📜 Manifesto
The Death Score Manifesto — Why 90% of Startup Ideas Should Never Exist
Every day, someone, somewhere, has an epiphany.
"I've got it!" they whisper, eyes wide at 2 AM. "What if there was an app that..."
They open their laptop. They draft a business plan. They start coding. Six months later, they launch to crickets. Zero users. Zero revenue. Zero interest.
Sound familiar?
Here's the uncomfortable truth: most startup ideas are terrible. Not just "needs more polish" terrible. "Should never have been born" terrible.
At DeathScore.ai, we built an AI specifically to validate this. Over the past few months, we've evaluated hundreds of startup ideas — from the absurd to the almost-viable. Our AI scores each idea on a 0-100 scale, analyzing market fit, unit economics, competitive moat, and execution feasibility.
The results are... grim. The median score across all submissions is a sobering reminder that most ideas aren't just bad — they're predictably bad. They fail in the same patterns, make the same mistakes, and delude their founders with the same cognitive biases.
This manifesto is about those patterns. Read it before you write another line of code.
I: The Seven Deadly Sins of Startup Ideas
We've identified seven patterns that account for ~85% of terrible startup ideas.
Sin #1: Solution In Search of a Problem
"I'm going to apply AI/blockchain/metaverse to..."
Stop right there. You've started with the technology and worked backward to find a use case. That's the startup equivalent of buying a Ferrari and then asking, "Where should I drive?"
The most successful companies — Stripe (payments were broken), Airbnb (hotels were full), Uber (taxis were terrible) — all started with a specific, painful problem.
Counter-test:
Can you describe the problem without mentioning your solution? If not, you don't understand the problem.
Sin #2: Uber for X
"Uber for dog walking!" / "Uber for laundry!" / "Uber for dental appointments!"
Here's the thing about Uber's model: it only works when three conditions align:
- High utilization (drivers can pick up another passenger immediately)
- Standardized service (a ride is a ride)
- Immediate demand (people need rides now)
Dental appointments violate all three. Dog walking violates at least two. Stop benchmarking your idea against the most successful platform business in history — unless you genuinely meet these conditions.
Sin #3: The "Everyone" Market
"Our target market is everyone who uses the internet."
No. No, it isn't. "Everyone" means you've done zero market segmentation. "Everyone" means you haven't identified who your first 100 customers are. "Everyone" means you're building for nobody.
Cold shower:
Find 10 real people who have this problem today and would pay to solve it. If you can't, your idea is dead.
Sin #4: Feature Posing as a Product
"A calendar app that schedules meetings for you." / "A smarter to-do list." / "A better note-taking app."
These aren't companies. They're features of existing products. Google Calendar, Notion, Todoist, and 47 other well-funded companies have entire teams working on exactly this.
Rule of thumb:
If a determined intern at an existing company could build your idea in a weekend, it's a feature, not a product.
Sin #5: The Revenue Fantasy
"We'll start free, get millions of users, and figure out monetization later."
This works for exactly two companies in history: Google and Facebook. Both had network effects so powerful that competitors couldn't replicate them. Unless your startup has the same gravitational pull — and it doesn't — you need a revenue model from day one.
Sin #6: The Sharpie-on-a-Napkin Economics
Every terrible idea has terrible unit economics. Here's the math most founders skip:
If your customer acquisition cost (CAC) is $50 and your lifetime value (LTV) is $30, you don't have a business — you have a charity.
We see founders who hand-wave this: "We'll optimize later!" No. You won't. Bad unit economics at launch rarely fix themselves — they metastasize.
Sin #7: Tech for Tech's Sake
"A blockchain-powered smart fork that vibrates when you eat too fast."
We actually saw this one. The AI scored it 8/100.
Just because you can add technology doesn't mean you should. The best startups use the simplest technology that solves the problem. Extra tech is debt, not equity.
II: The Red Flag Checklist
Before you pitch your idea to anyone, run this checklist. If you hit more than two red flags, it's time to pivot.
Scoring:
5+ red flags? Your idea belongs in the grave.
2-4 red flags? It needs major surgery.
0-1 red flags? Congratulations — you're in the top 10%.
III: The DeathScore Methodology
Here's how our AI actually evaluates your idea:
Scoring Dimensions (Each 0-10)
1. Problem Validity
Does this problem actually exist? Is it painful enough that people would pay to solve it? Is it a "vitamin" (nice to have) or "painkiller" (must have)?
2. Market Size
Is the addressable market large enough to build a business? Is it growing? Is it concentrated (easier to reach customers)?
3. Competitive Moat
Can competitors copy you? Do you have defensible advantages — network effects, data moats, regulatory barriers, brand?
4. Business Model
Do the unit economics work? Is the revenue model proven in adjacent markets? Can you reach profitability?
5. Execution Feasibility
Can you actually build this? Do you have the right team, resources, and timeline? Are there technical or regulatory showstoppers?
6. Timing
Is this the right moment for this idea? Too early (nobody's ready)? Too late (market saturated)? Or just right?
7. Founder-Idea Fit
Are you the right person to build this? Do you have domain expertise, personal pain, or unique insight?
The Balance Factor (-10 to +10)
A modifier for overall coherence. Great ideas have complementary dimensions (strong problem + strong model + strong founder). Bad ideas have internal contradictions.
The Final Formula
A perfect idea across all dimensions (70) plus the max balance factor (+10) = 100.
| Score | Verdict |
|---|---|
| 0-10 | DEAD ON ARRIVAL |
| 11-20 | CRITICAL CONDITION |
| 21-35 | NEEDS LIFE SUPPORT |
| 36-50 | MIGHT SURVIVE |
| 51-65 | SHOWING VITAL SIGNS |
| 66-80 | HEALTHY IDEA |
| 81-100 | UNICORN POTENTIAL |
IV: What Good Ideas Actually Look Like
After 500+ evaluations, here's what separated the 10% of ideas that scored above 50:
Pattern 1: Niche + Deep Expertise
The best ideas weren't broad — they were targeted. A tool for indie game developers to handle localization. Software for independent insurance brokers. A platform for wedding photographers to manage client galleries.
These founders didn't ask "what's a big market?" They asked "what specific pain do I understand better than anyone?"
Pattern 2: Pain > Pleasure
People pay more to solve problems than to gain benefits. B2B SaaS solves measurable pain (saving time, reducing cost). Consumer products that solve pain outperform those that add pleasure (another social network, another game).
Pattern 3: Existing Behavior + Better Experience
The most successful ideas don't create new behaviors — they make existing behaviors better. Airbnb didn't invent travel lodging. Stripe didn't invent payment processing. They took something people already did and made it dramatically easier.
Pattern 4: Clear First 100 Customers
Every founder of a successful idea I've evaluated could name their first 100 customers. Not "small businesses in Ohio." Actual names. They had a distribution plan before they had a product.
V: The Incubation Protocol
Your idea scored low. Now what?
Identify the lowest-scoring dimension
Is it market? Model? Timing? Focus your energy there.
Interview — don't survey
Talk to 20 people in your target market. Ask open-ended questions about their problems. Don't pitch your solution. Just listen.
Find the adjacent space
Your idea might be close to a good one. "AI bedtime stories for adults" scored 28, but "B2B sleep aid for hotels" could be viable. Don't kill the idea — reposition it.
Cut the scope by 90%
Your first version should be so small it's embarrassing. What's the minimum version that solves the core problem for 10 people? Build that. Launch it. Get feedback. Iterate.
Get a Death Score
Get your idea evaluated. Share the score. Let the roast humble you but not discourage you. The best founders are the ones who listen to brutal feedback and iterate.
VI: The Final Truth
Here's what I've learned from building DeathScore and evaluating hundreds of ideas:
"Good ideas are fragile. Bad ideas are stubborn."
A bad idea feels bulletproof at 2 AM. It has all the logic worked out. "Of course people want this."
A good idea feels terrifying. "What if I'm wrong? What if nobody cares? What if I waste a year of my life?"
That fear is the signal. If you're not afraid you're wrong, you haven't thought hard enough about your idea. The best founders we've seen are the ones who came in confident and left humbled — then went back to work with sharper assumptions.
"The DeathScore isn't the final word on your idea. But it should be the first."
Run your idea through before you pitch to investors. Before you incorporate. Before you buy the domain. Before you write a single line of code.
Because the cheapest feedback you can get is the one that hurts.
📖 More Startup Autopsies
About DeathScore
DeathScore.ai is a free AI-powered startup validator that evaluates your idea with brutal honesty. No signup required. No credit card. Just type your idea and get scored, roasted, and advised.
💀 Get Your Free Death ScoreWant more? Follow us for weekly startup idea autopsies.