How to pick a narrow niche to start in

1 The bottom line

Start with the narrowest group that feels real pain and that you can reach, in a market where others already earn money. Winning a small group first gives you proof and a base to expand from.

2 Key takeaways

• Start from a real pain, not an app idea. People pay to fix pain, and you do not have to convince them they need it.
  • Migraine Buddy makes over $50,000 a month from about 15,000 downloads a month Steve P. Young, Migraine Buddy1:18.
• Pick a group so narrow it feels too small. Content converts better there, and few rivals copy what works.
  • Knitting, ballet, and wrestling and MMA are examples Canyon (Kenyon Prasad), Sideshift21:21.
  • Course Hero began with Cornell alone Andrew Grauer, Learneo.
• Choose a group you can reach and that can pay. A narrow group is no use if you cannot find it or it cannot buy.
  • BoldVoice found working professionals beat college students Anada Lakra, BoldVoice34:10.
  • Shotsy started with its founder's own Reddit group Rik Haandrikman, RevenueCat14:40.
• Enter a market where others already earn money. An empty market may mean nobody wants to pay.
  • Look for one to three apps above $100k a month Caleb Dean, Runafy54:06.
• Win the niche first, then expand to what is next to it. Starting broad and narrowing down is the harder route.
  • Breathwrk started with breathing only and was acquired by Peloton for $2.2 million Steve P. Young, Breathwrk9:12.

Narrow until you can win it.

Josh, Blue Throne

2.1 Start from a real pain

The best niches begin with a problem people already hurt from and will pay to solve.

• Pick a painkiller, not a vitamin, as Steve P. Young, Migraine Buddy1:18 says.
  • A need-to-have gets higher subscription conversion and needs far fewer downloads to reach $50,000 a month Steve P. Young, Migraine Buddy1:18.
  • Look for incumbents with weak onboarding and paywalls Steve Young, App Masters12:05.
  • Migraine Buddy shows about $70K a month from 30K iOS downloads on Sensor Tower Steve Young, App Masters12:05.
  • Dating is a pain that does not go away until solved, so the landing page needs little explaining Dmitri, YourMove AI29:55. His one named exception is longevity, which people pay a lot for.
• Understand one customer's problem deeply before you build. With building getting faster, this matters most Alexander Embiricos, OpenAI (Codex)54:40.
  • Spend day one researching the niche, the market and competitor apps Nicole Cheung, Glam Up11:03.
  • Write one key user, one problem and one use case, so the goal rules out approaches Cat Wu, Anthropic7:00.
  • Canva started with school yearbook coordinators who had no design experience, and spoke to every customer Melanie Perkins, Canva50:38.
• Pick a field where you know something others do not.
  • Build the app around what you know that somebody else does not Jacob Eiting, RevenueCat53:43.
  • Cursor's founders spent four months on mechanical engineering because it looked uncompetitive. They lacked the expertise and the excitement Michael Truell, Cursor15:35.

2.2 Define a bullseye group you can reach

Name the smallest group most likely to adopt first, and set everyone else aside for later.

• Define a comically narrow bullseye customer before you build or research. This is advice from Michael Margolis, GV9:45.
  • Narrowing prioritizes your roadmap, focuses your research and filters the feedback you get Michael Margolis, GV9:45.
  • Decide their language, money, problem, technical facility and temperament Seth Godin24:05.
  • If no one in a narrow group is thrilled the product exists, that is a warning sign Michael Margolis, GV9:45.
  • Pitching the Humane Pin to about 400 tinkerers on Product Hunt would have given room to dance Seth Godin24:05.
• Serve the users who feel the most pain.
  • GPTZero started with teachers, whose jobs ChatGPT disrupted Edward Tian, GPTZero8:18.
  • Runna built product and marketing around race trainers, its highest value users Miranda Paine, Runna47:52.
  • Teachers shared GPTZero with students, who became the much larger share of users Edward Tian, GPTZero8:18.
• Choose a group you can find and that can pay.
  • Build for a group you can reach, such as a Reddit group or a conference Rik Haandrikman, RevenueCat14:40.
  • Check ability and need to pay. International students improve on their own and have not felt the career cost yet Anada Lakra, BoldVoice34:10.
  • Tolan's Gen Z users dropped off after a day or two. Users 30 to 35 plus kept higher retention and conversion Ajay Mehta, Tolan13:49.
  • Tolan first designed for younger kids. Testing pulled it toward late teens and early 20s Ajay Mehta, Tolan13:40.

2.3 Check that the market is real

A good niche already shows demand and money, without being a trap.

• Look for proof that people pay in this space.
  • Enter a space where one to three apps make at least $100k a month. Take the concept to an adjacent area, like ranked gyms to ranked running Caleb Dean, Runafy54:06.
  • Filter Flippa or Acquire.com listings to iOS apps, since sellers must show revenue Bechir Arfaoui0:15.
  • Find the keywords first, then build for that search demand Steve Young, App Masters20:30.
  • Liftoff had held $100k a month or more for at least 12 months when Caleb validated Caleb Dean, Runafy54:06. Five thousand a month is not enough validation.
  • A new AI app reached $5,000 in its first month through ASO alone Steve Young, App Masters20:30.
  • Benchmark against similar apps. The RevenueCat median is 2.6% iOS download to paid by day 35 Steve P. Young, App Masters4:54.
• Prefer a crowded market with room for niches over an empty one.
  • An empty market may mean users do not want to pay Leon Sasson, Rise39:10.
  • Apply a novel idea to an already popular niche. That tests fewer unknowns than a blank slate Daniel Heintzman, Sway AI37:30.
  • Most Rise users do not own a wearable, so it serves a different user than Oura and Whoop Leon Sasson, Rise39:10.
• Be suspicious of ideas that get easy praise.
  • A tarpit is an idea many people try that draws good early validation Dalton Caldwell, Y Combinator24:40.
  • Pick steady demand over lottery topics Casper, Payout35:46.
  • Friends' night-out apps, Foursquare clones and music discovery are tarpits Dalton Caldwell, Y Combinator24:40.
  • Saving money videos swing from 100,000 to 10 million views. Side hustle videos always get at least 300,000 Casper, Payout35:46.

2.4 Cut your niche out of a bigger category

Take one slice of a big market and become known for it.

• Pull one neglected tentacle from a successful app.
  • Pick a content type or keyword a big brand added but cannot focus on Steve P. Young, App Masters0:32.
  • Calm expanded into sleep stories, soundscapes and focus music, and does not optimize ASO for each Steve P. Young, App Masters0:32.
  • One of Steve's apps built this way made $550,000 last year Steve P. Young, App Masters0:32.
• Position around one use case, even if the features are general.
  • Name the app and design the screenshots around one use case. You get known for one thing Steve P. Young, Photo Revive AI2:50.
  • Photo Revive AI offers the same text to image and text to video as generic apps. It is estimated at about $500K a month Steve P. Young, Photo Revive AI2:50.
• Narrow until you can credibly claim number one.
  • Keep it believable. The category needs real competitors and its own ecosystem Josh, Blue Throne9:34.
  • Size the market by cutting Sensor Tower into logical categories. The top 30 apps are usually the lion's share Josh, Blue Throne9:34.
  • Flow narrowed health to women's health and raised $200M+ at a $1B+ valuation Josh, Blue Throne9:34.
• Listen when users ask for a specific vertical.
  • People kept asking if a martial arts app did wrestling, so George built Wrestle AI George, Wrestle AI3:33.
  • Watch what users push a demo to do Jack McDermott, ElevenLabs13:00.
  • Wrestle AI passed 100,000 downloads George, Wrestle AI3:33.
  • Users pasted hundreds of thousands of characters into a 200-character demo box. ElevenReader then targeted fan fiction and web novels Jack McDermott, ElevenLabs13:00.

2.5 Win it, then expand

Own the small group first, then widen along what your users already do.

• Start narrow and move up, not the reverse.
  • Build a base of downloads and sales in the niche, then go up to broader categories Steve P. Young, Breathwrk9:12.
  • Start with box breathing and move into meditation, rather than fighting Calm and Headspace CP Young, Coursiv15:18.
  • Course Hero started with Cornell, repeated it course by course, and later spread across schools and geographies Andrew Grauer, Learneo.
  • Honeydew spent six months only in New York. Adding Florida doubled the business overnight David Futoran, Honeydew18:20.
• Expand by following what users already do.
  • Watch for jobs inside your product that deserve their own surface. Do not sort by TAM Dylan Field, Figma34:00.
  • Watch users who go through hoops to use your product in unintended ways, as Discord communities did Peter Sellis, Snap and Discord57:50.
  • A school asked to use Canva for newsletters, which revealed a wider gap Melanie Perkins, Canva50:38.
• Design features that serve the core group and the next one.
  • Zero's protein and hydration features serve both fasting and GLP-1 users Mike Maser, Zero38:14.
  • Apt began with personality assessments, then added career path recommendations, job matching and an AI mentor Tomas Hoyos, Apt20:00.

2.6 Where they disagree

Advice splits on how crowded the market should be, how small the first group can be, and how hard to start.

• Crowded or uncrowded.
  • Crowded markets with room for niches are good Leon Sasson, Rise39:10. Red oceans like mental health and food scanners are very hard without a unique product Yuliya Lennox, Solid Starts1:12:50.
  • Others prefer boring niches like pest control or funeral homes. Less competition lets you charge real money Andrew Wilkinson, Tiny6:26.
  • Cal AI won a seemingly saturated market through hustle and marketing Yuliya Lennox, Solid Starts1:12:50.
• Easy or hard first market.
  • Speak stayed about four years in South Korea, the most demanding English-learning market, to prove real differentiation Connor Zwick, Speak23:17.
  • Praktika chose Brazil for low English penetration and high income upside, despite weak purchasing power Adam Turaev, Praktika10:03.
• Fast money or long-term brand.
  • Superpower targets tastemakers first. They bring less money early but create the brand and culture Max Marchione, Superpower28:06.
  • Business subscription apps can charge $200 to $300 a year when users make back $2000 Alex, Boost App Social39:40.
• Small business or enterprise.
  • Pick one and never blend them. Small business is marketing-led and enterprise is sales-led Jen Abel, Jellyfish6:07.
  • SMB gives short feedback loops and millions of potential buyers, and the market pulls you upmarket over time Dharmesh Shah, HubSpot49:06.