: “Reserve your spot — we’ll charge at launch, or refund in full if we don’t ship.”

The Indie Hackers playbook on this is direct: post in the relevant subreddit, describe what you’re building, say “I’ll charge $19 lifetime to the first 10 people who want it.” If 10 people pay, build. If nobody pays after three posts in different communities over two weeks, the problem isn’t a hair-on-fire problem for enough people to support a SaaS.

Flexport’s early signups from that experimental page were encouraging. But the signal that made them build was when shippers tried to pay. Vanta had customers finding them through word of mouth and emailing a barebones homepage — no ads, no launch — asking about the product. Money tells you something waitlist signups don’t.

Outcome: pre-order revenue, or a zero-revenue signal that tells you to pivot.

Time: 1 day to set up. Results within two weeks of promoting.

Step 5: Iterate message, not product

Most founders read a flat landing page conversion rate as “my idea is bad.” It usually means “my description is bad.”

The idea is the problem you’re solving. The message is how you describe the solution. These are different things, and they fail independently.

When your conversion rate is low:

Change the headline first. Swap it for one of the exact phrases from your interview transcripts — the ones Claude flagged as most common. A/B test two versions over 48 hours.
Change the target community. A React developer audience and a bootstrapped founder audience will respond to the same product differently. Don’t assume one community represents all potential buyers.
Change the price point. $9/month attracts a different signal than $49/month. If $49 converts and $9 doesn’t, you have something worth building at a real price. If neither converts after four iterations, go back to Step 2.

What you should not change in this phase: the core problem you’re solving. If five interviews confirmed the problem is real and at least one payment confirmed intent, the product direction is validated. Don’t pivot to a different problem because the first week’s message didn’t land.

Iterate on message for two weeks maximum. If you can’t reach 5%+ conversion from a relevant audience after four message variations, run five more interviews. You may have solved the wrong layer of the problem.

Outcome: a message with a conversion rate you can extrapolate from.

Time: 1–2 weeks.

Step 6: Decision gate — build, pivot, or kill

By the end of this process, you have three data points:

Problem validated by interviews (qualitative).
Landing page CTA conversion rate (quantitative).
Pre-order revenue (financial).

Make a binary decision.

Build if:

3+ interviews confirmed the problem is hair-on-fire.
CTA converts at 5%+ from a relevant audience.
At least 3 people paid at a price above $19.

Pivot if:

Interviews confirmed the problem but nobody paid — the pain is real but not urgent enough to pay for a solution today. Find the more acute version.
Money moved but interviews uncovered a different, sharper pain than the one you built the landing page for.
Conversion is 5%+ but only at a price point that won’t cover your costs.

Kill if:

Fewer than 2 interviews confirmed real, recurring pain.
CTA conversion stays under 2% after 4 message iterations from warm audiences.
Zero payments after a real pricing page was live for two weeks.

Killing an idea is not failure. Spending three months building something nobody wants is failure. These six steps take a weekend to a few weeks. That’s the full cost of knowing before you build.

TL;DR

Step	What	Tool	Time	Signal
1	Mine 1-star reviews	Perplexity, G2, Reddit	30 min	Recurring pain patterns
2	5 problem interviews	Cal.com + Claude	2–3 days	Real pain, real language
3	Smoke-test landing page	Carrd	2 hours	CTA conversion rate
4	Charge before you build	Lemon Squeezy	1 day	Pre-order revenue
5	Iterate message	Carrd (manual A/B)	1–2 weeks	Conversion improvement
6	Decision gate	—	30 min	Build / pivot / kill

Full stack cost: $0 (free tiers) to $31/year (Carrd Pro Standard + custom domain).

Once validation is done and you’re ready to build, see our best full-stack deploy platform roundup for where to host your app, and Stripe vs Paddle for picking your payment processor.