Validate a Business Idea With $0 in LATAM: 14-Day Plan

Table of Contents

• Summary: four methods, zero budget
• Why building before validating is expensive
• Step 1: interviews with The Mom Test

  • Three rules
  • How to rephrase a question
  • Message to ask for the interview
  • Where AI helps
• Step 2: the test landing page (smoke test)

  • What it needs
  • What to build it with, for free
  • Free analytics
• Step 3: concierge, or delivering by hand
• Step 4: presale or letter of intent
• How to get visits without paying for ads
• How to interpret the results

  • Painkiller or vitamin
  • Example thresholds
• 14-day plan
• Frequently asked questions

  • Can you validate a business idea without money?
  • How many interviews do I need?
  • How do I collect a presale if cards are rarely used in my country?
  • What conversion shows the idea works?
  • What if I collect a presale and the idea doesn't work?

Validating a business idea with $0 means testing your hypotheses with real people before paying for development, design or advertising. You don't need capital to do it. You need conversations, a page with a price, and someone who risks something, even a small transfer.

This plan is built for Latin America, where charging by international card is often the first obstacle. Here the buying test is done with Pix, Yape, Nequi or SPEI, which is how your customer actually pays.

If you want the shorter version focused on selling, it is in how to validate a business idea in one week. This post goes one step earlier and wider: it starts with interviews, uses AI to speed up the work, and ends with a written decision in 14 days.

Summary: four methods, zero budget

Method	Cost	Time	Evidence	What it tests
Mom Test interviews	$0	3 to 5 days	Qualitative: past facts and frustration	That the pain exists and is a priority
Landing page with smoke test	$0 on free plans	2 to 4 days	Quantitative: sign-ups, clicks, reservations	Interest and willingness to pay
Concierge MVP	$0	3 to 7 days	Operational: real effort to deliver	That you can deliver value, even by hand
Presale or letter of intent	$0	5 to 10 days	Money or written commitment	Demand at a price

The times are estimates for part-time work. What matters is the order: first you understand the problem, then you measure demand, and only then you ask for money.

Why building before validating is expensive

Most projects that never take off don't fail because of technology. They fail because they were built without knowing whether anyone would pay. Months of work behind closed doors, then a launch that meets silence. It is the same mistake we explain in Steve Blank's method.

The defense is to separate two kinds of signal:

• Verbal enthusiasm. "Great idea, let me know when it launches." It costs nothing and predicts nothing.
• Real commitment. The customer risks something: money, time, data they don't give to just anyone. A reservation, a presale, a transfer receipt or a signed letter.

In LATAM a local receipt is excellent proof. Paying through Pix, Yape, Nequi or SPEI means opening the app, typing a number and confirming. Nobody does that out of courtesy.

Step 1: interviews with The Mom Test

The Mom Test is Rob Fitzpatrick's book on how to ask questions without being lied to out of kindness. Its core idea: if your mom can give you a positive answer to your question, it is a bad question.

Three rules

1. Talk about their life, not your idea. Don't mention the project or ask for opinions on concepts.
2. Ask about the past. What they did, what they tried, what they paid, how much time they lost. What someone did is evidence. What they say they would do is noise.
3. Listen more than you talk. If you are explaining your solution, you stopped researching.

How to rephrase a question

Bad question	Why it fails	Good question
"Would you pay $15 a month for an app that organizes your finances?"	It makes you imagine a future where saying "yes" costs nothing	"How did you organize your finances last month? How much do you spend on tools today and what was most annoying?"
"Do you think it's a good idea?"	It asks for a compliment	"When was the last time this happened to you and what did you do?"
"What features would you like?"	It asks for design from someone who isn't a designer	"What solution did you try before and why did you drop it?"

Message to ask for the interview

It works for WhatsApp or LinkedIn. Adapt it and don't send it in bulk:

Hi [Name], I saw you lead [area] at [company] and your experience with [topic] caught my attention. I'm researching how [painful process] gets solved in [country or region]. I'm not selling anything: I want to learn from real cases. Would you have 10 minutes this week to tell me how you handle it today?

Where AI helps

ChatGPT or Claude are useful for three concrete things before and after the interviews:

• Reviewing your script. Ask it to flag questions that lead the answer or mention your solution.
• Rehearsing. You can simulate a customer to practice your tone, knowing it is a rehearsal and not evidence.
• Summarizing. With the transcript or your notes, look for patterns: which pain repeats, how much they spend, which alternatives they use.

What AI can't do is replace the real person. A synthetic customer can't tell you what they paid last month.

Step 2: the test landing page (smoke test)

A smoke test measures demand with a page that sells something that doesn't exist yet. With the interviews done, you already know who you are talking to and which pain they mentioned.

What it needs

• A headline with the result and the segment.
• Three concrete benefits.
• The price in plain sight. Without a price you measure curiosity, not demand.
• One button: "Reserve a spot" or "Buy the launch plan".
• A form for an email or for uploading the payment receipt.

What to build it with, for free

Any tool you can publish with today works: Carrd, Google Sites, a public Notion page, or an AI builder like Lovable or Emergent if you prefer to generate the page from a prompt. Check each free plan's limits before relying on it, because they change often.

A starter prompt:

Create a single-page landing for [value proposition] aimed at [segment] in [country]. Include a headline with the main benefit, three benefits in cards, a price block with a reserve button, and a frequently asked questions section. Do not invent testimonials or client logos.

The last line matters. A fake testimonial invalidates the test: you would be measuring how much they believe you, not how much they need what you offer.

Free analytics

A spreadsheet is enough: visits, traffic source, button clicks, emails and payments. If you want to see what people do on the page, Hotjar and similar tools offer free plans with heatmaps and recordings. Google Trends gives you an idea of search interest in your country. For email, Loops and other platforms have free plans that cover a small list. Confirm the current limits before choosing.

You don't need SEO or structured data on this page. Its job is to measure, not to rank.

Step 3: concierge, or delivering by hand

In a concierge MVP (or "Wizard of Oz") the customer receives the result and you produce it manually behind the scenes. If your idea is a recommendation app, you don't program the algorithm: you receive a form, analyze the answers and send the recommendation as a PDF over WhatsApp.

It does two things. It checks that the customer values the result, and it teaches you how much effort delivery takes. If five customers wear you out, you already know what to automate first.

Step 4: presale or letter of intent

This is the strongest test, because it asks for money or a written commitment.

Consumers (B2C). Offer a launch price with limited spots. Collect with the method your customer already uses:

• Pix in Brazil
• Yape or Plin in Peru
• Nequi or Daviplata in Colombia
• SPEI in Mexico
• A Mercado Pago link where available

Ask them to upload the receipt in a form. That way you record every payment without a payment gateway in between.

Businesses (B2B). A letter of intent (LOI) signed by the decision maker: the company commits to buying at a defined price and conditions if you deliver what was agreed. It is not a contract, but it separates real interest from "we'll look at it next quarter". For real legal matters, see essential contracts for freelancers and startups in LATAM.

An ethical rule. If you charge, you deliver or you refund. Taking money and disappearing destroys your reputation and can bring legal trouble. If you aren't ready to deliver anything, don't charge yet: use a waitlist with a visible price.

Before publishing the price, check that it leaves you a margin with the profit margin calculator.

How to get visits without paying for ads

With no budget, traffic comes from your network and from communities where your customer already is:

1. LinkedIn. Post what you learned in the interviews and link the page. Groups in your industry work better than your personal profile.
2. Reddit. Take part in subreddits on your topic by helping first. A bare link gets deleted.
3. WhatsApp and Facebook. Local groups in your niche, respecting their rules.
4. Direct messages. The interviewees who showed pain are the first people to write to.

Watch out for contamination: visits from friends and family inflate your conversion. Note where each visit comes from and calculate your numbers using cold traffic only.

How to interpret the results

Define the metrics before launching, not after. If you write them with the data in front of you, you will move the goal until your idea passes.

Painkiller or vitamin

Criterion	Painkiller (urgent pain)	Vitamin (secondary want)
Willingness to pay	Already spends time or money trying to solve it	Likes the idea but sets no budget
Adoption	Actively looks for an alternative	Needs convincing
In the tests	Signs up, reserves and pays	Clicks out of curiosity and leaves at the price

Example thresholds

These numbers are a reference for setting your own, not a universal rule:

• Persevere: on cold traffic with a visible price, 10% to 15% leave their email or reserve, and at least 3 people pay a presale or sign a letter of intent.
• Pivot: the pain showed up in the interviews, but the landing page doesn't convert. Change one variable only: the segment, the offer or the price, and repeat.
• Drop: nobody pays and the interviews showed no spending or frustration. It is a vitamin. Keep what you learned and move to another hypothesis.

If you landed close to the threshold, you can give it 72 more hours under the same rules. If you landed far, don't negotiate.

14-day plan

Week 1: understand the problem

• Days 1 and 2. Write your hypotheses in the Business Model Canvas, define who feels the pain and set your thresholds. Ask an AI to review your interview script.
• Days 3 to 7. Run 10 Mom Test interviews. Note spending, alternatives and frustration. Summarize the patterns.

Week 2: measure demand

• Days 8 to 10. Publish the landing page with a price, the reservation form and the local payment instructions.
• Days 11 to 13. Distribute through your free channels, write to the interviewees who showed pain and log everything in the spreadsheet.
• Day 14. Compare against your thresholds and write your decision in ten lines: persevere, pivot or drop.

If you decide to persevere, the next step is building a minimal first version. To structure it, see the agile business plan with template. And if your idea is software, the guide to a one-person Micro-SaaS with AI covers the full path.

Frequently asked questions

Can you validate a business idea without money?

Yes. Interviews, a free landing page, a form and a service delivered by hand cost nothing but your time. What must cost your customer something is the commitment: a payment, a reservation or a signed letter. Without that you are only measuring politeness.

How many interviews do I need?

Ten with people from the right segment are usually enough to see patterns. If after ten nobody tells you they already spend time or money on the problem, you don't need twenty more: you need another segment or another problem.

How do I collect a presale if cards are rarely used in my country?

With the local method your customer already has on their phone: Pix, Yape or Plin, Nequi or Daviplata, SPEI. Ask for the receipt in a form. It is more friction than a card button and, for that very reason, a stronger buying signal.

What conversion shows the idea works?

There is no universal number. Set it before launching based on your price and traffic. As a reference, on cold traffic that saw the price, 10% to 15% sign-ups with a few payments is a strong signal. A sign-up without a visible price is worth much less.

What if I collect a presale and the idea doesn't work?

You deliver what you promised by hand or refund the money. A presale is a commercial promise. If you can't fulfil it manually for ten customers, don't charge yet.

---

If you would rather do this with guidance, the Transforma Program works through validation and the first customer step by step. But the method works just as well on your own: write your thresholds today, before you have a single data point.