Tar Pit Ideas
Browse the pit ↓
A field guide

Startup ideas that sound obvious,
attract great teams,
and almost never work.

A tar pit idea is one that makes intuitive sense to everyone you pitch — and still buries a long list of well-funded, talented teams.

Every tar pit idea is plausible. That's the trap. Each one fails for a specific, structural reason that the pitch does not reveal — a hidden defeater you typically discover two years and several million dollars later.

Below: a ranked list, why each trap snaps shut, the graveyard of attempts, and a sober bottom line. Read it before quitting your job.

13 classics + 6 trending Updated regularly Tongue partly in cheek

The top tar pit ideas, ranked

By how often founders pitch them. "Heat" is our rough sense of how frequently the idea shows up in YC office hours, seed decks, and brunch conversations. The Trending tag flags ideas that became hot in the last few months — usually AI-flavored, always still on the way to the graveyard.

# Idea Heat Verdict
★ AI SDR / outbound sales agents Trending Avoid
★ Ambient AI medical scribes Trending 10× harder
★ Agent infrastructure / "build agents without code" Trending Avoid
★ AI recruiter / autonomous hiring agents Trending Avoid
★ B2B AI voice agents Trending 10× harder
★ AI companions / AI girlfriends Trending Avoid
— Perennial classics —
01 A better todo / productivity app Pitched constantly Avoid
02 A marketplace for local services Pitched constantly 10× harder
03 A better email client Pitched constantly 10× harder
04 A social network for [profession] Pitched often 10× harder
05 A better calendar app Pitched often Avoid
06 Instagram for [niche] Pitched often 10× harder
07 A better Craigslist Pitched often 10× harder
08 An app for making plans with friends Pitched often Avoid
09 A new consumer chat app Pitched often Avoid
10 A personalized news app Pitched periodically Avoid
11 A Q&A site Pitched periodically Avoid
12 A hyperlocal / neighborhood app Pitched periodically Avoid
13 A music discovery / better Spotify Pitched periodically Avoid

See trending AI tar pits ↓

01

A better todo / productivity app

"Notion meets Things meets my actual brain."

Why founders love it

You use a todo app every day, you hate it, you know exactly how to fix it. The TAM is every knowledge worker. The MVP is a weekend.

Why the tar pit closes around you

  • Productivity is identity, not product. Users want a system that matches how they already think. There are seven billion such systems.
  • Switching costs are zero but inertia is infinite. Everyone's "current" app is good enough. Migrating is procrastination cosplay.
  • Distribution is an Apple/Google problem. Reminders and Google Tasks are pre-installed, free, and good enough.
  • The truly engaged users are tweakers. They will demand features forever and pay $4 a month.

Graveyard

Cultured Code's Things and Todoist survive as lifestyle businesses, not venture outcomes. The graveyard: Wunderlist (acquired by Microsoft, killed). Astrid (acquired by Yahoo, killed). Any.do (zombie). Cluster. Producteev. Sandglaz. Doo. Clear (still pretty, still niche). TeuxDeux. About twenty Notion competitors a year, each.

Possible escape hatches

Sell to teams, not people. Sell a vertical workflow that happens to have tasks in it (Linear for engineering, Asana for ops). The horizontal consumer todo is a hobby, not a company.

Bottom line

We recommend against this as a venture-backed company. The structural problems — distribution, retention, monetization — are not fixable by a better team. It can be a fine lifestyle business; that's a different conversation.

02

A marketplace for local services

"Uber for plumbers / cleaners / handymen / dog walkers."

Why founders love it

Everyone has hired a flaky handyman. The market is enormous. Mom understands the pitch. Software clearly beats Craigslist. You will be the Uber of [verb].

Why the tar pit closes around you

  • Disintermediation is rational and immediate. The plumber gets your number on visit one. Your take rate funds your competitors.
  • Frequency is brutal. Most people hire a handyman once or twice a year. There is no habit to capture and no LTV to amortize CAC against.
  • Supply is a knife fight. Good providers are rare, busy, and skeptical of apps. Bad providers churn through you and tank your reviews.
  • Trust is local, not network. Yelp + a neighbor's text is hard to beat with a five-star average across strangers.

Graveyard

Homejoy (cleaning, $40M, dead 2015). Pro.com. Zaarly. Exec. Handybook/Handy (acquired by ANGI, struggled). Thumbtack (still grinding after a decade, repositioned). TaskRabbit (acquired by IKEA, never broke out). Porch (pivoted). Setter. Hello Alfred (pivoted). Helpling. The entire 2012–2016 on-demand cohort.

Possible escape hatches

Pick one vertical, own scheduling + billing as a SaaS for the provider (Jobber, ServiceTitan, Housecall Pro), and stop thinking of yourself as a consumer marketplace. Or be Angi and lose money for twenty-five years until you bore everyone else out.

Bottom line

Possible, but expect the degree of difficulty to be roughly 10× what the pitch suggests. Plan capital, timeline, and patience accordingly — and assume disintermediation from day one.

03

A better email client

"Inbox Zero, but real this time."

Why founders love it

Email is hated by everyone, used by everyone, and the incumbents are sluggish. Surely a small team with taste can take a bite.

Why the tar pit closes around you

  • You don't own the protocol. Gmail and Outlook can change an API, rate limit you, or ship your feature on Tuesday.
  • Email is a system of record. Trust costs are insanely high. Losing one message is a one-star review for life.
  • Power users churn. You'll spend two years on keyboard shortcuts and discover the user went back to Gmail because Calendar didn't sync.
  • Monetization is "Superhuman or bust." $30/month from a tiny pool, or ads you don't sell.

Graveyard

Mailbox (acquired by Dropbox, killed). Sparrow (acquired by Google, killed). Inbox by Gmail (killed by Google itself!). Astro (acquired by Slack, killed). Newton (died, resurrected, died). Hey is alive as a paid niche. Superhuman is alive as luxury. Shift, Polymail, Front (pivoted to support-ops). Spike. Edison. Boxer.

Possible escape hatches

Sell to a specific team that lives in shared inboxes (Front, Missive). Don't try to replace Gmail for the individual.

Bottom line

Possible at the margins (luxury individual, team workflows), but expect roughly 10× the difficulty the pitch implies. Do not enter this market assuming Gmail is sleepy.

04

A social network for [profession]

"LinkedIn for doctors / lawyers / architects / scientists."

Why founders love it

The audience is specific, lucrative, and underserved by LinkedIn's everything-bagel feed. They'd clearly pay for a clean, professional space of their own.

Why the tar pit closes around you

  • Social networks are cold-start machines. The product has no value until it's full. You need everyone in the niche at once, and they will not come.
  • The niche is on LinkedIn already. Begrudgingly, but with their actual coworkers, recruiters, and clients there too.
  • Low frequency = no habit. Professionals check in monthly, then quarterly, then never.
  • Monetization is small. Even if you win the niche, the niche has six figures of users, not nine.

Graveyard

Sermo and Doximity made it work for doctors (the rare exception, on the back of CME and pharma money). Everyone else: Doostang. LawLink. ArchDaily-as-a-network. ResearchGate is propped up by reluctant academia. The professional Path / professional Ello of every vertical you can name.

Possible escape hatches

Sell the profession a tool that produces a side-effect of a network (Doximity is a HIPAA pager that happens to have a directory). Don't sell the network.

Bottom line

A handful of category-defining companies exist here, but they did not win by building a social network. The degree of difficulty is roughly 10× what it looks like — proceed only if your wedge is a tool, not a feed.

05

A better calendar app

"Calendar, reinvented, with AI scheduling."

Why founders love it

Google Calendar hasn't changed in fifteen years. Scheduling is universally painful. Surely there's room for taste, AI, and a clean inbox-for-time.

Why the tar pit closes around you

  • Calendar is OS-level infrastructure. Your "app" is a skin on Google's API.
  • The other side of the meeting doesn't use your app. Coordination is a network effect of one.
  • Free incumbents are everywhere. Calendly is the dominant verb and has a free tier.
  • Big tech sherlocks you instantly. See: Sunrise → Outlook Mobile, natural language → every calendar app.

Graveyard

Sunrise (acquired by Microsoft, killed, parts shipped in Outlook). Tempo AI (acquired by Salesforce, killed). Cue. UpTo. Donna. Mynd. Timeful (acquired by Google, killed). Cal Friendly. x.ai (Amy/Andrew, pivoted, sold for parts). Fantastical and Cron survive as paid skins. Motion is alive but expensive.

Possible escape hatches

Sell scheduling-as-a-feature to a workflow that already exists (Calendly inside sales). Don't try to replace the calendar.

Bottom line

We recommend against the standalone consumer calendar. A B2B scheduling product attached to a real workflow is a different conversation — and roughly 10× the difficulty of "we'll just make a better Google Calendar."

06

Instagram for [niche]

"Photo / video sharing, but for runners / chefs / cars / dogs."

Why founders love it

Instagram is too broad, TikTok is too chaotic, the niche has dedicated creators and vendors who'd love a clean place to live.

Why the tar pit closes around you

  • Creators go where the audience is. The audience is on Instagram and TikTok. You have neither.
  • Niches are too small to be a venture business and too big to easily own.
  • The feature gap is closing forever. Reels, Stories, Notes, and the Explore page eat every wedge.
  • Moderation costs scale before monetization does.

Graveyard

Path. Hipstamatic. Color. Cluster. Frontback. EyeEm (bankrupt). Vero (a meme of failure). Poparazzi. Dispo. BeReal (huge spike, ~zero retention, sold for cheap). Foodspotting (acquired by OpenTable, killed). Strava is the rare survivor — and it took a decade and a hardware-attached habit (a watch) to get there.

Possible escape hatches

Tie the social to a hardware loop the niche already has (GPS watch, dashcam, fishfinder). Don't try to win attention on phones from Meta.

Bottom line

Possible with a non-obvious wedge — usually hardware, regulation, or a behavior change the incumbents can't ship. Expect roughly 10× the difficulty of "we'll be a nicer Instagram for X."

07

A better Craigslist

"Unbundle Craigslist." (We've all said it.)

Why founders love it

Craigslist looks like 2001 and runs the country. Surely a modern team with auth, photos, payments, and chat can take a slice — and there are many slices.

Why the tar pit closes around you

  • The good slices are taken. Airbnb (housing), Indeed (jobs), Tinder (personals), Etsy/eBay (stuff), Zillow (real estate). What's left is harder.
  • Liquidity is local. You have to win city by city, and Facebook Marketplace is free and already there.
  • Trust + safety eats your margin. Scammers love a fresh marketplace.
  • Take rates collapse to zero. Marketplaces with no payment hook can't monetize the transaction.

Graveyard

Yardsale. Zaarly (again). Listia. Krrb. Bonanza. Wallapop (alive in Spain). letgo (merged into OfferUp, struggling). OfferUp itself is a slog. Mercari is public and shrinking. Carousell lives in SEA. The "Craigslist of X" pitch deck is the most common document on Earth.

Possible escape hatches

Pick one vertical with payment in the loop and a clear take-rate (StockX, GOAT, Poshmark). Avoid anything where the transaction happens in cash at a parking lot.

Bottom line

Possible in one vertical with a strong payment hook. Expect roughly 10× the difficulty of "we'll unbundle Craigslist" — and remember Facebook Marketplace is free, pre-installed, and patient.

08

An app for making plans with friends

"Group chat plus a calendar plus a poll. Finally."

Why founders love it

Coordinating five friends across iMessage, Google Calendar, and a Doodle poll is notoriously miserable. The market is everyone with friends. The wedge is obvious: group polls, smart suggestions, a shared timeline, RSVPs that actually update.

Why the tar pit closes around you

  • You need all the friends. The product has zero value with three of five attendees. The other two will not download your app for one brunch.
  • The default is iMessage. A group thread is already a free, pre-installed, universally adopted "planning app." Yes, it's bad. It's also there.
  • Frequency is catastrophic. The median user plans a group thing twice a quarter. Retention curves are not curves — they are cliffs.
  • There is no monetization. You can't charge for invites, ads against a tiny audience are worthless, and "premium" planning is an oxymoron.
  • Coordination is a social problem, not a software problem. The blocker is "Sarah hasn't replied," not a missing feature.

Graveyard

Cluster. Plancast. Wave (acquired by Google, killed). Down To Lunch (briefly viral, dead within a year). Free. Klutch. Squad. Backchannel. Yubl ($30M+ raised, shut down). Wigo. Hobnob (zombie). Punchbowl is alive on the back of paid invitations. Evite has been dying gracefully for two decades. Partiful is the rare bright spot — and even Partiful is a feature, not yet a public-company business.

Possible escape hatches

Pick a single high-stakes event where coordination really hurts (weddings → The Knot; group trips → TripIt-adjacent). Make the event the product. Or build inside an existing network (group features in Snapchat, Discord, iMessage). Don't build a standalone planning app.

Bottom line

We recommend against this as a venture-backed company. Group coordination is one of the most-attempted, least-successful categories in consumer software for structural reasons no team has solved in twenty years.

09

A new consumer chat app

"WhatsApp but ___."

Why founders love it

Group chat is universal, the incumbents are clunky, and you have one (1) very specific feature that will change everything.

Why the tar pit closes around you

  • Messaging is the strongest network effect in software. Your friends will not switch, and a chat app without your friends is a notepad.
  • Apple and Google own the default. iMessage is pre-installed and now speaks RCS.
  • Your feature will ship in iMessage within 18 months. Stickers, reactions, edit, undo, threads, voice notes — all once-startups.
  • Monetization is bleak. Even WhatsApp gave up on per-user fees.

Graveyard

GroupMe (acquired by Skype, fading). Beluga (acquired by Facebook, became Messenger groups). Path. MessageMe. Yo. tbh (acquired by Facebook, killed). Houseparty (acquired by Epic, killed). Highlight. Color ($41M pre-launch, dead in a year). Peach. Frontback. IRL (faked the users, indicted). Clubhouse (pivoted toward chat, then dropped off).

Possible escape hatches

Pick a teen subculture that wants to escape parents (Snapchat) and pray. Or build chat inside another product where the network is already captive (Discord on top of gaming, Slack on top of work).

Bottom line

We recommend against this. Standalone consumer chat is the strongest network-effect market in software, and the once-a-decade exception will not be you on a first-time founder's timeline.

10

A personalized news app

"The Daily, but actually for me."

Why founders love it

Twitter is noise, the homepage is dead, the algorithm is broken — clearly there's room for a smart curated feed. Especially now, with AI.

Why the tar pit closes around you

  • You don't own the content. Publishers will sue you, throttle you, or starve you out, depending on the decade.
  • The ad model is a knife fight against Meta and Google. You will lose.
  • News is a low-LTV habit. Users tap once and leave. Retention curves look like a cliff.
  • Distribution is Twitter, TikTok, and push notifications. All three are not yours.

Graveyard

Circa. Summly (acquired by Yahoo, killed). Prismatic. Digg (the original, sold for parts). News360. Pulse (acquired by LinkedIn, killed). Trove. Flipboard (zombie). Apple News (alive only because it's pre-installed). Artifact (Instagram founders, shut down 2024). Inkl.

Possible escape hatches

Be a paid newsletter platform (Substack). Pick one vertical and own it (Punchbowl for Congress, The Information for tech). Don't try to be the front page of the internet.

Bottom line

We recommend against this as a horizontal consumer product. The economics defeat execution. Build a paid platform for someone else's content instead.

11

A Q&A site

"Quora is bad, Reddit is chaos, ours will be smart."

Why founders love it

Asking experts a question is a universal use case. The internet's existing answer is embarrassing. Now, with AI, you'll do it for real.

Why the tar pit closes around you

  • You need experts who answer for free. Experts have jobs.
  • SEO is dominated by Quora, Reddit, and Stack Exchange. Until LLMs swallow them all and you lose the long tail too.
  • Quality degrades the second you scale. See: every Q&A site.
  • AI ate the easy questions. What's left is the long tail your tiny community can't cover.

Graveyard

Aardvark (acquired by Google, killed). Mahalo. ChaCha. Jelly (Biz Stone, twice). Formspring (ask.fm cousin, dead). Hipster. Branch. Quibb. Brainly survives in homework. Stack Overflow exists but cratered in 2023 from AI.

Possible escape hatches

If you must, sell expert answers as a service to a vertical (HealthTap-style, JustAnswer) and live or die on supply. Or build a community whose answers are a side-effect of a more valuable activity (Discord servers, Slack communities).

Bottom line

We recommend against this. The general-purpose Q&A site was always hard, and large language models have now eaten the head of the demand curve. The long tail is not enough to build on.

12

A hyperlocal / neighborhood app

"Like Nextdoor but not insane."

Why founders love it

Real-world community has collapsed online. You'll bring back the friendly block. Plus, the lost-cat moat is huge.

Why the tar pit closes around you

  • You have to seed every ZIP code individually. A national launch is 50,000 cold starts.
  • Engaged users are HOA-coded. The product that wins selects for the user that ruins it.
  • Moderation scales linearly with users and superlinearly with drama.
  • Local advertising is a knife fight with Yelp, Meta, and Google.

Graveyard

Yik Yak (twice). Brigade. Outvote. Hey Neighbor. Fonolo. Peach. Front Porch Forum (lives in Vermont). EveryBlock (closed by NBC). Foursquare pivoted out of consumer. Nextdoor itself is a public-market disappointment.

Possible escape hatches

Latch onto an existing local primitive — schools (ClassDojo), kids' sports (TeamSnap), buildings (an HOA SaaS). Don't try to be "the neighborhood."

Bottom line

We recommend against this. Hyperlocal is a thousand cold starts in a trench coat, and the engaged users are not the ones you want.

13

A music discovery / better Spotify

"Spotify's algorithm sucks. Mine will be soulful."

Why founders love it

Music is universal, taste is sacred, Spotify is bland, and you know exactly which seven songs to play someone to convert them.

Why the tar pit closes around you

  • The labels. Three of them, and they own the catalog. They will take most of your gross.
  • Subscriber economics are punishing. Even Spotify barely makes money.
  • Discovery is downstream of catalog. Without catalog, no users.
  • Apple is structurally subsidized. They can lose money forever to keep you off the platform.

Graveyard

Rdio (filed Chapter 11). Beats Music (acquired by Apple, folded in). Songza (acquired by Google, killed). Mog. Rhapsody/Napster (zombie). Grooveshark (sued out of existence). Turntable.fm. This Is My Jam. 8tracks. Songkick consumer. Hype Machine. Last.fm exists in name only. Endel survives by being a meditation app.

Possible escape hatches

Be a creator tool (Splice, BandLab). Be a hardware company (Sonos). Don't try to be a consumer streaming service.

Bottom line

We recommend against this as a consumer streaming play. Three label counterparties will take most of your gross before you've shipped a feature.

How to use this list

This is not a list of bad ideas. It's a list of hard ideas — markets shaped in a way that defeats most teams that walk in. A few of these have produced billion-dollar companies. They were exceptions, not refutations.

If you're seriously considering one of these, three sober questions:

  • What's structurally different now? "AI" is not an answer by itself. New regulation, a new platform, a new behavior, a new cost curve — those are answers.
  • Who specifically has tried this and why did they fail? If you can't list five and explain each, you don't know enough yet.
  • What's your unfair distribution? Tar pits punish "if we build it, they will come." You need a channel the graveyard didn't have.

If the answer is "we'll just execute better," respect the graveyard. They thought that too.

Submit a tar pit

Missing a classic? Found a new one? Tell us the idea, why it doesn't work, and who's tried. We'll add the strong ones.