Is My Business Idea Good? 8 Kill Criteria to Check

Is this a good business idea? Learn the kill criteria to check first, spot weak assumptions early, and know when to build, rethink, or stop.

Is My Business Idea Good? 8 Kill Criteria to Check

When you ask “is my idea good?”, people liking or disliking your product is only part of the picture.

Let’s take a software that warns finance teams before an expensive contract renews. There maybe companies who already missed a cancellation deadline or the finance manager wants to make sure there are no surprises. The problem is genuine, your product could help.

Then you discover that it needs access to confidential contracts the team cannot approve for a new supplier. Without those documents, the product cannot do the job. What happens now? The customer’s problem is still very real but your proposed business has a problem of its own.

A business idea is worth investigating when a specific customer has a reason to choose it, you can deliver the outcome at workable economics, and the essential requirements are achievable with the resources available.

Kill criteria help you identify where that breaks before you commit heavily.

What are kill criteria for a business idea?

Kill criteria are specific conditions that would make you stop investing in the current version of an idea. They might concern the customer’s willingness to pay, the cost of serving them, a dependency you cannot secure or a market too limited for your goals.

The important word is specific. “People might not trust us” is a concern. “The product requires access that our intended customers cannot authorize under conditions we can meet” is a potential stopping condition.

There is also a difference between an unknown and a failure. Not knowing whether customers will share the necessary data means you need to investigate. Establishing that the required data is unavailable means something in the plan has to change. You might abandon a subscription, narrow the customer segment or redesign the product around different inputs. You don’t necessarily have to abandon the underlying problem.

Don’t let several positives hide one deal-breaker

Suppose your research supports the problem, market size and price. It still does not solve an essential data-access problem. Equally, a weak finding about competition does not automatically invalidate an idea when customers have a clear reason to choose your approach.

Treat these checks as connected business requirements, rather than points to add up. The question is whether the important conditions can hold at the same time.

If you’re validating a SaaS idea, for the practical process of collecting customer evidence and running tests, read our guide to how to validate a SaaS idea before you build. Here, we’ll focus on what should change your decision.

8 Kill Criteria to check first

The examples below are hypothetical. Use the stopping conditions to examine your own buyer, offer and constraints rather than treating them as universal pass marks.

Before applying these checks, write down who the customer is, what problem they face and what result your product promises. If those are still vague, a disappointing test may reflect an unclear offer rather than a bad business idea.

1. The problem matters to the user, but nobody will fund the outcome

People can dislike a task without the business being willing to buy software for it. Consider a tool that reduces administrative work. The person doing that work may be enthusiastic about efficacy but the budget owner still needs a reason to prioritize the purchase. Ask what changes when the task takes less time. Does the business avoid contractor costs? Can the team handle more customers? Does it remove a recurring source of mistakes?

Be careful with the answer. Multiplying hours saved by an employee’s salary describes the value of freed capacity, yes, but it also does not automatically mean the company will spend less cash.

Pause this offer for this segment when relevant buyers understand the benefit but consistently consider it too minor to fund at a viable price. An existing software budget is not required from day one. What you need to establish is a credible reason someone would allocate resources to this outcome.

2. The improvement does not justify changing how the customer works

A better feature is not necessarily a compelling purchase. For the contract-renewal product, the finance team might already manage deadlines through a shared calendar and periodic reviews. Your software could make reminders easier while requiring a substantial effort to collect documents, verify extracted dates and establish a new process.

The comparison that’s actually useful includes that work. Ask what would make the customer’s current approach no longer acceptable. Look for a specific gap your product can close without imposing a larger burden elsewhere.

Reconsider the proposed product when customers recognize the improvement but repeatedly find that their existing approach is good enough once setup, trust and disruption are considered. The finding might point toward a smaller product. Perhaps the customer needs help with one part of the workflow rather than a replacement for the whole system. Changing behavior is not automatically fatal. It needs a benefit substantial enough to justify it.

3. The price cannot support the way you need to sell and deliver

An idea can attract customers and still have weak economics. Suppose a product charges $100 a month. Variable delivery costs are $30 per customer, leaving $70 before fixed costs. If acquiring each customer costs $700, recovering that acquisition spend takes ten months of contribution, assuming the customer stays and costs remain unchanged.

Those figures are illustrative arithmetic, not a forecast or benchmark.

Now consider what might be missing: onboarding, founder-led sales calls, ongoing support or human review of AI output. Early manual work can be a reasonable investment in learning. The question is which parts can become more efficient, what that requires and whether customers value the result at the necessary price.

Stop expanding this version when it only works by treating unavoidable sales and delivery effort as free, with no credible route to improving the economics. Possible changes include a different price, a narrower service promise or a customer segment with more valuable use cases. Each creates a new assumption to test.

4. The customer’s need ends before the next subscription payment

Some products solve a valuable problem only once. A contract audit might identify renewal dates and populate a reliable calendar. After that, the customer could have little reason to keep using the original tool.

This does not make the audit worthless, rather it creates a question about the proposed revenue model. What continues to change after the first result? Are new contracts arriving? Do people need ongoing review or action? Would customers value those activities enough to keep paying? Do not add a monitoring feature merely to create a subscription. Establish whether the continuing job exists.

Drop the recurring-revenue assumption when customers achieve the promised outcome and have no meaningful reason to return. A one-time product or service may still be worth pursuing.

5. An essential dependency is outside your reach

Before building the polished product, identify what must be available for it to work. It could be customer data, access to an existing system, specialist expertise or a level of output accuracy the customer can rely on.

For our contract example, the difficult question might be whether you can obtain representative documents and extract the relevant terms reliably. A successful demonstration using a few clean sample contracts does not answer that. Investigate the dependency under realistic conditions. Who controls access? What inputs can they provide? What happens when the documents are incomplete or inconsistent?

Stop the current design when a non-negotiable requirement cannot be met at a cost, quality level or timeline your business can support. A narrower design may survive. Working with approved exports, selected documents or human review could change the requirements. But those changes also affect the customer experience and economics, so update the whole plan.

6. The reachable market cannot support the business you want

“Is my business idea good?” also depends on what you want the business to become.

In a Hacker News discussion about choosing ideas, the original poster explicitly wanted a profitable business that could support them rather than a unicorn. This is a different objective from building for venture-scale returns.

Start with your intended revenue and a plausible price. Then examine how many suitable customers you would need. For illustration, $20,000 in monthly revenue at $200 per customer requires 100 paying customers. If your qualification criteria leave only 120 potential buyers, the plan requires you to win nearly all of them. The arithmetic should prompt closer investigation but it should not be considered a market estimate.

Check whether your definition of a suitable buyer matches the evidence about urgency, purchasing ability and product fit. A directory full of companies in the industry is not the same as a pool of customers for this offer.

Reconsider the target or business model when the plan requires an implausibly large share of the genuinely addressable customers. A smaller opportunity can still support a worthwhile business. Judge it against the business you intend to build.

7. The plan requires resources you cannot realistically provide

Some ideas are viable businesses for a different team. You might be able to build the software but have limited access to the buyers. OR you might have the customer relationships but lack the expertise needed to deliver a reliable product.

Investigate the operating requirements, not just whether the idea interests you. Could you support the customer during their working day? Can you fund the work needed before a usable pilot? Can you obtain the specialist help the product requires? A gap is not automatically a reason to stop. A partner, smaller initial scope or different route to market might resolve it.

Also ask whether you want the business this idea would create. If your goal is a small business you can run independently, an offer that requires frequent custom work and a large support team may not fit. Compare the potential return with the time, money and other opportunities you would give up. A business can be commercially viable and still be the wrong use of your resources.

Pause when the next essential milestone depends on resources you cannot obtain, and the plan has no credible alternative.

8. The market is not ready, and you cannot afford to wait

A real problem can exist before customers are ready to adopt your solution. Ask whether your plan works under today’s conditions or depends on something changing first.

For the contract-renewal example, rising supplier costs could make finance teams more interested in controlling renewals. But interest in cutting costs is not enough. Are those teams actually assigning someone to review contracts, allocating a budget or evaluating tools?

The same check applies to the technology. If the product only becomes affordable to deliver after processing costs fall, that future reduction is an assumption your business depends on.

Being early is not automatically a reason to stop. A narrower group of buyers may already be ready, even when the wider market is not.

Pause when the business depends on a future change with no credible timeline, and you cannot afford to wait for it. Record what would justify revisiting the idea, such as customers adopting the necessary systems or delivery costs reaching a workable level. “The market will eventually be ready” is not a reason to keep spending indefinitely.

Before you stop, check what actually failed

Predetermined stopping conditions can prevent endless rationalization, but they only help if the test actually measures the assumption you care about.

For example, if you send 15 introductory messages and only two people reply, you have learned something about outreach but not whether customers would pay. So before killing an idea, think about what did this test actually prove or fail to prove?

Here’s how that plays out in practice:

Make sure the evidence describes the same customer

There is another way to make an idea look stronger than it is. You find urgent demand among small businesses, attractive pricing among enterprises and easy acquisition through friends. Then you combine those observations into one business plan. But you have not yet found a customer segment where those conditions coexist.

Keep the context attached to the finding. A price accepted by an enterprise buyer does not resolve the economics of selling to a smaller business. Access through a personal relationship does not establish a repeatable acquisition route. Your plan needs a customer for whom the problem, buying process, price and delivery requirements fit together.

Write a stopping rule that names the decision

Avoid rules such as “stop if feedback is weak.” They leave too much room for interpretation.

For our contract example, a more useful rule would be:

Before building automatic contract ingestion, we need to demonstrate that we can extract the required renewal information from representative documents customers are authorized to share. We’ll agree on acceptable accuracy with the pilot buyer and set a limit on manual correction time. If we cannot meet those conditions, we’ll stop the automatic-ingestion design and reconsider the workflow.

The rule names the requirement, the evidence needed and the part of the plan that would change. Set the review date and spending limit before the experiment. A deadline should end an open-ended commitment, even when the result is “we still don’t know.”

There is evidence that a more structured approach changes how founders make these decisions. A 2024 study of 759 firms found that entrepreneurs taught to formulate and test hypotheses were more likely to stop pursuing ideas and made more focused changes to their strategy.

Research also points to a useful role for AI in the evaluation itself. A separate study in Strategy Science found that AI assessments of startup business plans were positively correlated with experienced investors’ assessments. It supports using AI as an analytical input, while leaving the actual business outcome unproven.

At NELL, we help you put that kind of scrutiny into practice. DeepValidate brings together market evidence, competitor research and commercial assumptions, highlights gaps in the case, and identifies what needs testing with real buyers. You can examine what supports the idea and what still needs to be resolved before investing further.

Use NELL to investigate where your idea could break

The difficult part is bringing these questions together. You may find evidence of a painful problem while discovering that buyers expect it to be solved inside a tool they already use. A promising price may depend on a customer segment you cannot yet serve. A large market may shrink considerably once you apply the actual buying requirements.

We built NELL to help founders examine those assumptions before committing more time and money. DeepValidate researches the commercial opportunity, shows the evidence behind its findings and identifies weaknesses that need further investigation.

Start with the business you are actually proposing

Describe the customer, the outcome and the assumptions the idea depends on.

For example:

Software for finance teams at companies with 50–200 employees that identifies contract renewal deadlines and notice periods. Customers would provide approved contract documents. We’re considering a monthly subscription and need to investigate data access, extraction accuracy, ongoing value and the buying process.

This gives the assessment a clearer starting point than “AI contract management.”

Our free Quick Validate provides an initial assessment, including dimension scores, a biggest unknown and a suggested next step. It is a first pass without live web research. DeepValidate is the deeper, sourced research option.

Use the report to examine the assumptions that could change your decision

DeepValidate organizes the investigation around the problem, the market, your ability to compete, monetization and the decision about proceeding. For the kill-criteria checks, the relevant sections include:

Read the full walkthrough of a NELL DeepValidate report to see how the findings, risks and recommendations fit together.

Turn the finding into a decision about what happens next

Look closely at the sections that could change your plan. If the research leaves willingness to pay unresolved, the next investment should help answer that question. If the important uncertainty is access to suitable data, a landing-page redesign is unlikely to resolve it.

The worked report includes risk mitigations, warning signs, evidence gaps and sequenced next steps. You can also refine the report as the proposed direction changes. Use those findings to choose the next experiment rather than treating a score as permission to build everything. Our score is a structured assessment, not a probability that the business will succeed.

But why use AI if AI is not your customer?

That is a fair question. AI cannot buy your product, approve a budget, switch from a competitor or tell you whether the problem is painful enough to act on.

So NELL is not trying to replace customer validation.

DeepValidate helps with the work that comes before those conversations: researching the market, comparing alternatives, checking pricing and economics, identifying weak assumptions, and showing you what still needs to be proven with real buyers. The point is to go into customer validation with better questions and fewer blind spots.

NELL helps you figure out what needs proving. Your customers provide the proof.

Start with NELL’s free Quick Validate for a first assessment. When you want the deeper research behind the decision, move to DeepValidate.

Let the idea earn the next investment

You do not need to prove the entire business before taking another step. You do need to know what that step is supposed to establish, what result would change your mind and how much you are prepared to invest before reviewing the decision. Start with the condition that could make the rest of the work unnecessary.

Run your idea through NELL’s free Quick Validate and use the first assessment to identify what deserves closer investigation.

Frequently asked questions

Is my business idea bad if competitors already exist?

Competitors are a reason to investigate the buying decision. Find out why customers choose them, what remains unresolved and whether your proposed improvement justifies another purchase or a switch. The existence of competitors alone does not answer whether your particular offer is worth pursuing.

Should I abandon an idea if nobody prepays?

First check what you tested. A missing prepayment can reflect weak value, an unclear offer or a purchase that requires more evidence before approval. Investigate the actual buying process. Do not count vague enthusiasm as a commitment, but do not assume every legitimate buyer must purchase an unfinished product immediately.

Can a high AI validation score outweigh a failed kill criterion?

An overall score should not override an essential requirement you have established cannot be met. Investigate the contradiction. The research may be incomplete, the requirement may be changeable, or the proposed version of the business may need to stop. A favorable average does not resolve the underlying issue.