YT YouTube2Text

YouTube2Text → Transcripts → How to Get and Evaluate Startup Ideas | Startup School

How to Get and Evaluate Startup Ideas | Startup School — Transcript

by Y Combinator · 6,118 words · 902 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:01[Music]
0:08all right hello everyone
0:11I've got a lot of content to get through
0:13so I'm gonna move fast Buckle in if you
0:17are looking for a startup idea right now
0:19I'm going to try to help but more
0:21importantly I'm going to try to give you
0:22the conceptual tools to think about
0:24startup ideas in a sophisticated way the
0:27way that we think about them at YC
0:28here's the thing uh no one not even YC
0:31knows for sure which ideas will succeed
0:34and in any case if your idea succeeds
0:37has much to do with how well you execute
0:38as your initial idea anyway but certain
0:41ideas are much more likely to succeed
0:43than others and so my goal here is to
0:46help you stack the deck in your favor by
0:48starting with a promising idea the
0:50advice in this talk came from several
0:52places first I analyzed the top 100 YC
0:55companies by evaluation and I looked at
0:57how they all got their idea so I started
0:59with some hard quantitative data on how
1:01recent billion dollar companies actually
1:03came up with their idea it also draws on
1:06a classic essay by Paul Graham that I
1:08really recommend it's called how to get
1:09startup ideas
1:11um it also comes from helping YC
1:12companies that pivot in the middle of
1:15the batch and learning over the years
1:16what advice helps them to find a new
1:19good startup idea
1:20and then finally it comes from Reading
1:23thousands of YC applications that we
1:25rejected and looking at the mistakes
1:27that caused good Founders to come up
1:29with bad ideas and those are the
1:31mistakes I want to help you all avoid
1:33this Talk's got three parts first I'm
1:35going to tell you the most common
1:36mistakes Founders make with startup
1:38ideas
1:39then I'm going to talk about how to know
1:40if your idea is good and then about how
1:42to come up with new ones okay the four
1:45most common mistakes with startup ideas
1:47so the most common mistake is just
1:49building something that doesn't solve a
1:51real problem for your users typically
1:54you can
1:55articulate the problem that you're
1:57solving you can put it in words but when
1:59you actually go and talk to the users
2:00it's just not something that they really
2:03care about
2:04we call it a solution in search of a
2:06problem or a cisp
2:09let's go through an example so a lot of
2:11Founders come up with an idea with this
2:13kind of thought process they go hmm AI
2:16is cool what could I apply AI to
2:20and then they go look for a problem that
2:21they could solve with AI That's a
2:24solution in search of a problem and the
2:25reason it's dangerous is that if you do
2:27that you'll probably find a problem but
2:30it will it will be a superficially
2:31plausible problem it'll be a made-up
2:33problem that people don't really care
2:35about rather than a real problem that
2:37people actually care about
2:39and if people don't really care about
2:40the problem they won't really care about
2:43your solution so instead you want to
2:45fall in love with a problem
2:48the best way to find a startup idea is
2:50to start with a high quality problem now
2:53Sunrise Founders hear this and they
2:55decide to interpret that as guidance to
2:58work on some like huge societal problem
3:00like I don't know Global poverty or
3:02something
3:03um no doubt those are real problems but
3:05they're they're too abstract to make
3:07good starting points for startup ideas
3:08you need something that's more specific
3:10something that's tractable with a
3:11startup the next mistake is getting
3:13stuck on what we call tar pit ideas
3:16what's a tar pit idea so there's this
3:19certain set of common startup ideas that
3:22have been around for forever they have
3:24been applying inroves to YC batch after
3:26batch for years and when founders start
3:29working on these ideas it's like they've
3:31gotten stuck in tar they never seem to
3:33go anywhere so we call them tar pit
3:35ideas here's what causes tarpet ideas
3:37they all form around some like
3:40widespread problem that lots of
3:41potential Founders encounter and it's a
3:44problem it seems like it could be easily
3:46solved with a startup but it's an
3:48illusion there's actually a structural
3:50reason why it's very hard or impossible
3:52to solve which is why after all these
3:54years no one has solved it
3:56and you can see why ideas like this
3:58would be so dangerous why they will
4:00cause so many Founders to waste months
4:02of their life stuck on a tarpet like
4:04they're very tantalizing from a distance
4:05because they're so superficially
4:07plausible as startup ideas here's a
4:09concrete example this is a very common
4:11tarpet idea that's been applied to YC
4:12for like 20 years this is like the
4:14stereotypical college student idea and
4:17it goes like this you think
4:19man every Friday or Saturday Saturday
4:22night when I'm making plans to meet up
4:24with my friends it's so inefficient I'm
4:26in all these different text threads and
4:28chat groups and we're like trying to
4:30make plans to meet up I'm just going to
4:32make an app to make it more efficient
4:34Well turns out that there are some
4:35structural reasons why this idea is hard
4:37which is why in like 20 years of people
4:40applying to YC with this idea nobody has
4:42actually pulled it off you can see why
4:44so many people have been attracted to it
4:45it's like it's a problem that almost
4:47everyone encounters at some point and it
4:49seems like it would be so easy to solve
4:51like you can just imagine the app it's
4:52just got like a list of events and you
4:54invite friends to it like it seems so
4:55simple the thing about tarpet ideas is
4:58that they are not necessarily impossible
5:00like I'm even open-minded that somebody
5:02will eventually make the like app to
5:04meet up with your friends idea of work
5:06is more accurate to think of them as
5:08common ideas that are much harder than
5:11they seem
5:13so if you want to work on one
5:15um here's my advice first Google it it's
5:19amazing how many Founders skip the step
5:21of just like Googling for their own
5:23startup idea to see who has worked on it
5:25in the past you should find who's worked
5:27on this in the past and actually talk to
5:29them if you can try to figure out what
5:32the hard part of this idea is that has
5:35caused other people to not be able to
5:37solve it yet the next mistake is simple
5:39it is amazing how many Founders will
5:43basically just like jump into the first
5:44idea they have without even stopping to
5:46consider whether it would actually make
5:48a good business
5:49but more dangerous is the founders on
5:52the opposite side of the spectrum who
5:54sit around waiting for the perfect
5:55startup idea and of course there is no
5:57such thing so these people just never
5:59actually start a company so if you
6:00imagine that there's like a spectrum
6:02between picking the first idea that
6:04comes to mind and waiting for the
6:05perfect idea and you know somewhere in
6:08the middle there is this like happy
6:09place which is the place that you want
6:11to be right and the way that Paul Graham
6:13put this is that you should think of
6:15your idea as a good starting point
6:19new startup idea is perfect and no
6:21matter what you start with it's probably
6:22going to morph anyway so you just want
6:24to have an initial idea that has enough
6:26interesting qualities that can morph in
6:29the right direction so now suppose you
6:30have a startup idea and you want to know
6:32if it's good I'm going to give you a
6:34framework for this and the format of the
6:37framework is 10 key questions to ask
6:39about any startup idea so the first one
6:41is do you have founder market fit if I
6:44depict like one most important criteria
6:46it'd probably be this one and what I
6:48mean by founder market fit is just are
6:50you the right team to be working on this
6:52idea
6:53and a great example of what good founder
6:56market fit looks like is plan grid so
6:58plan grid makes an iPad app to view
7:01construction Blueprints and to the
7:03founders of plan grid were Tracy and
7:05Ralph and Tracy had worked in the
7:08construction industry and she knew a lot
7:09about construction and Ralph was an
7:12awesome developer who was like the
7:14perfect person to build this iPad out if
7:15you're going to imagine a team to start
7:17playing grid the team that you would
7:19imagine would look you know something
7:21like that and that's what good founder
7:22market fit looks like it's like this
7:24team is obviously the right team to work
7:26on the idea in fact founder market fit
7:28is so important
7:30that I would recast your search for a
7:33startup idea when most people go to pick
7:36a startup idea they try to look for a
7:39good startup idea like in the abstract
7:42and instead I would think about this
7:44exercise as an exercise to pick a good
7:46idea for your team you with me it
7:49doesn't matter if something is a good
7:51startup idea for someone else if it's
7:53not a good idea for your team so you may
7:56as well just look for ideas that you
7:58would actually be good at executing okay
7:59number two how big is the market
8:01obviously you need a big Market which
8:03for startups typically means like a
8:04billion dollar market but actually less
8:07obviously there are two kinds of markets
8:09for startups that are good
8:11um ones that are big now and ones that
8:13are small but rapidly growing
8:15and an example of the second one is
8:17coinbase so when coinbase got started in
8:202012 the Bitcoin trading Market was
8:22minuscule but even at that time it was
8:24pretty obvious that if Bitcoin succeeded
8:27the way people hoped that it would that
8:29this would eventually be a billion
8:30dollar market
8:31number three how acute is this problem
8:34so as I said earlier the most common
8:36mistake is just like working on
8:38something that just isn't really a
8:39problem or it's just not a problem that
8:41people care enough about
8:43um here's an example of the opposite
8:45here's an example of what a good problem
8:47looks like Rex
8:50so Brax from Winter 2017 makes a credit
8:53card for startups
8:55and before brex if a startup NYC wanted
8:58a corporate credit card they literally
9:00could not get one because a no bank
9:02would give a credit card to a startup
9:05that's a good problem like if you're if
9:07the alternative to your solution is
9:09literally nothing that's what a good
9:10problem looks like okay next do you have
9:13competition
9:14now most Founders think that if you have
9:16competition that that's bad but
9:18counter-intuitively it is the opposite
9:21most good startup ideas have competition
9:24but if you were going up against
9:26especially entrenched competition you
9:28typically need a new insight next one is
9:31like do you want this personally
9:34do you know people personally who want
9:36this it's amazing how often people start
9:39companies where the answer to both these
9:41questions is no if that's the case you
9:44definitely got to worry that you know
9:46maybe nobody wants this so definitely
9:48time to go talk to some users
9:51only recently become possible or only
9:53recently become necessary
9:55so something has recently changed in the
9:57world like a new technology regulatory
10:00change or a new problem that is often
10:04what creates a new opportunity
10:06and a great example of this is company
10:08called Checker which does background
10:10checks via VIA an API so it's an API for
10:14doing background checks on people and
10:16roughly the story of Checker is delivery
10:20services like doordash and instacart and
10:23Uber started to take off
10:25and they were all hiring huge pools of
10:27delivery people and workers and they
10:29needed to run background checks on all
10:30of these people and there were at the
10:33time already a bunch of large existing
10:35companies that run background checks but
10:37they weren't well suited for this very
10:39new use case and that is like exactly
10:42the kind of change in the world that
10:44creates a new opportunity let's talk
10:47about proxies
10:48so a proxy is a large company that does
10:52something similar to your startup but it
10:54is not a direct competitor
10:56and so um
10:58a good example of this in practice is a
11:00company called rapid which does food
11:02delivery in Latin America and when
11:04rapping got started there were already a
11:06few delivery companies in other parts of
11:07the world like doordash they're doing
11:09very well they just hadn't caught on in
11:11Latin America yet and so doordash was a
11:14great proxy to show that this idea of
11:16doing food delivery in Latin America
11:17would probably work is this an idea
11:19you'd want to work on for years but this
11:22is a tricky one like sure if the answer
11:24to this question is yes that's a good
11:26sign but often it's not often an idea
11:29grows on Founders over time as it starts
11:32to work as I'm going to talk about in a
11:34moment a lot of the best startup ideas
11:35are in boring spaces like tax accounting
11:39software or something like that like no
11:40one is particularly passionate about
11:41like nobody starts off being passionate
11:44about tax accounting software
11:46um but tax accounting software is
11:48probably a good business and if you're
11:50actually running a successful business
11:52you tend to become passionate about it
11:54over time okay is this a scalable
11:57business
11:57so if you're building pure software the
12:00answer is yes because software skills
12:01infinitely and you can just like check
12:03this one off
12:04um the place for Founders most often get
12:06into trouble here is with Services
12:08businesses like agencies or Dev shops
12:10anything that requires like high skill
12:12human labor in order to serve your
12:14customers okay and my last question is
12:17is this a good idea space which of
12:20course means I need to tell you what an
12:22idea space is this is a concept from my
12:24colleague Dalton who you'll hear from
12:26later in this course
12:27an idea space is like one level of
12:30abstraction out from a particular
12:32startup idea it is a class of closely
12:35related startup ideas
12:37like software for hospitals or
12:40infrastructure monitoring tools or food
12:42delivery services and here's the thing
12:45different idea spaces have wildly
12:48different hit rates
12:49over the last 10 years if you started a
12:52company that did like fintech
12:54infrastructure or vertical SAS for
12:57Enterprise the the probability that your
12:59company became a billion dollar company
13:01was astonishingly High
13:03whereas if you started something in
13:05consumer Hardware or social networks or
13:08ad Tech the success rate was like orders
13:12of magnitude loader and I can't say that
13:13that will continue to be the case for
13:15those specific areas because spaces
13:17flipped from hot to cold over time but
13:19it is still worth thinking about picking
13:22a good idea space
13:24and a good idea Space is really just one
13:26that like you expect is going to have a
13:28reasonable hit rate for new startup
13:30ideas and one that has founder market
13:32fit
13:33that way even if your initial idea isn't
13:35quite right there are probably good
13:37adjacent ideas that you can sort of like
13:39drift into
13:40I'm going to give a good example of like
13:42how this played out in practice so a
13:43good example of picking a good idea
13:45space is this company called 5tran from
13:47yc1213 and basically it's already 5tran
13:50is they started making this tool for
13:52data analysis
13:54and they went to some companies and they
13:56tried to sell it and the companies
13:58didn't want it so they pivoted and they
14:00built a different tool for data analysis
14:02and they went back to the same companies
14:04and they tried to sell them that one and
14:06the companies didn't want that either
14:08but each time they went to companies and
14:10tried to sell them some tool for data
14:12analysis they would learn more about
14:13what those companies actually wanted and
14:15So eventually they sort of stumbled into
14:17an actual problem into an actual tool
14:19for data analysis that companies
14:21actually wanted
14:22and
14:23this is why picking a good idea space to
14:26start with is so important like because
14:28the five Tran Founders were shopping for
14:30ideas and a fertile idea space they put
14:33themselves in a good position to like
14:35bump into a good startup idea if they
14:37had picked a bad idea space they
14:39probably wouldn't have found anything
14:40okay before I talk about how to generate
14:43startup ideas I've got one important
14:45topic to tell you about
14:47these are three things that make your
14:50startup idea seem bad
14:53but actually make them good and the
14:55reason that they make them good is that
14:57most Founders will shy away from ideas
15:00like these which leaves them on the
15:02table for smarter Founders to go and
15:04grab them and here they are
15:06ideas that are hard to get started ideas
15:09that are in a boring space and ideas
15:12that have existing competitors
15:14ideas that are hard to get started so
15:17Paul Graham wrote a terrific article
15:19about this called schlep blindness which
15:21I'd really recommend reading and the
15:22example that Paul discusses is stripe
15:24you all know stripe they make it easy to
15:26integrate credit card payments to your
15:28website
15:28and the fascinating thing about stripe
15:31is that when stripe launched there were
15:33thousands of developers who already knew
15:36that this was a problem they had tried
15:39to integrate credit card payments to
15:40their site and they realized that the
15:41existing options sucked
15:43but strangely not one of them even tried
15:47to start striping and it's kind of a
15:50kind of a fascinating question why
15:51nobody else tried when so many people
15:53were in a perfect position to see this
15:55problem
15:56and the reason is that getting
16:00started building stripe required some
16:03things that seemed really hard you had
16:05to get a special deal with a bank you
16:08get to learn a lot about the
16:09nitty-gritty details of credit card
16:11infrastructure those things seemed so
16:14hard that they scared off all the other
16:17people who might have started stripe
16:18which caused them to leave this like 100
16:21billion dollar opportunity on the table
16:22for the stripe Founders to go and pick
16:24up okay the second one is ideas that are
16:27in a boring space and a great example of
16:30this is Gusto which makes payroll
16:32software
16:34payroll software pretty boring right
16:37the thing is there are thousands of
16:39people who must have realized that
16:40payroll software sucked but because it
16:42was kind of a boring problem nobody
16:44tried to fix it until the Gusto Founders
16:45came along and the thing is that because
16:47most Founders shy away from ideas like
16:50this
16:51boring ideas like payroll software have
16:53a much higher hit rate than fun ideas
16:56like
16:57apps to find new restaurants to eat at
17:00or like apps to find the next song to
17:02listen to something like that like fun
17:04ideas get picked over boring ideas get
17:07left on the table for a long time
17:10now you might be thinking Jared why
17:13would I want to work on a boring idea
17:15that sounds you know boring
17:17but here's the thing even if you work on
17:19an idea that sounds fun at the outset
17:22the day-to-day reality of your startup
17:24is going to be mostly the same anyway
17:27either way you'll be mostly writing code
17:30fixing bugs talking to users like pretty
17:34much the same stuff and so I would argue
17:36that once the initial excitement of your
17:39idea has worn off and you are 6 or 12
17:43months in and you are grinding out the
17:45execution that makes your idea actually
17:47work
17:49how fun the initial idea sounded
17:51actually has little to no correlation
17:53with how much fun you will actually be
17:56having working on your company
17:59and the last one is that Founders
18:01incorrectly shy away from spaces where
18:03there are existing competitors
18:04counter-intuitively most startup ideas
18:07most good startup ideas have existing
18:09competitors when founders go into spaces
18:11with no existing competitors they often
18:13find out that the reason that there are
18:15no competitors is because no one wants
18:17the product
18:18a great situation is actually a market
18:21where there are existing competitors but
18:23you've noticed something that they all
18:24seem to have missed or they all just
18:26kind of suck a classic example of this
18:28is Dropbox so when Dropbox launched they
18:31were already about 20.
18:34cloud-based file storage companies
18:37Dropbox was like the 20th company and
18:39it's space to launch naively you might
18:41have thought that that made this a bad
18:42Market I mean you would have gone and
18:44looked and said like oh there are
18:45already 20 competitors it seems like a
18:46bad Market to go into but if you were
18:48Savvy about startup ideas you would have
18:50realized that it actually made it a
18:52great Market and the reason is that even
18:54though there were like 20 companies
18:56doing this most people didn't use any of
18:59them and that strongly suggests that
19:02there actually was a problem here but
19:04the existing products hadn't solved it
19:07and Drew the founder of Dropbox had a
19:10very specific insight about what all of
19:12them were missing his Insight was
19:14basically their UI sucked and the reason
19:16their UI stopped was that at the time
19:18the way that they all worked is you had
19:20to like go to their website and manually
19:22upload your files one at a time into
19:24their website which sucked of course and
19:27Drew had really a technical Insight
19:29which was that if he integrated directly
19:31into the host operating system he could
19:33just sync your files automatically
19:34without you having to do anything and
19:36that was a real step function change in
19:38how convenient these Services were to
19:40use and that was the right insight okay
19:43let's talk about how to come up with
19:46startup ideas so it is possible to sit
19:49down and explicitly think of startup
19:51ideas and in a moment I'm going to talk
19:53about how to do this but it is actually
19:55not the best way the best way to have
19:57startup ideas is to just notice them
19:59organically and if you look at the YC
20:01top 100 companies at least 70 percent of
20:04them have their startup ideas
20:05organically rather than by like sitting
20:07down and explicitly trying to think of a
20:09startup idea and the problem is that
20:11when people sit down and try to think of
20:13startup ideas they tend to think of bad
20:15ones they're especially likely to think
20:18of the same set of tar pit ideas that I
20:20talked about earlier whereas startup
20:22ideas that occur to you organically are
20:24more likely to be good ones so if you're
20:27not planning on starting a company
20:28imminently and you just want to put
20:30yourself in a position to have organic
20:32startup ideas in the future here are
20:34three ways to do that this is like
20:36playing the long game to set yourself up
20:38for future success
20:39um first is just like become an expert
20:41on something valuable if you're working
20:44at the Forefront of some field you'll
20:46see good startup ideas in that field
20:48and a great way to do that is to go work
20:50at a startup Harge talked about this in
20:52his talk last week if you're working at
20:54a startup you will become an expert in
20:56the thing that that startup does and
20:57that is really putting yourself in a
20:59position to have great startup ideas and
21:01finally if you're a programmer one thing
21:03that can work is to just build things
21:05that you find interesting even if
21:07they're not businesses they're not
21:08clearly startup ideas sometimes they
21:10like
21:12turn into them over time and a really
21:16striking example of this is the story of
21:18replica this is exactly how replica
21:20started it was just something that amjad
21:22found interesting it wasn't supposed to
21:24be a startup originally okay but if you
21:26want to generate ideas for startups
21:28right now I'm going to walk through
21:29seven recipes for doing that I've tried
21:32to list these in order of How likely
21:34they are to lead to actually good ideas
21:37so start with the first ones okay here's
21:40the first one and the best one
21:43start with what your team is especially
21:46good at
21:47and think of ideas that take advantage
21:49of your expertise the reason that this
21:52is so effective is that any idea you
21:54come up with this way has automatic
21:56founder market fit do you see how this
21:58is almost like a hack to generate the
22:00set of ideas that has founder market fit
22:02here's a great example of like how this
22:04worked in practice
22:05so a good example is this company resi
22:08which is like open door for rental
22:10apartments and before starting resi the
22:13founders have worked in real estate and
22:15debt financing and they were experts in
22:17those areas
22:18and where they got into YC they spent
22:21the first month looking for ideas
22:23but the smart thing that they did is
22:25they only looked at ideas in that idea
22:28space in like the rough idea space at
22:30like the intersection of like real
22:31estate and fintech and that was a smart
22:34move because that is a very fertile idea
22:36space for startups like many billion
22:39dollar companies have come out of that
22:40idea space and the resi founders were
22:42experts in that idea space and so
22:44because of that
22:45their search for a startup idea was
22:47pretty quick and painless and pretty
22:49quickly they came up with the idea for
22:51Rezi which is like an excellent idea and
22:53has perfect founder market fit so if you
22:56have specific expertise like the resi
22:58founders did you should definitely start
23:00by looking at ideas in you know the
23:03things that you're experts in
23:05um it is weird how many startups apply
23:08to YC and we look at their applications
23:10and the founders like are actually
23:11legitimate experts in something but the
23:13idea that they're applying with is like
23:15something completely different
23:17now if you're if you're a young founder
23:19if you're in college or something you
23:21may not have had a chance to develop
23:23it's like the level of domain expertise
23:25that the resi founders had so this may
23:27not be the right recipe for you so let's
23:29talk about some other ones
23:31the next recipe is to start with a
23:33problem you've personally encountered
23:35ideally one that you're in an unusual
23:38position to see so vetco is a website
23:43for veterinarians to order supplies so
23:46you can think about it like amazon.com
23:48for vets
23:49and the short story of vetcove is that
23:52the founders are brothers
23:53and their dad is a veterinarian
23:56and growing up they would notice that
23:58the way he ordered supplies was like
24:00super old-fashioned like you'd have to
24:02like call up a supplier on the phone and
24:05like order stuff through like a 1-800
24:07number or something it was like very
24:10obvious that you could build an
24:11amazon.com kind of thing that would
24:13replace that
24:14and I I love the story of vetco because
24:16it's such a great example of what a
24:20great startup opportunity looks like
24:22because thousands of veterinarians must
24:25have known that this was a real problem
24:27that it was really annoying there wasn't
24:28like a basic website where you could
24:30just go to order supplies but the thing
24:32is veterinarians don't start Tech
24:34startups very often and then on the
24:36other hand you had like thousands of
24:38programmers in Silicon Valley who were
24:40begging their heads against cisps and
24:42tarpet ideas totally unaware that over
24:45here there was like a really great
24:47genuine problem to work on
24:50and because of this when the vet
24:52co-founders got started they had no
24:54competition so this amazing idea just
24:56got left on the table for years so if
24:59you want to try to actually use recipes
25:01one and two here's a specific set of
25:03instructions for how to do it this is
25:05advice I often give to Founders NYC who
25:07are pivoting and who are looking for a
25:09new idea here's how it works
25:12for each founder on your team
25:14go through every job you've ever had
25:17plus all your internships plus like
25:20other life experiences
25:23and think really carefully about each of
25:25them
25:26what problems did you come across
25:30what did you learn that other people
25:32don't know
25:34What are problems or opportunities that
25:36you've been in kind of a special
25:38position to see
25:40those are the best places to start
25:42looking for startup ideas okay the next
25:45one is to just think of things you
25:47personally wish existed this is like a
25:49really classic recipe this is very
25:50common advice great example of this is
25:52doordash the short story of doordash is
25:55that the founders of doordash were
25:56undergrads at Stanford and the thing
25:58they really wanted was to be able to
26:00just order food from local restaurants
26:01and have it delivered to their dorm and
26:03before doordash you couldn't do that so
26:05they started doordash and this is a
26:08great recipe but
26:10this is the recipe that is most
26:13dangerous and potentially leading to tar
26:15pit ideas so if you're using this recipe
26:17you just gotta stop and think for a
26:19second is there a reason why this thing
26:20doesn't exist yet okay the next one is
26:23to look for things in the world that
26:24have changed recently that might have
26:26created a new opportunity a great
26:28example of this is covet
26:30so when the pandemic started it changed
26:32daily life for all of us and many
26:35Founders realize that this created the
26:37opportunity for new companies and some
26:39very successful startups came out of it
26:41and
26:42one of them is this company called
26:44gather town which builds like this fun
26:46way to hang out with other people online
26:48and the founders of gathered town were
26:50actually working on a different idea
26:51that wasn't going so well and when the
26:53pandemic started they pivoted to this
26:55because it was obvious that like the
26:57change in behavior of the pandemic had
26:59created a bunch of new opportunities you
27:01can also look for companies that have
27:03been successful recently and look for
27:06new variants on them
27:08a good example of this is a company
27:10called Nuvo cargo Nuvo cargo is a good
27:13example of explicitly sitting down to
27:15try to think about startup ideas and
27:16actually finding a good one which is the
27:17thing that I told you is hard to do
27:19pneumo cargo is basically flexport for
27:21Latin America they help U.S companies to
27:24import stuff from Mexico
27:26and the story of nuva cargo is that
27:28Deepak the founder was working on a
27:30different idea NYC and he realized that
27:32his idea wasn't going to work
27:35and he went on like a systematic search
27:37for better ideas
27:39and he picked Nuvo cargo for for very
27:42analytical reasons he picked it because
27:44it was a large Market
27:45because they were good
27:47proxies from other companies flexport
27:49and he picked it even though he didn't
27:52have deep domain expertise in the import
27:54export space because he had some
27:57connections that would enable him to get
27:58started and he felt like he would just
28:00be very good at running this kind of
28:02operationally intensive business
28:04and that worked really well Niva cargo
28:06is doing super well you can also go and
28:09talk to people and just ask them what
28:11problems they have this this can work
28:13the the downside with this recipe is it
28:15actually requires a lot of skill if you
28:17want to do this I would recommend first
28:19picking a fertile idea space and then
28:23going and talking to people within that
28:25idea space and I would also recommend
28:27talking not just to potential customers
28:30but also to potent but also to founders
28:33of companies in that idea space to get
28:36advice about what ideas are actually
28:38worth pursuing and a good example of
28:40doing this successfully is a to B and A
28:43to B was in the position that a lot of
28:45Founders that struggle to find good
28:47startup ideas are in which is that the
28:48founders were pretty young and they
28:50hadn't acquired a lot of specific domain
28:52expertise like the resi founders had yet
28:54but they really wanted to do a startup
28:56and they wanted to do a startup that had
28:57like a genuinely good idea the way that
28:59they came up with the idea for ADB was
29:01very systematic and so I'm going to walk
29:03you through how they did it and really
29:05break it down for you so A to B makes
29:08fuel cards and if you haven't heard of a
29:10fuel card it's like a special kind of
29:12credit card for truck drivers when the A
29:15to B Founders got into YC they pivoted
29:17and they spent the whole YC batch
29:19looking for a new idea and here's how
29:21they did it first they picked an idea
29:24space and the idea space that they
29:26picked was essentially software for the
29:28trucking industry
29:30and they picked this idea space despite
29:33not being experts at the trucking
29:35industry because they felt that it just
29:38should be a fertile idea space to go
29:40hunting for startup ideas the trucking
29:42industry is as like Big Industry hasn't
29:45been that disrupted by startups and
29:47software yet so they just felt like
29:48there were probably some good problems
29:50to work on in the trucking industry but
29:53the problem is
29:54they didn't know that much about the
29:55trucking industry so they didn't know
29:57what those problems were and so they
29:58decided that they would turn themselves
30:01into experts in the trucking industry
30:04and what they did is they actually
30:06physically drove two truck stops which
30:08were places where truck drivers are just
30:10kind of
30:11Milling about and they would just walk
30:13up to truck drivers and like start
30:16talking to them and ask them questions
30:18about what their problems were
30:20they also talked to a lot of Founders
30:23who had started companies in the
30:24trucking space to get ideas for what
30:28problems were actually worth working on
30:30they would basically talk to like anyone
30:32who knew anything about Trucking that
30:34was willing to talk to them and as they
30:36did that they began to put together like
30:38a mental map of the space and where the
30:41good ideas were where the bad ideas were
30:43and they actually went through a whole
30:45bunch of different potential ideas
30:47before eventually deciding to work on
30:49fuel cards and I love the example of a
30:52to B Because A to B is one of the best
30:55new ideas to come out of YC in several
30:57years A to B is a phenomenal company and
31:00the approach that they used to find this
31:01idea is something that really anyone
31:03could do and most Founders don't do this
31:07because it just sounds like too much
31:10work so if you're willing to put in the
31:12work this is an amazing way to find a
31:14startup idea okay and my last recipe is
31:17to just like look for big industries
31:19that seem broken
31:21any Big Industry that seems broken is
31:23probably right for disruption and
31:25finally I've got kind of a bonus recipe
31:26which is to just find a co-founder that
31:30already has an idea they're actually
31:31like a lot of people just on Startup
31:33School co-founder matching right now
31:34that already have an idea and are
31:36looking for a co-founder so if you don't
31:38have a co-founder and you don't have an
31:39idea that could be a great hack to
31:41getting both at the same time the last
31:43point that I want to leave you with is
31:45just to remember that it's often hard to
31:47tell if a startup idea is good or not
31:49and so while I hope the concepts that I
31:52talked about here will help
31:54typically the only way to know for sure
31:57if your startup idea is good is to Just
31:59Launch it and find out so if after all
32:02this you've got a startup idea and
32:04you're still kind of on the fence about
32:05whether it's actually a good idea or not
32:07that is my advice for you just launch it
32:09and find it
32:12[Music]

About this transcript

This page contains the full transcript of How to Get and Evaluate Startup Ideas | Startup School by Y Combinator, generated from the public captions YouTube serves with the video. The transcript has 6,118 words across 902 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.