Write Your Stop Conditions Before You Spend the First Dollar

The reason people stay too long on a dead product is not stupidity. It is that the decision to quit gets made in the same emotional state that the sunk cost created. The fix is procedural: define, in writing, at the start of the project, the specific results that will end it. Something like "if we cannot land three signed letters of intent at the target price within ninety days of having a working prototype, we stop," or "if landed unit cost exceeds $14 after the second value engineering pass, we stop."

Written stop conditions do two things. They turn an identity question into a data question, and they give co-founders and investors a shared standard so the conversation is not about who believes harder. Review them at each gate, and treat changing them as a formal decision requiring a reason, not something you do quietly at 11 p.m.

Five Signals the Idea Is Finished

The unit economics will not close. You have quoted real tooling and real parts, run at least one serious cost reduction pass, and the number still does not leave room for distribution margin. Cost problems that survive a redesign are usually structural, not a sourcing failure. The diagnosis path is in your prototype costs too much to manufacture.

Nobody will pay, only praise. You have shown it to fifty people in the target segment, they all said it was a great idea, and not one has put money down or signed anything. Enthusiasm without commitment is the single most reliable predictor of failure. If your validation so far has been conversations rather than transactions, redo it properly using how to validate a product idea before concluding either way.

The technical risk did not retire. Every project has one hard problem. If after two or three serious attempts the core mechanism still does not work reliably, and each attempt cost more than the last, you are not converging. Note the difference between "we have not solved it yet" and "each iteration teaches us less than the previous one." The second is the terminal condition.

The regulatory or IP path is closed. A blocking patent with broad claims that you cannot design around, or a classification that turns a $60,000 consumer product into a $900,000 regulated device with a multi-year approval cycle, changes what the venture is. Handle the first case with you found a similar patent, now what, and check the second early using does your product need regulatory approval.

You are out of runway and the next milestone is not fundable. Money is a hard constraint, and running a company down to zero to prove a point destroys the salvage value of everything you built. Stop while you can still pay to close things properly.

Quitting Versus Changing Direction

Most of what looks like abandonment should be redirection, and the distinction turns on which assumption failed.

  • Wrong customer, right product. Consumers will not pay $200 for it, but facilities managers will pay $600. Change the segment, keep the engineering.
  • Right problem, wrong solution. The pain is real, your mechanism is not the answer. Keep the market knowledge, rebuild the product.
  • Right product, wrong business model. Selling units fails, leasing or a consumable model works.
  • Right everything, wrong timing. A component is too expensive today and will not be in three years. This is a shelf decision, not a kill decision. Document it and move on.
  • Nothing is right. No paying customer, no working mechanism, no path to cost. That is a genuine stop.

The pivot mechanics, including how to do it without losing your team and investors, are covered in when to pivot a hardware startup.

Test Cheaply Before You Decide Either Way

Before killing a project, spend a small amount to remove ambiguity, because a kill decision made on guesses is as bad as a continue decision made on hope. A structured feasibility pass, described in feasibility study for a product idea, typically costs a fraction of one engineering month and answers the cost, technical, and market questions in parallel.

Cheap tests that produce real signal include a single-cavity soft tool quote instead of a full production tool quote, a landing page with a checkout to measure conversion rather than interest, one paid pilot with a single serious customer, and a quoted bill of materials from two contract manufacturers rather than an internal estimate. Each costs hundreds to low thousands of dollars and converts an argument into a number.

What to Salvage on the Way Out

A stopped project still contains value, and most founders walk away from it. Before you close the folder, extract:

  • The customer list and the interview notes. You now know a segment and its problems better than almost anyone. That knowledge transfers to the next idea.
  • The IP. A filed provisional or an issued patent can be sold or licensed even when you never shipped.
  • Reusable engineering. Mechanisms, firmware modules, test fixtures, and vendor relationships all carry forward.
  • The failure analysis. Write down which assumption killed it, in one page, and check every future idea against that list. The recurring patterns are catalogued in why new products fail.

Then choose the next thing deliberately rather than on rebound, using the ranking approach in how to choose which product idea to pursue. Founders who stop one project cleanly and start the next with what they learned tend to succeed on the second or third attempt. Founders who spend everything defending the first one rarely get a second.

Get an Outside Read Before You Decide

Projects House is regularly asked to assess whether a stalled project has a viable path left, and we say no when the answer is no. Describe where your project stands through our contact form and we will give you a direct assessment.