How to plan your MVP

so if i skip to the end here is the tldr of this entire presentation launch something quickly then iterate it's that simple i'm gonna go into so much more detail but if you just want to get the major takeaway right now launch something quickly then iterate so to communicate this idea this is the midwit meme and it's proven to be one of the most helpful tools to advise early stage startups what's great about this meme is that the genius founder and the idiot founder both arrive at the correct answer faster than the midwife the hard-working

smart founder the hard-working smart founder will distract themselves with doing customer surveys and fundraising and t building and building out a massive product roadmap and hiring a pr firm everything that's not building an mvp and launching it they will convince themselves that these actions will have a significant positive impact on trajectory of their company but i promise you that none of these things can have a bigger impact than launching your mvp talking to users and iterating so if you want to be the genius founder here's what you need to do first launch quickly use the

early momentum that you and your teammates have to get your product out into the world second get your first customers oftentimes the process of trying to get your first 10 customers teaches you so much about the problem you're trying to solve whether the hypothesis behind your solution is flawed next talk to those initial customers one-on-one to get feedback try to understand the pain they have why they're trying to use your product to solve that pain and how your product is failing them and i promise you your product will be failing them and last but not

least iterate improve your product over time so this seems so simple but this is the problem this is the line of thinking that goes through founders heads sure sounding and building an mvp sounds like a great idea but wouldn't even be better to make my minimum product something that users would love screw that the product in my head would work even better than that why don't i just build a thing in my head and then finally you know i think i'm a product genius and product genius has built the perfect product on day one for

example steve jobs and the iphone going through this progression feels unique and interesting to many early stage founders but what's the problem is it's extremely common and extremely flawed way of thinking what it does indicate though are two emotions that all of us had when starting our companies and i want to dig into these because this is how you don't fall into this trap the first emotion is fear every single founder believes that if they launch something that people don't love immediately their company will be doomed most founders want to spend a ton of time

building the perfect product upfront out of fear and what's interesting is so many yc founders make decisions out of fear of what happens when things will go wrong my favorite tactic to attack this fear is to ask a founder to detail specifically how their company would die if their mvp turns out to be bad the problem is founders have trouble doing this because almost all of the time it's not true almost every single time a founder believes that their company be dead when their mvp doesn't work out they can always iterate their mvp or build

a new one i call this strategy the boogeyman tactic so when giving advice i like to think of a founder in the in their bed late at night it's dark room and they think they hear something in the closet the typical strategy is to hide under the covers and be fearful until you fall asleep but obviously the best strategy is to get out of bed open the closet and see that there's nothing there so if there are things about your company or about your mvp that you really are afraid man if we do something bad

something bad will happen just take a second to actually write down what you think bad could happen and i bet it'll be like opening up the closet and realizing bogeyman's not there now here's a second emotion that causes founders to avoid building mvps confidence specifically overconfidence what goes through their mind is just like steve jobs 000 people just like steve jobs is building the iphone and that iphone was perfect i know what to build as well and so just let me build it let me just build the perfect thing i call these people fake steve

jobs they believe the hype and the story behind steve jobs but they don't actually understand what really happened the real steve jobs didn't build the perfect iphone on day one mostly we don't remember but the first iphone had no 3g so internet was basically not usable unless you're on wi-fi you couldn't take videos with the first iphone there was no app store so all you were using were apple's crappy native apps it had horrible battery life battery life it had no security so you couldn't use it in workplaces you couldn't copy paste it was expensive

and the screen cracked easily that was the quote-unquote perfect day one iphone so the reality is you aren't steve jobs steve jobs wasn't even this fantasy version of himself and so don't use your confidence to fall into the trap of trying to build something perfect table all right so here is a basic spec for your mvp and then we're going to follow up with some examples from yc companies first it should be very fast to build it should take you weeks not months to get the first version of your product live second it should have

very limited functionality you're actually trying to figure out what is the least that someone needs to get some value and so oftentimes the best way of doing that is like massively under built as opposed to massively overfilled and then you should be targeting a very small set of users so often i see yc founder is trying to get the entire market of all of the users all at once when the reality is they should be just going after the folks who have the most pain and they only need the smallest number of those people to

get started so let's look at the first example this is an early version of airbnb what most people don't realize about airbnb is the following v1 of airbnb had no payments you had to pay the host when you arrived or arrange paying in some other way it had no map view you couldn't see literally where the properties were in a city as a host you had to have an airbed how hilarious is that you actually had to have an airbed if you wanted to rent out your whole house or even a room in your house

you couldn't do it you had to have an airbed and then last it was only for conferences they would shut cities down after the conference would be over because the whole theory of airbnb was we're extra housing when the hotels are already full so that's v1 airbnb pretty crappy this is the first version of twitch so v1 there was only one channel the video was incredibly low resolution we weren't doing streaming video games yet and it was so expensive to stream live video that we almost went out of business within the first three months because

of the cost finally a company that everyone knows stripe back then it was known as dev payments stripe started with no banking deals very few features stripe had so few features that many yc companies that were further along like us couldn't even use it it was really built for the extremely early stage founders and then finally most people don't realize is that while it looked like this amazing api in front the founders were actually doing manual paperwork every night to onboard new customers and get the damn thing run so if these mvps were so bad

how did they become great companies why did people use them what was the core that they figured out well here was a core insight that all of these mvps had in common they were solving a hair on fire problem for their users a hair on fire problem is a common startup term but most people don't really think about it enough so i want to work you through a little bit of a thought experiment here imagine right now your hair was on fire right now while you're watching this let's think about if i was sitting next

to you and i was a founder trying to solve that problem for you right now what is the product you would want at this exact moment i think everyone would say some version of a bucket of water i took that bucket of water put it over my head bam hair is not on fire anymore that's perfect product exciting what if sitting right next to you selling you a product i didn't have a bucket of water what happens if the only thing i had to sell you was a brick what would you do now first a

lot of founders say to themselves well i would you know go find some other solution i would go try to find someone slowing a bucket of water remember your hair is on fire this isn't something that you can just walk around until you find something better you want a solution right now so you know what you would do you would buy my brick and you would beat yourself over the head with the brick to try to smother the fire that's an mvp it solved the problem but it was basically the shittiest possible way to solve

that problem so if you think about those examples i just gave all of their mvps solved a user's hair on fire problem so airbnb two-sided marketplace for the guests in the first version of airbnb their guests actually bought non-refundable multiple thousands of dollars tickets to a design conference in san francisco only to find out that all the hotel rooms were sold out so these guests could either lose thousands of dollars on conference tickets they couldn't use or sleep on an air mattress on a stranger's floor that's what they chose to do for the hosts the

airbnb founders their rent had just got increased by their landlord they didn't have jobs and they had no money in their bank accounts hosting strangers on their floor and charging them money was the thing that prevented them from being evicted from their apartment both sides had a hair and fire problem for justin tv and twitch justin khan one of the founders was the first streamer we served and he was desperate to get famous that was his goal and what's interesting is that that crappy mvp you saw caused justin's face to be on the front page

of the san francisco chronicle within two weeks of us launching and causing them invited to the today show in front of millions of viewers it kind of worked um was it perfect no but he was a lot more famous after justin tv started than before and then finally stripe yc startups 10 years ago literally had to spend thousands of dollars and often execute a three-month biz dev deal so they could simply accept credit cards on their site without stripe these companies would literally go out of business before they could start charging their customers and making

revenue and so that's why they used a crappy very early version of stripe so with everything i've said so far i know some of you will still have some doubts some of you are thinking if my mvp is going to suck anyways isn't it just a waste of time why can't i just survey my users figure out what they want and then build something they love surveys take way less time than mvps but here is a non-obvious truth it is so much easier to understand the user's problem and how to solve that problem when a

user is actually using an early version of your product the issue is that users are so much better at telling you whether your product solves their problem then what set of features would solve their problem it turns out users are not great product people they know the result that they want in their head but they don't have the spec for the product that would solve their problem in their head that's your job even worse users will often talk about problems they have as if their hair is on fire but when you build all the features

they ask for they won't even try to use your product because the problem wasn't really that important to them so don't fall into this trap that just talking to people alone or surveying them is sufficient to figure out what problems they have and how to solve them talking to users with an mvp that is the trick so here are some hacks for building your mvp quickly so you can start learning first time box your spec tell yourselves that we're going to build an mvp within three weeks or within six weeks and we're going to get

the best thing we can do within that specific period of time that will prevent you from just letting feature creep make that turn into three months six months nine months number two write your spec be organized it's so much easier to get work done when you've written down what you want to do why and how you're going to measure it it's so much easier to coordinate with your co-founder when you've written this down the worst case is to give yourself three weeks to build something and waste one of those weeks because you didn't write down

what you wanted to build next cut your spec i guarantee you that the initial mvp spec that you built is not going to fit in the amount of time you've allotted so after you wrote down all those features and things you want to build go out there and cut 30 50 of that and finally don't fall in love with your mvp remember it's just a starting point you should be in love with your customer you should be in love with your customers problem you should recognize that your mvp is a very very early starting point

and you're going to have to change it massively so in summary build the mvp quickly get it in the hands of desperate users folks with their hair on fire talk to them and learn exactly why your mvp doesn't solve their problem and iterate until it does also always remember it's better to have a hundred people who love your product than a hundred thousand who kind of like it remember to focus on delighting a small number of desperate users by actually solving their problem it's most often the case that there will be more people like them

in the world it is very rare to see a yc company who has a hundred extremely happy customers and doesn't end up building a great product for many many more all right that is mvps and i am happy to take some questions