6. Finding Startup Ideas

Most startup ideas do not arrive as lightning. They appear as repeated irritation, waste, delay, distrust, manual work, or a change in behavior that incumbents have not adjusted to yet. A founder’s job is not to sit in a room and invent clever ideas. The job is to notice where the world is already pulling, then test whether that pull is strong enough to build a company around.

The core idea-generation question is: where is a specific group of people already spending time, money, trust, or reputation to work around a problem that should not exist anymore?

A good idea is not just “interesting.” It has a painful problem, a reachable customer, a current workaround, a reason now is a better time than before, and some reason you can learn faster than other people.

Ideas Are Observed Before They Are Invented

Founders often talk about ideas as if they are nouns: “a CRM for schools,” “AI for lawyers,” “a marketplace for local services.” That is too thin. Useful startup ideas begin as observed situations.

Thin idea:

AI for Indian SMB finance.

Observed situation:

Multi-location cloud kitchen owners spend Sunday nights reconciling Swiggy, Zomato, UPI, refunds, commissions, cash expenses, and GST-ready records. They do it in spreadsheets, miss mismatches, and cannot see outlet-level leakage until month end.

The second version has a customer, workflow, trigger, cost, and current workaround. It can be tested. The first version mostly invites opinions.

When you write ideas, write the situation around the idea. Who is hurting? When does it happen? What do they do today? What does it cost? Who pays? What changed that makes now interesting?

The Four Best Sources Of Ideas

1. Pain You Understand Deeply

The easiest place to begin is a problem you have lived with. This could come from your job, your family business, your customers, your community, or a workflow you have personally suffered through.

Personal pain is useful because you already know the language, context, and hidden friction. But it is dangerous if you assume everyone else feels it as strongly as you do. Your first test is simple: find ten other people with the same pain and ask how they solve it today.

Good signs:

People already spend money or time on a workaround.
The pain repeats often.
The current solution is messy, expensive, slow, or unreliable.
Users can describe the problem without you explaining it.
The person affected can introduce you to others with the same issue.

Bad signs:

People agree politely but do nothing.
The problem is annoying but not urgent.
The buyer and user are different, and neither feels responsible.
You mainly want to build it because it sounds cool.
The problem disappears when you leave your own unusual context.

Personal pain gives you a starting hypothesis, not a verdict.

2. Repeated Customer Complaints

Customers are constantly telling the market where opportunities exist. They complain in support tickets, app reviews, Reddit threads, WhatsApp groups, LinkedIn posts, G2/Capterra reviews, sales calls, job descriptions, and informal operator conversations.

Do not look only for “I wish X existed.” Look for repeated evidence of struggle:

“We still do this in Excel.”
“We hired a person just to manage this.”
“This tool is too expensive for us.”
“The software works, but nobody in our team uses it.”
“We lose customers because this step is slow.”
“We cannot trust the data.”
“Compliance is always a last-minute panic.”

The best ideas often hide inside boring complaints. Boring is good. Boring usually means budget, frequency, operational pain, and less fashion-driven competition.

3. Market Or Technology Shifts

Some ideas become possible because something changed. UPI changed payment behavior. Smartphones changed distribution. AI changes the cost of support, research, writing, coding, analysis, and workflow automation. GST, Account Aggregator, DigiLocker, ONDC, FASTag, Aadhaar, India Stack APIs, and other infrastructure changes create new rails, new data availability, and new expectations.

But a shift is not an idea by itself. “AI for finance” is not an idea. “AI assistant that helps Indian SMB owners reconcile payments, invoices, collections, WhatsApp orders, and GST-ready records” is closer to an idea.

Use this filter:

What changed?
Who feels the change?
What old workflow is now obviously inefficient?
Who has budget or authority to fix it?
Why would the customer act now?

Technology shifts create opportunity when they collide with a painful workflow. They create noise when founders chase the technology without a customer.

4. India-Specific Friction

Indian founders often copy ideas from the US too quickly. Sometimes that works. Often, the better opportunity is to notice what is uniquely difficult here.

Look for friction in:

Trust: customers do not know whom to believe.
Payments: collections, refunds, reconciliation, credit, cash flow.
Compliance: GST, invoicing, payroll, vendor records, sector rules.
Language: support, onboarding, learning, discovery in regional languages.
Distribution: fragmented channels, offline relationships, reseller networks.
SMB operations: WhatsApp, Excel, phone calls, informal processes.
Talent gaps: training, hiring, quality control, productivity.
Access: healthcare, education, finance, legal help, logistics.
Status and aspiration: products that help people signal progress, safety, quality, or professionalism.

India rewards founders who understand messy reality. The opportunity is often not “build a better app.” It is “make a broken workflow trustworthy, usable, and economically possible for a specific customer segment.”

Idea Sources Worth Mining

Use these as research sources, not as places to copy. The point is to find recurring pain, vocabulary, workflows, and complaints.

Source	What to look for
Your job	Repeated work, internal hacks, approvals, handoffs, spreadsheet rituals.
Customer calls	Unsolved objections, manual work after purchase, gaps between promise and use.
Family business	Cash flow issues, vendor chaos, local distribution, compliance, collections.
Your city	Local services, logistics, real estate, education, healthcare, transport, civic friction.
Failed startups	Was the idea bad, timing early, distribution wrong, or business model weak?
App store reviews	Missing features, trust issues, bugs, support pain, price objections.
G2/Capterra reviews	What users hate about existing B2B tools.
Job descriptions	Work companies are hiring people to do manually.
Government tenders	Categories where institutions already budget for outcomes.
VC portfolios	Crowded themes, but also underserved adjacent niches.
SaaS directories	Categories with existing spend and unhappy customers.
WhatsApp groups	Real informal behavior, especially in Indian SMB and professional communities.
LinkedIn posts	Operator pain, hiring needs, market shifts, unpopular complaints.
Agency services	Repeatable services that might become productized.

Keep asking: is this a loud complaint or a costly workflow?

AI Startup Ideas

AI creates many idea-shaped distractions. A useful AI startup usually does one of three things:

Reduces the cost of a workflow that already exists.
Improves the quality or speed of a decision.
Makes a previously expensive capability available to a new customer segment.

Promising AI idea areas include:

AI copilots for support, sales, finance, compliance, research, recruiting, legal workflows, and developer productivity.
Vertical AI for specific industries such as clinics, accounting firms, exporters, schools, logistics operators, or insurance teams.
AI workflow automation where the product owns the full job, not just the chat layer.
AI for Indian languages, voice interfaces, document extraction, multilingual support, and regional onboarding.
AI for small businesses that cannot hire specialized teams.

Red flags:

The product is just a thin wrapper over a model.
The customer does not have a workflow where AI output creates measurable value.
Accuracy requirements are high but your data advantage is weak.
You cannot explain why the customer would keep using it after the novelty fades.
The distribution problem is harder than the AI problem, and you have no distribution edge.

The better question is not “What can AI do?” It is “Which expensive workflow becomes dramatically cheaper or faster because AI exists?”

Build An Idea Log

Do not trust memory. Create a simple idea log.

Use this format:

Field	What to write
Customer	Who exactly has this problem?
Trigger	When does the problem happen?
Current workaround	What do they do today?
Cost of pain	Time, money, risk, stress, lost revenue, churn.
Buyer	Who can approve or pay?
Frequency	Daily, weekly, monthly, rare.
Evidence	What did you see or hear?
Your advantage	Why can you learn faster here?
Next person to call	Name or segment, not “users.”

If you cannot fill this table, you do not yet have an idea. You have a topic.

Ten Useful Idea Prompts

Use these prompts when your mind is blank:

What do people still do in Excel even though the stakes are high?
Which workflows require WhatsApp follow-ups because software has failed?
Where do customers pay agencies because no product is good enough?
Which compliance task creates recurring panic?
What do founders, doctors, teachers, accountants, exporters, recruiters, or shop owners complain about privately?
Which job descriptions reveal manual work that software could reduce?
Which new regulation, platform, or infrastructure change creates confusion?
Which expensive tool is hated by a smaller customer segment?
Which offline trust network could be made more transparent?
Which repeated service could become a product if the workflow were standardized?

The answer to one prompt is not enough. Use it to generate conversations.

Common Mistakes

Starting with a technology instead of a customer pain.
Mistaking your own curiosity for market demand.
Copying a US startup without checking Indian buying behavior.
Building for “SMBs” without choosing a narrow segment.
Ignoring who pays.
Confusing a problem people discuss with a problem people act on.
Falling in love with the first idea because it feels productive.
Choosing a fashionable market where you have no access.
Writing feature lists before understanding current workarounds.

A Simple Weekly Idea Practice

Keep a running idea log.
Add five observations every week from real conversations, reviews, work, or market behavior.
Pick one idea and speak to five people who might have the problem.
Ask what they do today, how often it happens, what it costs, and why they have not solved it.
Kill, park, or continue the idea based on evidence.
At the end of the month, review which customer segments produced the strongest repeated pain.

This ritual makes idea generation less mystical. It turns it into observation, conversation, and pattern recognition.

The 30-Day Idea Mining Sprint

If you do not yet have a strong idea, do not spend a month brainstorming in private. Spend a month collecting field evidence.

Use this sprint:

Week	Work	Output
1	Observe your own work, family business, old customers, and professional network.	25 raw pain observations.
2	Mine reviews, job posts, communities, WhatsApp groups, and service businesses.	25 more observations with source links or notes.
3	Speak to 10 people across 2-3 promising customer segments.	Stories of recent problems and current workarounds.
4	Choose 3 ideas, define the first wedge for each, and run a small validation test.	One idea to continue, one to park, one to kill.

The sprint is deliberately evidence-heavy. A founder who writes 50 observations will usually see patterns that a founder with five polished ideas will miss.

Look For Paid Pain, Not Cleverness

When reviewing your idea log, mark every observation with one of these labels:

Paid pain: the customer already spends money, hires people, pays agencies, buys tools, or loses revenue because of the problem.
Workflow pain: the customer spends repeated time, creates manual processes, checks things twice, or uses messy tools to avoid mistakes.
Trust pain: the customer does not know whom to trust, how to verify quality, or how to reduce risk.
Access pain: the customer wants an outcome but cannot reach the right supplier, expert, lender, buyer, doctor, teacher, or service.
Status pain: the customer wants to look professional, safe, credible, modern, compliant, or successful.

Paid pain is usually easiest to turn into revenue. Workflow pain is often easiest to observe. Trust pain can create strong companies in India because customers pay to reduce risk. Access pain can create marketplaces or managed services, but only if you can solve liquidity and quality. Status pain can be powerful in consumer and education markets, but it needs careful pricing and distribution.

Idea Anti-Patterns

Be suspicious when the idea begins with one of these shapes:

Anti-pattern	Why it is risky	Better question
”Uber for…”	It borrows a model before understanding local supply, demand, trust, and economics.	Which specific transaction is painful today?
”AI for everyone in X”	It starts with technology and market label, not workflow.	Which expensive task becomes cheaper or faster?
”A community for…”	Communities need identity, repeated participation, and value before scale.	What urgent job brings people back weekly?
”Marketplace for all local services”	Liquidity, quality, fraud, and operations explode quickly.	Which narrow service has urgent demand and constrained supply?
”App for SMBs”	SMBs are not one segment.	Which business type, workflow, owner, and trigger?
”Better version of big company”	Better features rarely beat incumbents without a wedge.	Which ignored segment is overpaying or underserved?

Anti-patterns are not forbidden. They are warnings to go one level deeper before building.

From Observation To Startup Idea

Convert every promising observation into this sentence:

[Customer] struggles with [specific problem] when [trigger], causing [cost], and today they solve it by [workaround]. Because [change or insight], we may be able to help them achieve [outcome].

Examples:

CA firms struggle to collect client invoices before GST filing deadlines, causing deadline panic and junior staff follow-up, and today they solve it through WhatsApp reminders and spreadsheets. Because clients are already used to digital document sharing, a workflow tool may reduce follow-up and improve filing readiness.
Small D2C brands struggle to answer repetitive order-status questions after campaigns, causing support overload and poor customer experience, and today they solve it by hiring temporary support agents. Because AI and commerce integrations can automate common responses, a narrow support assistant may reduce cost during spikes.
Coaching institutes struggle to follow up with parent enquiries during admission season, causing missed admissions and owner anxiety, and today they solve it through phone calls, registers, and staff memory. Because WhatsApp is already the communication channel, a lightweight lead-follow-up system may fit the workflow.

The sentence forces specificity. If it sounds vague, the idea is still vague.

Idea Mining Map

When founders say “I do not have ideas,” they often mean “I am not looking in enough places with enough structure.” Use an idea mining map and force yourself to collect raw observations before judging them.

Mine	What to look for	Example signals
Your work history	Repeated frustration, spreadsheet-heavy work, handoffs, approvals, reporting, compliance, support.	”Every month we manually reconcile…” “Only one person knows how…”
Customer conversations	Complaints, workarounds, budget owners, urgent deadlines, switching stories.	”We already pay an agency for this.”
Family business	Informal processes, owner dependence, staff training gaps, vendor trust, cash collection.	”The owner checks everything on WhatsApp.”
City or local market	Fragmented supply, poor discovery, quality uncertainty, local logistics, language gaps.	”Everyone asks a friend for a reliable provider.”
Job descriptions	Tasks companies hire for repeatedly, tools required, manual responsibilities, painful workflows.	”Must manage MIS reports, vendor follow-up, GST documents.”
Review sites	Complaints about existing software, missing features, poor support, high price, bad onboarding.	”Too complicated for small teams.”
Communities	Repeated questions, templates being shared, manual hacks, peer recommendations.	”Which tool do you use for…”
Regulation and government	New reporting, documentation, filings, licenses, audit requirements, subsidies, procurement.	”Deadline changed; everyone is confused.”
Technology shifts	New capability that changes cost, speed, quality, trust, or distribution.	”AI can now draft/check/summarize what used to take hours.”
Service businesses	Agencies doing repeatable work that could become a productized service or software.	”Every client needs the same onboarding checklist.”

Do not start by asking “is this a unicorn?” Start by asking, “What is the repeated pain, who feels it, and how do they solve it today?”

Idea Mining Interview Script

When you are mining for ideas, do not pitch. Your goal is to understand the customer’s world well enough that problems reveal themselves.

Use a 20-minute operator conversation:

Context:
I am trying to understand how [role/business type] handles [broad workflow]. I am not pitching anything. I am looking for real operational friction.

Recent work:
What did the last week/month look like for this workflow?

Breakdowns:
Where did things slow down, break, need follow-up, or require owner attention?

Workarounds:
What tools, spreadsheets, WhatsApp groups, people, agencies, or manual checks do you use today?

Cost:
What happens when this goes wrong? Time, money, customer complaints, compliance, cash flow, owner stress?

History:
Have you tried to fix this before? Why did the solution not stick?

Budget/ownership:
Who owns this problem? Who approves spend if it becomes painful enough?

Pattern:
Who else has this problem in a similar form?

After the conversation, write only observations first. Do not immediately turn every complaint into a product. A good idea usually appears after the same pain shows up across several people in the same segment.

What To Listen For

Strong idea-mining signals sound concrete:

“Every Friday we…”
“The owner still has to check…”
“Only one person knows how…”
“We pay someone just to…”
“When this fails, customers…”
“Before the deadline, everyone…”
“We tried [tool/service], but…”

Weak signals sound abstract:

“This industry is inefficient.”
“People need better software.”
“AI can help here.”
“Everyone will want this.”
“There is no good app.”

Do not dismiss weak signals, but push them toward a recent story. If the person cannot describe the last occurrence, you probably do not yet understand the problem.

Source Signal Rubric

Not every idea source is equally strong. Score each source by the behavior it reveals.

Source signal	Strength	How to use it
One complaint in a forum	Low	Add to backlog; look for repeats.
Many similar complaints across sources	Medium	Define a segment and interview people.
Job descriptions hiring for the work	Medium	Inspect the workflow and owner.
Existing spend on agencies/tools/people	Strong	Map buyer, budget, and switching pain.
Customer shares files, screenshots, or workflow artifacts	Strong	Move from idea mining to validation.
Customer asks for help or introduction after the conversation	Very strong	Test a manual service, pilot, or concrete offer.

Use the rubric to avoid overreacting to internet noise. A loud thread can be useful, but a quiet workflow with existing spend is often more valuable.

Observation Quality

Raw ideas are cheap. High-quality observations are specific.

Weak observation:

SMBs need better software.

Better observation:

Small distributors reconcile sales, returns, UPI payments, cash, and WhatsApp orders manually every evening. The owner checks the numbers personally because staff mistakes affect cash and trust.

Strong observations include:

A specific customer type.
A repeated situation.
A current workaround.
A cost or risk.
A buyer or owner.
A trigger.
A reason existing solutions are insufficient.

Every week, collect 10 raw observations and rewrite the best three into this stronger format.

Idea Backlog Hygiene

Keep an idea backlog with four statuses:

Status	Meaning	Action
Raw	Interesting observation, not yet understood.	Add source and context.
Promising	Repeated pattern with possible pain.	Find 5 people to interview.
Testing	Clear customer, problem, and riskiest assumption.	Run a validation test.
Parked	Interesting but weak access, timing, buyer, or founder fit.	Revisit monthly.

Founders often keep everything in their head. That makes ideas feel more developed than they are. A written backlog creates honesty: either an idea has evidence, or it does not.

India-Specific Idea Filters

For Indian founders, ask these additional questions:

Is the problem hidden inside informal work, not visible in software categories?
Does trust matter more than convenience?
Is the buyer different from the user because of family, owner, finance, or hierarchy?
Does WhatsApp already coordinate the workflow?
Does UPI, GST, Aadhaar, DigiLocker, ONDC, Account Aggregator, or another public/digital rail change behavior?
Are language, city, income, or channel differences creating underserved segments?
Is the first product actually a service layer that earns trust before software?

India rewards founders who notice operational reality, not only internet trends.

The Founder Observation Loop

The best founders do not wait for an idea. They build an observation loop that keeps exposing them to market pain.

Use this weekly loop:

Day	Activity	Output
Monday	Read complaints, reviews, job posts, communities, and support threads.	10 raw observations.
Tuesday	Speak to 2-3 operators, buyers, or users.	Recent stories and current workarounds.
Wednesday	Inspect artifacts: spreadsheets, invoices, screenshots, forms, WhatsApp flows, dashboards.	Workflow evidence.
Thursday	Rewrite the best ideas into customer/problem/workaround format.	3 usable hypotheses.
Friday	Pick one hypothesis and ask who can be reached next week.	Interview list and outreach angle.

The loop matters because founders often mistake thinking for learning. Thinking improves the idea in your head. Observation improves your contact with reality.

In India, artifact inspection is especially useful. A customer may describe a process politely, but the real process may live in Excel sheets, Tally exports, WhatsApp forwards, paper registers, Google Forms, call recordings, UPI screenshots, and late-night founder follow-ups. Ask to see the workflow. The mess is often the market map.

From Complaint To Startup Hypothesis

Do not write “people complain about X” and stop. Convert every repeated complaint into a testable hypothesis.

Use this format:

We believe [specific customer segment]
struggles with [specific repeated situation]
because [current workaround or constraint]
which causes [measurable cost, risk, delay, or lost opportunity].
They currently solve it by [existing workaround],
and we can test this by [next conversation, workflow inspection, or small offer].

Example:

We believe small Indian exporters doing 20-100 shipments per month
struggle with invoice, documentation, and payment follow-up coordination
because the work is split between founders, accountants, logistics partners, and overseas buyers.
This causes delayed payments, compliance anxiety, and avoidable follow-up work.
They currently solve it through WhatsApp, Excel, email threads, and individual memory.
We can test this by interviewing 10 exporters and asking to inspect the last shipment workflow.

This is still not a validated idea. But it is now specific enough to test.

The Hidden Spend Test

Good ideas often have hidden spend before they have software spend. Look for places where customers already pay with:

Founder time.
Staff time.
Agency retainers.
Consultants.
Manual operations teams.
Error correction.
Delayed collections.
Compliance penalties or panic.
Lost customers.
Extra inventory.
Refunds, credits, or support load.

If customers say “we do not spend on this,” ask what happens when the problem is not solved. Someone may be paying indirectly through delays, mistakes, stress, churn, or management attention.

Hidden spend is important because it tells you the problem may already have economic weight. The first product does not need to capture all of it. It only needs to remove enough pain for a narrow customer to act.

Idea Mining Interview Questions

When you are still finding ideas, avoid asking “what startup should I build?” Ask questions that reveal repeated pain.

Useful questions:

What part of your work still feels unnecessarily manual?
Which task do you delay because it is painful?
Which spreadsheet, WhatsApp group, or manual tracker is most important to your business?
What breaks when volume doubles?
What do you pay an agency, consultant, junior employee, or freelancer to handle?
What do you check personally because mistakes are costly?
Which tool do people in your team avoid using?
What causes customer complaints even though everyone knows it should be fixed?
What compliance, reporting, finance, hiring, support, or sales task creates recurring stress?
What did you recently try to improve but give up on?

Listen for recent stories. A founder who collects stories will find better ideas than a founder who collects opinions.

Opportunity Map

After two or three weeks of idea mining, create an opportunity map instead of choosing immediately.

Opportunity	Customer	Current workaround	Hidden spend	Reachability	Why now	Next test

Fill 10-20 rows. Then look for clusters:

Several opportunities in the same customer segment.
Several problems caused by the same workflow.
Several complaints connected to the same buyer.
Several cases where the current workaround is expensive and visible.
Several ideas where you already have access to users or buyers.

The cluster is often more important than the single idea. It shows where your learning advantage may compound.

The Founder-Advantage Filter

Some ideas are good ideas for someone, but not good ideas for you. A founder should judge an idea through both market pain and personal advantage.

Use this filter before committing months of life to an idea:

Question	Strong signal	Weak signal
Can you reach the customer?	You can name 30 buyers or users and get conversations without heroic effort.	You admire the market from far away.
Do you understand the workflow?	You know the tools, language, incentives, and failure points.	You mainly know the buzzwords.
Can you earn trust?	Customers will share artifacts, data, intros, or real constraints with you.	Conversations stay polite and generic.
Can you learn faster than competitors?	You have domain access, distribution, technical depth, speed, or a personal network.	Your only advantage is enthusiasm.
Can you survive the sales cycle?	The business model matches your runway and temperament.	The buyer may take 9-12 months and you have 4 months of cash.
Would you still care after the novelty fades?	You can imagine doing customer calls in this market for years.	You mainly like the idea because the market is fashionable.

This filter is not meant to make founders timid. It is meant to make them honest. Startups are already hard. Choosing a market where you have no access, no trust, no stamina, and no learning advantage makes the game unnecessarily brutal.

The best early idea is often not the largest possible market. It is the painful, reachable wedge where the founder can learn faster than almost anyone else.

The 50-Observation Field Notebook

If you are still searching for an idea, give yourself an input target instead of a creativity target. Do not say, “I need one great idea.” Say, “I need 50 real observations from people doing painful work.”

An observation is not a startup idea. It is a small piece of reality:

A founder manually reconciles payments every Friday.
A school administrator keeps three different attendance records because parents, teachers, and regulators all need different views.
A sales manager spends two hours before every review cleaning CRM data because the team does not update it.
A clinic receptionist uses WhatsApp, paper, and one old desktop system to coordinate appointments, reports, and collections.
An exporter calls five people to know whether one shipment document is ready.

Collect observations in this structure:

Field	What to capture
Person	Role, company type, city, segment, and seniority.
Moment	When the pain appears: daily close, month end, audit, renewal, launch, hiring, shipment, exam season.
Workflow	The actual steps they take today.
Tool stack	WhatsApp, Excel, Tally, Zoho, Google Sheets, paper, custom software, people, agencies.
Failure	What goes wrong if the work is delayed, wrong, or ignored.
Cost	Time, money, lost revenue, risk, reputation, customer anger, founder attention.
Owner	Who gets blamed or interrupted.
Workaround	What they already do to survive.
Quote	The customer’s exact words.
Next question	What you still do not understand.

Fifty observations may produce only five serious idea candidates. That is normal. The point is to train your mind to see patterns instead of falling in love with the first clever thought.

Pattern Extraction

After 50 observations, group them by:

Customer segment: who repeats the pain?
Trigger: when does the pain become urgent?
Workflow: which step keeps breaking?
Workaround: what do people already pay for with money, time, or headcount?
Buyer: who owns the budget or consequence?
Trust barrier: why have existing tools not solved it?
Channel: where do these people already gather or buy?

The strongest idea is rarely the loudest single complaint. It is usually the pattern where the same pain, same owner, same trigger, and same workaround appear across multiple customers.

The “Show Me” Rule

When someone describes a painful workflow, ask politely if they can show you the artifact:

Spreadsheet.
WhatsApp thread.
Invoice.
Dashboard.
Support ticket.
Manual checklist.
Job description.
Email chain.
Report format.
Customer complaint.

You are not asking for confidential data. You are trying to understand the shape of the work. Artifacts reveal reality that interviews hide. They show what columns matter, where errors happen, which people are involved, and how messy the workflow really is.

If a customer will not show anything, that is not automatically bad. But if ten customers can only speak in abstractions and nobody can show a current workaround, the idea is probably still too vague.

Idea Source Quality Board

Not every source of ideas deserves the same trust. A founder can waste months chasing ideas from places that produce excitement but not buying behavior.

Use this board to rank your idea sources before you rank your ideas:

Source	What it is good for	What to watch for	Quality test
Your own repeated work	Problems you understand from inside the workflow.	You may assume everyone feels the pain as strongly as you do.	Can you find 10 similar people with the same problem?
Customer conversations	Live detail about pain, context, budget, and urgency.	Customers may be polite, speculative, or inconsistent.	Do they describe recent behavior without prompting?
Services delivery	Real paid work, messy workflows, and repeated manual tasks.	Service revenue can hide a product that will not generalize.	Does the same task repeat across multiple clients?
Support tickets and complaints	Language customers use when the pain is hot.	Loud complaints may come from edge cases.	Is the complaint tied to lost money, delay, risk, or churn?
Sales calls	Buyer objections, procurement friction, and willingness to pay.	Prospects may ask for custom features to avoid saying no.	Does the buyer agree to a next step with cost or effort?
Communities and forums	Early pattern discovery and vocabulary.	Upvotes, comments, and rants are not demand.	Can you turn the pattern into direct customer calls?
Regulatory or platform shifts	New deadlines, costs, compliance work, and workflow changes.	The window may be crowded or temporary.	Can you identify the role that now owns the new burden?
Investor and media trends	Macro direction and vocabulary.	Trend heat often attracts shallow competition.	Is there a specific customer pain underneath the trend?

Prefer sources where the customer is close to the pain and the pain is close to action. A founder should trust a messy spreadsheet shown by a customer more than a polished market map made in isolation.

Source Scoring

Give each source a score from 1 to 5 on:

Recency: did the pain happen in the last 30 days?
Repetition: have you seen it more than three times?
Cost: does it waste money, time, revenue, trust, or compliance effort?
Owner: is there a person clearly responsible for fixing it?
Access: can you reach more people like this without heroic effort?
Workaround: are people already doing something imperfect to survive?
Budget path: is there a plausible way this becomes paid?

If an idea comes from a weak source, do not kill it immediately. Treat it as unproven. Your next job is to move it from weak source to strong source by finding real users, real artifacts, and real behavior.

Problem Frequency Map

A good idea is usually attached to a repeated moment. The more precisely you understand that moment, the easier it becomes to build, sell, price, and position.

Map the problem by frequency:

Frequency	What it usually means	Founder question
Daily	The pain may support habit, workflow software, automation, or operations tooling.	Who touches this every day and what would save them 30 minutes?
Weekly	The pain may support reporting, review, coordination, or recurring service.	What meeting, delivery, payment, or review makes this urgent?
Monthly	The pain may be tied to accounting, payroll, billing, compliance, or targets.	What deadline creates pressure and who gets blamed?
Quarterly or annual	The pain may need enterprise sales, advisory, compliance, or high-ticket packaging.	Is the value large enough to survive a slow buying cycle?
Event-triggered	The pain appears during hiring, fundraising, audits, launches, outages, exams, festivals, exports, renewals, or disputes.	Can you predict the trigger and reach the customer before it happens?

Then map the owner:

Owner type	What to learn
Founder or owner	Is this stealing attention from sales, hiring, cash, or customers?
Functional head	Does this affect targets, team output, or reputation?
Operator or admin	Is the pain frequent but budget controlled elsewhere?
Finance or compliance	Does accuracy, auditability, or deadline risk matter?
Sales or customer success	Does the problem block revenue, retention, or response time?
End user	Is usage strong enough to influence the buyer?

The best early wedge is often a frequent pain owned by a reachable person who already has an ugly workaround. That combination gives you learning speed. Without learning speed, even a good idea can feel dead because the founder cannot get enough truth fast enough.

Idea Triage Queue

Once ideas start appearing, do not treat all of them equally. Put every idea into a triage queue so the founder’s calendar follows evidence, not excitement.

Queue	What belongs here	Founder action
Observe	Interesting pattern, but weak evidence or vague customer.	Collect more examples without building.
Interview	Specific customer and painful situation appear repeatedly.	Book 5-10 conversations with qualified people.
Artifact review	Customers mention a current workaround.	Ask to see spreadsheet, workflow, ticket, invoice, report, or process.
Buyer test	User pain is visible but commercial path is unclear.	Speak to buyer, owner, approver, or budget holder.
Manual test	Pain, buyer, and current workaround are credible.	Offer a concierge/manual version before software.
Park	Real but poor timing, weak access, or low founder advantage.	Save notes and revisit only on a trigger.
Kill	No recent pain, no owner, no action, or no reachable segment.	Preserve learning and remove from active review.

Weekly triage rule:

No idea moves forward without a stronger signal than last week.

Stronger signals include a customer showing a workaround, introducing a stakeholder, discussing budget, sharing data, agreeing to a paid/manual test, or repeating the same pain as others in the same segment. Weaker signals include likes, encouragement, generic interest, and “let me know when it is ready.”

This queue helps founders avoid two common traps: killing ideas too early because they are not polished, and keeping weak ideas alive because they are emotionally pleasant.

Reader Action

Pick one idea you are currently excited about. Write one page with: customer, problem, current workaround, cost of pain, buyer, evidence, and why now. Then schedule five customer conversations before you write another feature list.

If you cannot name five people to speak with, your next task is not product. It is access.