How to Validate a Business Idea Before You Build
How to validate a business idea before you build: check complaints, paid demand, search volume and competitors per function, then test with a kill criterion.
If you’re working out how to validate a business idea, the short version is this: before you build, find independent evidence for four things. People have the problem, they already pay to get it solved, they search for a fix, and the products on sale today don’t serve them well.
The method: stop judging the idea as a whole. Write it down as the separate jobs the product would do and check each one. A pitch can read well while only one of its parts has proof, and validating a business idea this way shows which part to build first and which to cut. If you have no candidate yet, start with where business ideas come from.
This guide is for indie developers and micro SaaS builders choosing what to build for an app marketplace or as a Chrome extension. It covers the evidence checks and cheap experiments that come before code; building the MVP, running ads and writing a business plan are out of scope.
Key takeaways
- Grade each function from E0 (not observed) to E4 (people pay for a product).
- Split the idea into functions and check each one. In our public sample report, one function out of four had direct evidence.
- Money already spent, such as freelance gigs with buyer reviews, is stronger proof than anyone saying they would pay.
- Search volume shows interest in a task, not willingness to pay for your version.
- Write the kill criterion for every experiment before you run it.
Write the idea as functions, then check each one
Business idea validation is the work of collecting outside evidence that a product’s jobs are wanted and paid for, before you write code. It goes wrong at the first step when the pitch is treated as one thing. If you ask how to validate my business idea and answer with a survey about the whole pitch, people are polite.
A function is one sentence with a verb describing one job of the product: “compares stock levels after a failed sync” is a function. “Inventory peace of mind” isn’t one, because nobody can search for it or pay a freelancer to do it.
Then grade what you can observe for each function. Our validator uses five evidence levels, and they work just as well in a spreadsheet:
| Level | What was observed | Where you see it |
|---|---|---|
| E0 | Not observed | Nothing found, or nobody looked |
| E1 | People said it | Complaints in reviews, forums, Reddit |
| E2 | People search for it | Monthly searches for task phrases |
| E3 | People pay someone to do it | Freelance gigs with buyer reviews |
| E4 | People pay for a product | Paid apps with real customers |
A level records what was observed, never how convincing the idea sounds; E0 counts neither for nor against. Build first the function with the highest level, test E1 and E2 functions with interviews and a commitment test (only E3 and E4 mean money has moved), and park E0 functions until a customer raises them unprompted. The checks behind the levels are described on how it works.
A worked example with a Needs proof verdict
Our sample validation report runs a real idea against the Shopify App Store and its 254 pain clusters (groups of complaints about the same problem). The idea is a tool for merchants running several warehouses, submitted as four functions:
| Function | Level | Complaints | Paid demand | Search |
|---|---|---|---|---|
| Compares stock levels after a failed sync | E2 | 1 pain, 2 distinct people | Searched, nothing sold | 190/mo |
| Shows each discrepancy with its money value | E0 | No matching pain | Searched, nothing sold | 0/mo |
| Holds orders for out-of-sync SKUs | E0 | No matching pain | Catalogue section not crawled | 0/mo |
| Emails the bookkeeper a daily list | E0 | No matching pain | Catalogue section not crawled | 0/mo |
Only the first function had direct evidence, from two people. One wrote: “overselling on amazon when shopify already sold the last unit is a nightmare to fix” (source). The other: “Having to email a customer to apologize and cancel their order because the stock didn’t actually exist completely ruins brand trust.” (source) Its 190 monthly searches came from three phrases, led by shopify inventory sync at 90.
The verdict was Needs proof, with medium confidence and partial evidence. Even the supported function is a near match: the pain behind it is multichannel overselling, so the report warns that the multi-warehouse wedge may be narrower than the evidence.
Where to look for complaints, and how many people count
Complaints are E1 evidence, and the useful ones are unprompted: app marketplace reviews, vendor community forums, Reddit threads and one-star browser extension reviews. That’s also how we build pain cards: every day we read those sources, drop praise, off-topic chatter and support noise, and group posts about the same underlying frustration into one card with the quotes linked. As of November 2026 the index holds 7,668 pain cards across 78 marketplaces.
Count people, not posts. One person posting in five threads is one data point. A strong pattern looks like card 4136 on the Shopify App Store: 304 posts from 256 people, dropshippers whose Meta ad campaigns burn budget with no sales, with explicit willingness to pay. One of them wrote: “At these CPM numbers I can’t see this ever being profitable.” (source)
A smaller pattern can still be clean. Card 3981, Shopify merchants hit by Google Merchant misrepresentation suspensions, has 19 posts from 19 people, so every post is a different merchant. One wrote in an app review: “Same misrepresentation issue as another user, but no clear answer as to what specific rule my specific store was failing at.” (source)
The worked example rested on two people, which makes it a lead. Look for several unrelated people in more than one source describing the same failure in their own words. Check the date of the newest post too, since an old complaint may already be fixed. And check that each complaint is about your function: the sample report rejected a currency display complaint because it described a different problem.
Paid demand is proof of willingness to pay
Saying “I would pay for that” costs nothing, so a survey is a weak answer to how to validate an idea. Money that has already moved is a better test. When people hire a freelancer to do a job by hand, the job has a price and buyers with a budget, which is what E3 records.
Our paid demand catalogue groups Fiverr gigs by the concrete task and lists 78,710 clusters. On its first page, creating fillable PDF forms has 87 gigs from 79 sellers, 12,877 buyer reviews and an average starting price of $14.75. Read buyer reviews as a conservative lower bound of paid demand, not a count of orders.
In the worked example the catalogue held nothing for the first two functions and was never crawled for the last two, so proof of payment there has to come from a commitment test.
The strongest case is when independent signals agree. Our published Squarespace export idea rests on 10 people who complained, about 2,220 Google searches a month for “squarespace alternatives” and related phrases, and 37+ paid Fiverr orders with a median order of $92. One complaint reads: “300 a year for a site you barely touch is wild when free options exist.” (source) Even there, no paid orders for an exit-specific tool were observed, so the price still needs a test.
Search volume shows interest, not a buyer
Search is E2. Check monthly volume for short, task-shaped phrases, not your product name. Google Keyword Planner gives average monthly searches and advertiser competition, and competition adds a commercial-intent signal.
Volume proves people look for a fix often enough to measure. It doesn’t prove they want your fix, would pay, or mean the same problem. The sample report sets a bar of 50 searches a month and counts no search volume for a function unless a confirmed pain backs it. If you grade ideas by hand, use the same two rules.
Check competitors before you trust the gap
An empty market can mean nobody has built the product yet, or that nobody can serve the buyer at a profit. Search the marketplace and the web for the job in plain words, install the closest products and write down what each one refuses to do.
In the sample report, the first function already had three observed competitors: LitCommerce, Sellbrite and Trunk (Stock Sync & Bundling). A competitor with paying customers is E4 evidence for the job: good news for demand, harder for entry. Your opening is the complaint their users still write.
Run experiments with a kill criterion
To move a function from “people said it” to “people pay”, run your own experiment and write down, before you start, the number that ends the idea. That number is the kill criterion. The What to check next section of the sample report lists three:
- Existing solutions. Search the App Store and the web for each function. Kill criterion: an existing app already offers all four functions together with good reviews and active users.
- Problem interviews. Interview 10 Shopify merchants who run several warehouses and have had sync failures, and ask how they find and fix stock discrepancies today. Kill criterion: fewer than 3 of 10 express strong interest or willingness to pay.
- Commitment. Describe the four functions on a landing page or mockup and ask 20 multi-warehouse merchants to pre-order or commit to a paid pilot at $29 a month. Kill criterion: fewer than 2 commit.
In interviews, ask what happened the last time the problem hit rather than whether they would buy a future product. The commitment test is the only step that produces payment evidence for your own product, so “sounds useful” does not count as a pass.
None of this needs a product. A mockup is enough for the commitment test, which is how to validate business ideas cheaply: the fastest way to test business ideas is to ask for money before an MVP exists. Market validation of ideas usually fails after the test, when a result lands under the line and the founder moves the line.
How to validate a business idea without money
Everything above except the paid pilot is free. If you need to know how to validate your business idea on no budget, read reviews and forum threads, count distinct people and save the links, then check freelance marketplaces for the job. Interview people from those threads, since they’ve already said they have the problem. The cost of validating business ideas this way is your time.
The Free plan on our pricing page includes Shopify and Atlassian pains with full evidence, and the sample report is public. Validating your own idea is a Pro feature at $15 a month, with 5 validations a day.
For a physical product or a local service, the rules on how to validate ideas stay the same and only the sources change: product reviews, local groups and conversations with customers. Our complaint corpus covers software only, so for a bakery or a skincare line you run these checks by hand.
FAQ
How do you validate a business idea when only one function has evidence?
Test that function alone. For the sample idea, that means interviews about comparing stock after a failed sync, with the other three functions parked until a merchant raises them unprompted.
Can I skip validation if I am the customer?
Being your own customer gives you a head start on the problem, but you are still one person. Knowing how to validate your idea matters most when you are already sure of it, because you will read every neutral result as a yes.
I have three ideas. Which one should I validate first?
Start with the one whose buyers you can already reach. Choosing between options is covered in what business should I start.
Your next step
That is how to validate a business idea in practice: split it into functions, find evidence for each, then run experiments with a line that ends them. Write your idea as functions tonight and grade each from E0 to E4. To run the same checks against real marketplace data, paste the idea into idea validation and read the report function by function.