Quick Answer (TL;DR)
No users does not mean no research. It means you need different research methods. The most effective pre-launch discovery techniques are competitor user interviews (talk to people who use similar products), fake door tests (measure demand before building), concierge MVPs (deliver the value manually), landing page experiments (test messaging with real traffic), and community mining (find unmet needs in public conversations). All of these give you real behavioral signal without requiring a single existing customer.
The principles of product discovery stay the same at every stage. The methods just adapt to what you have available.
The Zero-User Problem
Most discovery advice assumes you have users to talk to. "Interview your customers." "Watch users interact with your product." "Analyze usage data." Helpful if you are at a company with millions of users. Useless if you are pre-launch, entering a new market, or building something that does not exist yet.
The zero-user problem is not actually a discovery problem. It is a sourcing problem. The research methods work fine. You just need to find the right people to research and the right signals to measure.
Here are seven methods, ordered from highest signal to lowest effort.
Method 1: Competitor User Interviews
Talk to people who use products adjacent to what you are building. They have the problem you want to solve. They just solve it differently today.
How to find them:
- Search G2, Capterra, or Product Hunt reviews for competing products. Reviewers often include their name and company. Reach out on LinkedIn.
- Post in relevant subreddits or Slack communities: "I am researching how teams handle [problem]. Would anyone be open to a 20-minute call?"
- Ask your network for introductions to people in your target role at your target company size.
What to ask:
- "Walk me through how you currently handle [problem]." (Open-ended, behavioral)
- "What is the most frustrating part of your current workflow?" (Pain discovery)
- "Have you tried other tools for this? What made you switch or stay?" (Switching costs)
- "If this problem magically disappeared tomorrow, what would change for you?" (Value sizing)
What NOT to ask: "Would you use a product that does X?" This is a hypothetical question and the answer is meaningless. Everyone says yes to hypothetical products.
Eight to twelve interviews will reveal patterns. If three people describe the same workaround for the same frustration, you have a discovery worth pursuing.
Method 2: Fake Door Tests
A fake door test puts a button, link, or signup form for a feature that does not exist yet in front of real people. You measure how many click. The click is the signal, not a survey response.
Examples:
- A landing page with a "Request Early Access" button. The page describes your product. The button collects an email. You measure the conversion rate from visitor to email signup.
- A feature button inside an existing product that leads to a "Coming Soon" message. You measure the click rate compared to existing features.
- A Google Ad for your product concept. You measure click-through rate and cost per signup.
Why it works: People vote with their clicks, not their words. A 5% conversion rate on a landing page with cold traffic is stronger validation than 50 people telling you "yeah, I would probably use that."
How to set one up cheaply:
- Write a landing page that describes the value proposition in one sentence
- Add a clear call to action ("Get Early Access" or "Join the Waitlist")
- Drive 500-1,000 visitors via targeted ads ($200-500 budget)
- Measure the conversion rate
A conversion rate above 3% from cold paid traffic suggests real demand. Below 1% means your positioning is off or the problem is not painful enough. Iterate on the messaging before concluding the idea is bad. Sometimes the product is right but the framing is wrong.
Method 3: Concierge MVP
A concierge MVP delivers the value your product would provide, but you do it manually. No code. No automation. Just you, doing the work for the customer. This is one form of minimum viable experiment that validates demand before writing production code.
The classic example: Food on the Table started by personally shopping for groceries for a single customer based on her preferences and local store sales. No app. No algorithm. Just a founder with a shopping list. This validated that people would pay for personalized meal planning before writing a line of code.
How to run one:
- Find 3-5 people willing to be your first customers (paid or free)
- Deliver your product's core value manually. Use spreadsheets, email, and personal effort.
- Observe what they actually use and what they ignore
- Note where they push back or ask for changes
- Track whether they come back for more
What you learn: A concierge MVP tests two things simultaneously. First, whether people want the outcome your product delivers. Second, what the actual workflow looks like in practice. The second insight is more valuable. You will discover steps, edge cases, and preferences that no amount of upfront planning would reveal.
This is the most effort-intensive method on this list, but it produces the highest-fidelity signal. If you can serve 5 customers manually, you understand the problem deeply enough to automate it. For more on MVP approaches, see our guide on what an MVP actually is and when each type is appropriate.
Method 4: Landing Page Experiments
Similar to fake door tests, but more structured. Run multiple landing pages with different value propositions, targeting the same audience, and see which one converts best.
What to test:
- Problem framing. Does "save 10 hours per week on reporting" convert better than "never miss a key metric again"? The winning frame tells you which pain point resonates most.
- Audience targeting. Does the same page convert better with product managers or engineering managers? The winning audience tells you who to build for first.
- Pricing signals. Does a page mentioning "$49/month" convert differently than one saying "free to start"? Price sensitivity data before you build saves painful pivots later.
Run each variation for at least 500 unique visitors before drawing conclusions. Use the results to inform both your product direction and your go-to-market messaging.
Method 5: Community Research
Your target users are already talking about their problems. You just need to find the conversations.
Where to look:
- Reddit. Search for your problem domain in relevant subreddits. Sort by top posts. Read the comments. People on Reddit are blunt about what works and what does not.
- Slack and Discord communities. Industry-specific communities (Lenny's Newsletter Slack, Mind the Product, etc.) have channels where people ask for tool recommendations and complain about existing solutions.
- Stack Overflow and GitHub Issues. For developer tools, these are goldmines. Issues labeled "feature request" tell you what users want. Issues labeled "help wanted" tell you where they struggle.
- Twitter/X. Search for complaints about competing products. "I wish [competitor] would..." is a product requirement in disguise.
- G2 and Capterra reviews. Three-star reviews are the most useful. Five-star reviews are vague praise. One-star reviews are often about unrelated issues (pricing, support). Three-star reviews say "this is good but it is missing X." That X is your opportunity.
How to synthesize: Create a spreadsheet with columns for source, pain point, current workaround, and frequency. After reading 50-100 posts, patterns emerge. The pain points mentioned most often, with the worst current workarounds, are your strongest candidates.
This is passive research. It tells you what problems exist but not how badly people want them solved. Combine community research with one of the active methods (interviews, fake doors) to validate the intensity.
Method 6: Proxy Metrics
When you cannot measure your own product's usage, measure the market's behavior instead.
Search volume. Use Google Trends or keyword research tools to see if people are searching for solutions to the problem you want to solve. Growing search volume for "[your problem] tool" or "how to [solve your problem]" signals rising demand.
Job postings. If companies are hiring for roles that manually do what your product would automate, the problem is real and the budget exists. A company paying $80K/year for a role your product could partially replace is a company with willingness to pay.
Funding signals. If competitors in your space are raising money, investors have validated that the market exists. You still need to validate that your specific approach works, but the market-level risk is lower.
Adjacent product growth. If tools that sit next to your product in the workflow are growing quickly, the ecosystem is expanding. A rising tide of adjacent products suggests your target users are investing in the problem space.
These are weaker signals than direct user research, but they are free and fast. Use proxy metrics to decide which ideas are worth deeper investigation, then validate with interviews or experiments.
Method 7: Founding Customer Programs
A founding customer program gives early adopters special access and influence in exchange for their time and feedback.
The offer:
- Free or heavily discounted access for 12 months
- Direct access to the product team (weekly calls, shared Slack channel)
- Influence over the roadmap (their use cases get prioritized)
- "Founding Customer" badge or recognition
What you get:
- Committed users who will actually use the product and give honest feedback
- Real usage data from day one
- Case studies and testimonials for your launch
- Word-of-mouth referrals to their peers
How to recruit founding customers: Start with the people from your competitor interviews and concierge MVP. They already know you and the problem. Offer them the founding customer deal. Aim for 5-10 founding customers, not 50. You want deep relationships, not a crowd.
Founding customer programs bridge the gap between discovery and launch. They give you real users before you have a product in market. Use their input to shape your product roadmap and prioritize features using a framework like RICE based on actual customer feedback rather than assumptions.
Combining Methods for Confidence
No single method gives you enough confidence to build. Combine them.
High-confidence stack:
- Community research to identify candidate problems (1-2 days)
- Competitor user interviews to validate pain intensity (2-3 weeks)
- Fake door test to measure demand with real behavior (1-2 weeks)
- Concierge MVP with 3-5 customers to learn the workflow (2-4 weeks)
If all four methods point in the same direction, you have strong validation. If they contradict each other, dig into why. The contradiction itself is a discovery.
Low-budget stack:
- Community research (free, 2 days)
- Landing page experiment with $300 in ads (1 week)
- Five interviews sourced from landing page signups (1 week)
Total cost: $300 and three weeks. That is a small price for knowing whether your idea has legs before spending months building it.
The absence of users is not the absence of signal. It just means you need to look in different places. Every method in this guide produces real behavioral data. None of them require asking people to imagine a product that does not exist. Start with the method that fits your timeline and budget, then layer on additional methods as your conviction grows.
Explore More
- Discovery for New Product Managers - Learn product discovery fundamentals as a new PM.
- What Is a Concierge MVP? - Expert answer on concierge MVPs for product validation.
- Top 10 Customer Feedback Tools and Methods (2026) - 10 tools and methods for collecting, organizing, and acting on customer feedback.
- Top 10 User Research Methods for PMs (2026) - 10 user research methods every PM should know.