Killing ideas early without ego damage
The signals that say "stop." How to walk away from your own idea before you sink three months into it.
Builders romanticize "grit" and "persistence" until those words quietly turn into "stubbornness" and "sunk cost." The most valuable skill in building isn't shipping. It's killing the idea before it eats your year.
Nobody talks about this part because it doesn't fit on a hoodie. "Always be shipping" is louder than "sometimes be stopping." But every builder you admire has a graveyard behind them, and the reason they got to the good thing is they didn't stay on the bad thing for nine extra months out of pride.
Here are the four signals that say stop. If you see two of these clearly, take it seriously. Three and you already know.
1. Nobody has a workaround
You've done ten user interviews and not one person currently solves this problem in a clunky way. No spreadsheet. No Zapier duct tape. No janky Notion template. No "I just deal with it."
That sounds counterintuitive. You'd think "no existing solution" means open market. It usually means no real pain. People route around problems that hurt. If they aren't routing, it doesn't hurt enough to pay for.
Example: you're building an AI tool to help freelancers track time across clients. Ten freelancers later, none of them track time at all. They invoice from memory. The "problem" is yours, not theirs.
2. The yes responses are polite, not specific
"Cool, I'd use that" is not a yes. "Send me a link when it's live" is not a yes. "I'd pay $50 a month for that today and here's my email" is a yes.
The test is specificity. A real buyer names a price, a timeline, or a current alternative they'd switch from. A polite person names nothing and smiles. If your interview notes are full of "interesting" and "neat," you have feedback that feels like validation and isn't.
Example: twelve people said your tool sounded great. Zero of them pre-paid when you offered a 50% discount for a deposit. The signal isn't the twelve. It's the zero.
3. The category leader has 100x your distribution
The market is real. Someone owns it. They have a million users, a sales team, and a marketing budget bigger than your rent.
"X for Y" against an entrenched leader is a long, losing fight unless you have a structural reason to win (a new platform, a pricing model they can't match without cannibalizing, a wedge they literally can't take). Building a slightly better version is not a wedge.
Example: you're building "Notion but for product managers." Notion has 30 million users and a PM template gallery. You have a Twitter account with 400 followers. The product can be better. It won't matter.
4. You can't say why YOU should build this
Close your laptop. Out loud, finish this sentence: "I'm the right person to build this because ___."
If the answer is "I had the idea" or "nobody else is doing it," that's not an edge. That's a starting line. An edge is a specific insight, a domain you've lived in for years, a network that gives you the first hundred customers, a technical capability that's rare. Without one, you're competing on pure execution against people with more time, more capital, and more reps.
Example: you want to build legal software but you've never worked in legal, don't know any lawyers, and can't articulate why the existing tools are wrong beyond "they look old." Pick a different fight.
The 60-day kill clause
Set a date before you start. Put it in your calendar. On day 60 you sit down honestly and check: do I have signal? Not "could I have signal eventually." Do I have it now.
The reason this works is the date exists before you're emotionally invested. Future-you will rationalize. Past-you was clear. Without a date, the project just slides, because every individual week feels too early to quit.
The ego problem
Killing is hard because you told people. They asked how it's going. Your mom retweeted you. Your group chat is invested. Walking away feels like admitting you were wrong in public.
Reframe: serial killers of their own ideas ship more, not less. The builders with multiple hits aren't the ones who white-knuckled one bad idea into the ground. They're the ones who killed quickly and freed up the year. Quitting badly conceived projects is a competitive advantage. Most people can't do it.
The one-thing-learned rule
Before you walk away, write down the single insight this attempt gave you. One sentence. "Freelancers don't track time at all." "PMs won't switch tools without team buy-in." "AI agents in legal need explainability before accuracy."
That sentence is not consolation. It's the input to the next idea. Killing isn't waste. It's research with a price tag.
When it's a pivot, not a kill
Sometimes the validation says the market is real but the product is wrong. People have the pain, they're paying for bad alternatives, they want something, just not what you built. That's a pivot.
Be honest about which one is happening. The trap is calling a kill a pivot to protect your ego, then spending another three months pivoting in circles. Rule of thumb: if you're keeping the customer and changing the product, it's a pivot. If you're changing the customer and the product, you're starting over and you should admit it.
Closer
The point of killing fast isn't pessimism. It's making room for the idea that actually works. Once you've decided what you're building (and what you're not), the next thing is showing up about it in a way that doesn't make you cringe.
That's the next guide: "Building in public without cringe."