Good startup ideas are often discovered rather than invented. They appear when you notice a painful repeated problem, understand who has it, and see that existing alternatives are expensive, awkward or incomplete.
Good ideas are more often found than invented
Instead of starting with “I want to build an app”, start by collecting situations where people lose time, money, attention or control. A project is stronger when the technology follows the problem.
Why is the problem usually more valuable than the idea?
An idea can change many times. A strong problem gives you a stable direction. If customers urgently need an outcome, there may be several possible solutions. If the problem is weak, even a clever solution struggles.
How to recognise a valuable problem
It is painful
The consequences are meaningful enough that people care.
It repeats
Recurring problems create recurring value.
People already try to solve it
Spreadsheets, assistants, manual work, competitor products or ugly workarounds prove that effort already exists.
You know who has it
A defined segment is easier to interview, reach and serve.
The audience is reachable
A great problem in an inaccessible market can still be a bad startup opportunity.
There is budget
The person benefiting may not be the person paying. Understand how money moves.
The timing is right
Technology, regulation, behaviour or cost changes can make an old problem newly solvable.
Where to find startup ideas
1. Keep a problem journal
Write down annoyances, repeated manual work and moments when you think “why is this still done this way?”
2. Observe work processes
B2B opportunities often hide in handoffs, spreadsheets, duplicate entry and approval chains.
3. Look for workarounds
A complex workaround is evidence that the problem matters.
4. Read complaints, not only trend reports
Reviews, forums, support threads and professional groups contain direct language from users.
5. Watch changes
New regulation, AI capabilities, devices, costs or demographics create new constraints and opportunities.
6. Find niches ignored by broad products
A horizontal tool may serve everyone adequately and one segment poorly.
7. Start with a service
Manual service work reveals repeated tasks that can later become software.
8. Combine two worlds you understand
Unusual founder advantages often come from overlapping domains.
Practical brainstorming techniques
Ten problems a day
Quantity trains observation. Do not judge each item immediately.
“Why is this still done this way?”
Question old processes that survive through inertia.
“Ten times better”
Imagine a solution dramatically faster, cheaper, easier or more reliable.
Jobs to be done
Focus on what the person is trying to accomplish, not the current product category.
Opposite assumption
Take a standard industry rule and ask what happens if the opposite is true.
Build for one person - but choose the right person
A specific early user helps you make decisions. The danger is choosing someone whose needs are unique. Look for a person who represents a broader repeated pattern.
Market research: what to examine
1. Customer
Who experiences the problem and who buys?
2. Existing alternative
What happens today?
3. Intensity and frequency
How painful and how common is the problem?
4. Market size from the bottom up
Estimate number of realistic customers multiplied by plausible annual value rather than relying only on huge top-down industry numbers.
5. Market dynamics
Is the segment growing? What is changing?
6. Competition and category
Competitors reveal demand and customer expectations.
7. Reachability
Where can you actually contact customers?
How to evaluate several ideas
Score them across problem intensity, frequency, willingness to pay, founder advantage, reachability, competition, timing, technical feasibility and potential scale. Do not treat the score as truth; use it to expose assumptions.
How do you know whether an idea is good enough to build?
You do not know with certainty. You look for enough evidence that building the next small test is justified.
What is idea validation?
Validation is the process of replacing assumptions with evidence. It is not collecting compliments. The strongest evidence comes from behaviour under realistic conditions.
A ladder of validation evidence
Opinion < email signup < scheduled call < data provided < pilot commitment < pre-order < payment < repeated paid use. The exact ladder varies, but commitment matters.
How to talk to potential customers
Ask about the past
“When did this last happen?” is better than “would you use this?”
Do not reveal the solution too early
Otherwise people begin trying to please you.
Look for behaviour and detail
Tools used, time spent, money lost, internal workarounds and frequency matter.
Ask for the next step
A pilot, introduction, pre-order or access to real data creates stronger evidence.
Prototype, MVP and final project are not the same
Prototype
Demonstrates an interaction or concept.
MVP
Creates enough real value to test a business assumption.
Final project
A mature system built for reliability, scale and a broader set of needs.
Choose the MVP according to the largest risk
If demand is uncertain, test demand. If technology is uncertain, build a technical experiment. If willingness to pay is uncertain, ask for money. The MVP should attack the risk, not simply be a smaller feature list.
Where to find first users and early adopters
Start with people who feel the problem most strongly
Early adopters tolerate imperfection when the pain is high.
Specific places to look
Professional communities, niche subreddits, industry events, LinkedIn, existing clients, newsletters, local associations, marketplaces and direct outreach.
Do not only ask “do you want to test?”
Explain the problem and the expected commitment. Serious testers are more useful than a long list of curious signups.
Do things that do not scale
Personally onboard users, migrate data and observe them. The purpose is learning.
What is product-market fit?
Product-market fit is the point where a defined market strongly values the project and usage becomes sustainable. It is not one moment or one metric.
Product-market fit is not a universal percentage
Popular surveys such as asking how disappointed users would be if a product disappeared can be useful, but no percentage replaces retention, payment behaviour and market context.
Signs of product-market fit
Retention stabilises
A meaningful group keeps using the project.
Customers pay without extraordinary persuasion
The value is understood.
Organic referrals appear
Users bring others.
Users complain when the project is unavailable
It has become important enough to notice.
Customers ask for more
Expansion demand appears.
One segment stands out
Strong fit is often narrow before it is broad.
What is NOT product-market fit?
High traffic, many signups, a successful Product Hunt day, an investment round, one large customer, positive compliments, a founder manually rescuing every account, or rapid top-line growth with extreme churn are not sufficient proof.
Validation versus product-market fit
Validation says the idea deserves to be built further. Product-market fit says a specific market repeatedly values the working project.
What if validation fails?
When to continue
Continue when the problem is strong but the segment, positioning, channel or solution appears wrong.
When to stop or change direction
If nobody shows meaningful behaviour after repeated good tests, continuing can become attachment rather than evidence-based entrepreneurship.
A practical path from idea to PMF
- Write down 20–50 problems.
- Select the three strongest.
- Run at least ten problem conversations for each.
- Identify the riskiest assumption.
- Create the smallest credible test.
- Find 5–20 early adopters.
- Ask for action or payment.
- Build a narrow MVP.
- Measure activation and retention.
- Focus on the segment where signals are strongest.
A twenty-question idea test
Who has the problem? How often? What happens if they do nothing? What do they use now? What does that cost? Who pays? Can you reach them? Why now? Why you? Is the market growing? Is there competition? What do competitors miss? What is the riskiest assumption? Can it be tested without full development? What action proves demand? What is the smallest MVP? What would make users return? What can become defensible? What would make you stop? What evidence would justify investing more?
Problem interviews: what not to ask
Avoid questions such as “Do you think this is a good idea?”, “Would you pay €20 for this?” or “Would you use an app that...?” People are poor at predicting future behaviour and often try to be supportive.
Ask what they did last time, which tools they used, who was involved, how much time or money it cost and what finally happened.
Segment before you generalise
If ten people describe ten different problems, you may be looking at several markets. Group interviews by context and look for repeated patterns. A smaller coherent segment is more useful than a large vague one.
Competitor research
Read reviews, pricing pages, support forums, release notes and cancellation complaints. The goal is not to copy features. It is to understand which expectations are already established and where customers remain dissatisfied.
Demand tests
A landing page can test whether a message attracts the right people. A waitlist can test intent. A pre-order tests stronger commitment. A paid pilot tests both value and buying process. Choose the test that matches the current uncertainty.
Concierge MVP
Deliver the result manually behind a simple interface. If customers value the result, you can later automate the repeated parts. This is especially useful for workflow and AI projects.
Wizard-of-Oz MVP
The user experiences what looks like an automated service while some operations are performed manually behind the scenes. Use this ethically and do not misrepresent capabilities where trust or safety would be affected.
Pricing is part of validation
A free test can prove usage but not willingness to pay. Introduce real pricing before you assume the business model works. Even a small paid pilot creates stronger evidence than a large free beta.
Retention cohorts
Measure groups of users who started in the same period and track how many remain active. This prevents new acquisition from hiding the fact that older users are leaving.
Why product-market fit can disappear
Markets change, competitors improve, acquisition channels shift and customer expectations rise. PMF is not a permanent achievement. Mature companies continue measuring retention and relevance.
False positives in validation
Friends, investors, innovation programmes and startup communities may praise an idea without being customers. Media attention can also create temporary curiosity. Keep separate signals from people who actually experience the problem.
False negatives in validation
A good problem can appear weak if the test is badly designed. Wrong segment, confusing copy, poor prototype or inappropriate channel can all hide demand. Failure should tell you which assumption failed, not automatically that the entire market is impossible.
When to pivot
A pivot changes one major assumption while preserving useful learning. You might keep the technology but change the customer, keep the problem but change the solution, or keep the audience but change the business model. Randomly changing everything is not a pivot; it resets the experiment.
Early adopter characteristics
Strong early adopters usually feel the problem more intensely, already spend time or money on a workaround, can make a decision relatively quickly and are willing to tolerate an imperfect first version in exchange for a better outcome.
A simple PMF dashboard
Track activation, weekly or monthly retention by cohort, paid conversion, churn, expansion, referral rate and qualitative reasons for staying or leaving. Add segment breakdowns so a strong niche is not hidden by weak overall averages.
TAM, SAM and SOM
Top-down market size can be useful context, but early founders need a reachable market. TAM is the broad theoretical total, SAM is the segment your model can serve, and SOM is the portion you can realistically reach in the near term. Bottom-up calculations usually force more realistic assumptions.
Market timing
A strong idea can fail when technology, regulation, cost or behaviour is not ready. Ask what changed recently that makes the project possible now. “Why now?” is as important as “why this?”
Switching behaviour
Even if your solution is better, customers must leave something. Estimate the cost of migration, learning, lost data, internal approvals and risk. The project may need to be much better than the current alternative, not merely slightly better.
Pre-orders and letters of intent
These can create stronger evidence than surveys, but interpret them carefully. A non-binding letter is weaker than money transferred. The closer the commitment is to a real purchase, the stronger the signal.
Design partners
A small number of early customers can work closely with the team to shape the first version. Choose design partners that represent the target segment and avoid turning the project into custom software for one company.
Churn interviews
Users who leave can reveal why product-market fit is weak. Ask what they expected, what happened, what they returned to and what would have needed to be different.
Expansion as a PMF signal
When customers voluntarily add seats, usage or locations, it suggests the project is creating enough value to deserve a larger role. Expansion can be a stronger signal than new signups.
Word of mouth
Organic referrals indicate that users understand the value well enough to explain it to others. Track how new users heard about the project rather than assuming all direct traffic is truly direct.
Conclusion
Do not try to prove that your idea is brilliant. Try to discover whether a real problem, reachable customer and strong enough behaviour exist. Validation is not about being right; it is about becoming less wrong before the expensive part begins.