Founder Resources

How to find a Startup Idea that Works

Akshay Singh Oct 23, 2026 15 min read
How to find a Startup Idea that Works

A note before you read this: This guide shares a method for finding and testing startup ideas, not a guarantee of success. The examples are illustrative, not real customer research, and the score in the worksheet is a thinking tool, not a prediction. If you ever collect money or pre-orders, be transparent about what you’re offering and follow the rules that apply to you.

Quick answer: You find a startup idea by observing real problems in places you already have access to, writing them down precisely, checking how people cope with them today, mapping the strongest one with the PAIN framework (Person, Action, Impact, Now), talking to the people affected, scoring the evidence out of 30, and running one small test before you build anything.

Most people don’t get stuck on starting a business. They get stuck on the first question: what should I build? Every idea they think of already seems to exist, so they wait for a flash of inspiration that rarely comes.

Here’s what I’ve learned covering early-stage founders: the wait is unnecessary. Strong businesses rarely begin with a brilliant invention. They begin when someone notices a problem people face repeatedly, studies how those people cope today, and finds where the existing options fall short. That’s a method, not a moment of genius, and you can run it in about a week.

This guide walks through that method step by step, with a free worksheet and a 7-day plan at the end. No code, and very little money, required.

What you’ll walk away with:

  • A list of roughly ten real frustrations worth investigating
  • One problem mapped clearly with the PAIN framework
  • Notes from at least five customer conversations
  • A comparable opportunity score out of 30
  • One small, specific test to run next

Jump to: Why “what should I build?” fails · The 7 steps · 7-day plan · Free worksheet · FAQ

Startup idea discovery processSeven steps: Observe, Record, Research, Interview, Map, Compare, Test.From problem to test in 7 steps 1. Observe — watch real places2. Record — write 10 frustrations3. Research — see what exists4. Interview — ask about last time5. Map — fill in a PAIN Map6. Compare — score your top 37. Test — run one small experimentvrishbhanuja.com
The seven steps covered in this guide. Original diagram by vrishbhanuja.

Why does “what business should I start?” lead nowhere?

Because it skips straight to a solution before you understand a problem. Sit down and brainstorm business ideas cold, and you’ll usually end up copying a trend or falling for a solution nobody asked for.

Paul Graham, co-founder of the startup accelerator Y Combinator, makes a similar point in his essay How to Get Startup Ideas: the way to find ideas isn’t to try to think of them, but to actively look for problems, ideally ones you have yourself. He also warns that one of the most common startup mistakes is solving a problem nobody actually has.

It helps to keep four things apart, because people often use them interchangeably:

  • An idea is just a thought. “An app for hostel laundry” is an idea, nothing more.
  • A problem is something real people feel. “Students wait 40 minutes for a free washing machine” is a problem.
  • An opportunity is a problem that happens often, causes real pain, and that people already try to solve on their own.
  • A tested assumption is when you’ve confirmed, through real conversations or real behaviour, that part of this is actually true.

Finding a problem doesn’t prove a business exists inside it. It only tells you where to look next, which is exactly what the seven steps below are for.


What are the 7 steps to finding a startup idea?

The seven steps are: pick three places to observe, turn frustrations into specific problems, research what already exists, map the problem with PAIN, talk to real customers, score the opportunity, and test before you build. Each one is covered in detail below.

Step 1: Pick three places you can observe closely

Choose three environments where you can watch people run into problems up close. Your college or hostel. A local shop street. Freelancers, job seekers or a small office you have access to. A family business. Common paperwork or admin tasks you deal with yourself.

These are starting points to look, not proven markets. What matters is access: you can walk in, watch, and ask questions without cold-emailing a stranger. For a first-time founder with limited capital, that access is a genuine advantage over larger companies working from the outside.

Step 2: Turn frustrations into specific problems

Spend a few days noting what goes wrong around you. Aim for roughly ten frustrations, and look specifically for:

  • Tasks that repeatedly take too long
  • Complaints you hear more than once
  • Expensive or clumsy workarounds
  • Manual work people wish was easier
  • Mistakes, delays or missed opportunities
  • Problems people already spend time or money managing

Write the problem, not a product. “An app for X” is a solution in disguise; it tells you nothing about who’s affected or why it matters.

Worked example (illustrative, not real research). A vague observation like “students are always asking for notes” is too thin to act on. A specific version looks like this: “Before internal exams, second-year students in my hostel spend about an hour in WhatsApp groups asking for last year’s question papers, and often still don’t find the right one.” The second version names who is affected, when it happens, and exactly what goes wrong, which means you can go and check it.

Step 3: Research what already exists

Before you commit to a problem, find out how people cope with it today. This takes an hour or two and costs nothing:

  • Search engines and autocomplete. Type the problem in and see what people actually ask around it.
  • Forums and public discussions. Read the complaints people post. Never collect anyone’s private information.
  • Reviews of existing products. Look for the same shortcoming mentioned repeatedly.
  • Competitor websites and pricing pages. Note what they charge and which customers they seem to overlook.
  • Google Trends. Shows whether interest in a search term is rising, flat or falling over time.

Look specifically for repeated complaints, underserved groups, confusing workflows and clear gaps. Respect each site’s terms of use.

A popular search or an existing competitor is not proof of demand. Both are clues worth noting, not evidence that people will pay. A crowded market can even be encouraging, since it shows people already want something, as long as you can identify exactly what the existing options miss.

Step 4: Map the problem with PAIN

The PAIN framework turns a vague frustration into a specific, checkable problem. It has four parts:

  • P — Person. Who experiences this, specifically?
  • A — Action. What are they actually trying to accomplish?
  • I — Impact. What does it cost them in time, money or stress?
  • N — Now. What do they do today to manage or work around it?
PAIN Map with an examplePerson, Action, Impact and Now, filled in with an illustrative example about students finding old question papers.The PAIN Map (example, not real research) P — PersonA — ActionI — ImpactN — Now Second-year B.Tech studentsliving in one hostelFind last year’s question papersbefore internal exams1-2 hours lost searching;often cannot find the right paperAsk in WhatsApp groups andseniors; share photos of papers vrishbhanuja.com
A completed PAIN Map example. Illustrative only, not real customer research. Original diagram by vrishbhanuja.

If you can’t fill in the “Now” box with any confidence, that’s a signal you don’t yet understand the problem well enough, and it’s worth returning to Step 2 or Step 3 before going further. Once you’re comfortable with PAIN, Strategyzer’s Value Proposition Canvas is a useful next tool for connecting a customer’s needs to what you plan to offer.

Step 5: Talk to real customers

Interview five to ten people who genuinely experience the problem. Ask about the past, not the future; questions that work well include:

  • Tell me about the last time this happened.
  • What were you trying to accomplish?
  • What did you do about it?
  • Which part was hardest?
  • How often does this happen, roughly?
  • How much time, money or effort does it cost you?
  • Have you tried other solutions? What happened?
  • Have you ever paid for something, or changed how you work, because of this?

Avoid “Would you use my app?” and “Do you like my idea?” People tend to answer those kindly rather than honestly, and a guess about future behaviour is weak evidence on its own. What someone actually did is far more reliable. Rob Fitzpatrick’s book The Mom Test is a useful reference on this exact problem.

Reading an interview without over-interpreting it (illustrative example). Five students tell you, “Sure, I’d use that.” One tells you, “Last week I spent an hour hunting for a paper, so I paid a senior ₹50 to send me photos of it.” The first four answers are polite interest. The fifth describes real behaviour. Don’t draw a conclusion from a single story either way; note it, and look for the same pattern across several more conversations.

Step 6: Score the opportunity

Compare your strongest problems with a simple, transparent score. Rate each of these six factors from 1 (weak evidence) to 5 (strong evidence):

  1. How often does the problem happen?
  2. How serious are the consequences?
  3. Do people already use workarounds or pay to manage it?
  4. Can you reach the people affected?
  5. Have your conversations shown real demand, not just polite interest?
  6. Can you test this with the time and resources you currently have?
Score (out of 30)What it suggests
6–15Evidence is early or thin. Keep learning before committing further; this is not a rejection.
16–23A mixed picture. Identify which factors are still guesses, and test those specifically.
24–30Strong evidence on paper. Still confirm with real people and a small test before building.

Treat this as a comparison and reflection tool, not a prediction. A high score doesn’t guarantee success, and a low one doesn’t mean you should quit. Its real value is surfacing which assumptions you’ve actually tested and which you’ve only guessed at.

Step 7: Test before you build

Identify the single assumption that, if wrong, would sink the idea, then pick the cheapest test that could prove it wrong:

  • More interviews tell you whether the problem is common, though words are not the same as behaviour.
  • Watching the existing workflow shows you what really happens, step by step, though usually with a small sample.
  • A simple landing page tests whether your offer interests the right people, though a click isn’t a payment.
  • Doing the service by hand (“concierge”) shows whether people use and value the result, though it’s hard to scale.
  • A clickable mock-up checks whether the flow is clear, though it doesn’t prove demand on its own.
  • A small pilot with a narrow group shows whether it works for them specifically, though it may not generalize.
  • A willingness-to-pay test is the strongest signal available, provided you’re transparent about what’s being offered.

Likes, page views and enthusiastic survey answers don’t prove a sustainable business. If you collect money or pre-orders at this stage, state clearly what the person is buying and when it will arrive, then honor that commitment or refund it. Before you run the test, write down the specific result that would change your mind, so you aren’t tempted to read a “yes” into ambiguous evidence afterward.


What does a 7-day startup idea discovery plan look like?

The seven steps above map directly onto a seven-day plan, with each day taking roughly 30 to 60 minutes. It’s also exactly how the free worksheet below is structured, so you can track your progress as you go.

DayTask
1Choose three places or groups you can observe closely.
2Record about ten frustrations. Write problems, not products.
3Research existing alternatives: search, reviews, forums, competitors.
4Talk to at least three potential customers, more if you can manage it.
5Look for recurring patterns and map your strongest problems with PAIN.
6Shortlist up to three opportunities and score each one out of 30.
7Choose one small test and write down what evidence would change your mind.

What mistakes should you avoid when looking for a startup idea?

MistakeFix
Starting with a technology, not a problemStart with a specific person and what’s going wrong for them. Choose the tool later.
Assuming a trendy market guarantees demandFind a specific group that already struggles with the problem.
Copying a business without understanding its customersFind out who its customers actually are and what they still complain about.
Asking leading interview questionsAsk about the last time the problem actually happened.
Treating polite interest as proofLook for time spent, money paid or workarounds people have already built.
Building too much before testing assumptionsRun the smallest possible test first.
Defining the target customer too broadlyName one small group you can realistically reach this month.
Ignoring distribution and pricing until laterWrite down how you’d reach ten customers and what they might pay, early.

Your free startup idea discovery worksheet

The worksheet below follows the same seven steps covered in this guide. Tick off each step as you complete it, fill in your PAIN Map, score your opportunity and plan your test, all in one place. Your entries stay in your own browser; nothing is sent to vrishbhanuja.com. Use the download button if you’d like to keep a copy.

Interactive worksheet

Find your startup opportunity

Move from a vague idea to a problem worth investigating. Tick the steps, map the problem and choose one small test. You can return to this worksheet later.

Your discovery progress0 of 7 steps

1. The seven discovery steps

2. Map one problem with PAIN

Describe the customer and their current behaviour—not just the app you want to build.

3. Rate the opportunity

Score each factor from 1 (weak evidence) to 5 (strong evidence). This helps compare ideas; it does not predict success.

Score: 18 / 30Mixed picture (16–23). Find which scores are still guesses and test those first.

4. Choose your next experiment

Save to keep your entries in this browser. Your entries are not sent to Vrishbhanuja.

Your score is a reflection tool, not a guarantee of market demand. Validate assumptions with real customers.


What should you do once a promising problem emerges?

Finding a real problem is the beginning, not the end. From here, dig into six further questions: who exactly the customer is, whether the demand is genuine, how you’ll reach them, how you’ll earn money, what makes you different from existing options, and whether you can realistically deliver.

Ready to examine your idea in more detail? Go beyond this quick score. Check demand, customer clarity, distribution, monetization, differentiation and execution fit with vrishbhanuja’s Startup Idea Validator. It helps you look at your idea from several angles; it can’t promise demand, funding or profit.

Check my detailed idea ↗

Already confident you want to move forward? Read our guide on how to register a startup in India, or browse the founder tools hub.


Frequently asked questions about finding a startup idea

Q: What should I do if I have no startup idea?
A: Don’t try to invent one. Pick three places you already know well, write down ten recurring frustrations you notice there, and talk to the people who have them. Startup ideas usually come from problems you observe, not from brainstorming in isolation.

Q: Can I find a startup idea without technical skills?
A: Yes. Finding and validating a problem requires observation and conversation, not code. Most early tests, such as interviews, a landing page or doing the service by hand, need no technical skill at all.

Q: Do I need a completely original idea to start a business?
A: No. Many successful businesses improve on something that already exists rather than invent something new. A competitor’s presence can actually show that demand is real; your job is to find what they miss.

Q: How can a student find a startup idea with little money?
A: Use your campus, hostel and classes as a free research environment. Observing, interviewing classmates and running a small manual test cost little more than your time.

Q: How do I know if a problem is worth solving?
A: Check whether the problem happens often, causes real consequences, and whether people already use workarounds or pay to manage it. Then confirm you can actually reach the people who have it.

Q: Can I validate a startup idea without building an app?
A: Yes. Customer interviews, watching an existing workflow, a simple landing page, or doing the service manually for a few customers all generate evidence without writing a line of code.

Q: What should I do after I find a promising startup idea?
A: Write down your biggest untested assumption, run the smallest experiment that could disprove it, and then examine customer clarity, demand, distribution, monetization, differentiation and your ability to execute.


Start with Step 1 today

You don’t need a perfect idea this week. You need a real problem, honest evidence, and one small test to run next, which is a far more achievable goal and the one that actually leads somewhere.

Open the worksheet above and write down your three places to start. Once you have an idea you’re serious about, take it to the Startup Idea Validator.

Stuck on a PAIN Map, or unsure how to approach your first conversation? Drop a comment and tell me where you got stuck.

📂 Category: Founder Resources
🏷️ Tags: startup-ideas, customer-discovery, idea-validation, student-founders, india-startup

Related Articles

Read next

Start Building Today

Read smarter. Build faster. Use the right tools.

Explore Founder Tools
Scroll to Top