SamAltman

Lecture1:Welcome,andIdeas,Products,TeamsandExecutionPartI
Welcome
WelcometoCS183B.IamSamAltman,I'mthePresidentofYCombinator.Nineyearsago,Iwasa
Stanfordstudent,andthenIdroppedouttostartacompanyandthenI'vebeenaninvestorforthelast
few.SoYC,we'vebeenteachingpeoplehowtostartstartupsfornineyears.Mostofit'sprettyspecific
tothestartupsbutthirtypercentofitisprettygenerallyapplicable.Andsowethinkwecanteachthat
thirtypercentinthisclass.Andeventhoughthat'sonlythirtypercentofthewaythere,hopefullyitwill
stillbereallyhelpful.

We'vetaughtalotofthisclassatYCandit'sallbeenofftherecord.Andthisisthefirsttimealotof
whatweteachisgoingtobeontherecord.We'veinvitedsomeofourguestspeakerstocomeandgive
thesametalkstheygiveatYC.We'venowfunded725companiesandsowe'reprettysurealotofthis
advicewegiveisprettygood.Wecan'tfundeverystartupyet,butwecanhopefullymakethisadvice
verygenerallyavailable.

I'monlyteachingthree.CountingYCitself,everyguestspeakerhasbeeninvolvedinthecreationofa
billionplusdollarcompany.Sotheadviceshouldn'tbethattheoretical,it'sallbeenpeoplewhohave
doneit.

Alloftheadviceinthisclassisgearedtowardspeoplestartingabusinesswherethegoalishyper
growthandeventuallybuildingaverylargecompany.Muchofitdoesn'tapplyinothercasesandI
wanttowarnpeopleupfront,thatifyoutrytodothesethingsinalotofbigcompaniesornon-
startups,itwon'twork.Itshouldstillbeinteresting,Ireallythinkthatstartupsarethewayofthefuture
andit'sworthtryingtounderstandthem,butstartupsareverydifferentthannormalcompanies.So
overthecourseoftodayandThursday,I'mgoingtotrytogiveanoverviewofthefourareasyouneed
toexcelatinordertomaximizeyoursuccessasastartup.Andthenthroughoutthecourse,theguest
speakersaregoingtodrillintoalloftheseinmoredetail.

Ideas,Products,TeamsandExecutionPartI
Sothefourareas:Youneedagreatidea,agreatproduct,agreatteam,andgreatexecution.These
overlapsomewhat,butI'mgoingtohavetotalkaboutthemsomewhatindividuallytomakeitmake
sense.

Youmaystillfail.Theoutcomeissomethinglikeideaxproductxexecutionxteamxluck,whereluck
isarandomnumberbetweenzeroandtenthousand.Literallythatmuch.Butifyoudoreallywellin
thefourareasyoucancontrol,youhaveagoodchanceatatleastsomeamountofsuccess.

Oneoftheexcitingthingsaboutstartupsisthattheyareasurprisinglyevenplayingfield.Youngand
inexperienced,youcandothis.Oldandexperienced,youcandothis,too.AndoneofthethingsthatI
particularlylikeaboutstartupsisthatsomeofthethingsthatarebadinotherworksituations,like
beingpoorandunknown,areactuallyhugeassetswhenitcomestostartingastartup.

Beforewejumpinonthehow,Iwanttotalkaboutwhyyoushouldstartastartup.I'msomewhat
hesitanttobedoingthisclassatallbecauseyoushouldneverstartastartupjustforthesakeofdoing
so.Therearemucheasierwaystobecomerichandeveryonewhostartsastartupalwayssays,always,
thattheycouldn'thaveimaginedhowhardandpainfulitwasgoingtobe.Youshouldonlystarta
startupifyoufeelcompelledbyaparticularproblemandthatyouthinkstartingacompanyisthebest
waytosolveit.

Thespecificpassionshouldcomefirst,andthestartupsecond.Infact,alloftheclasseswehaveatYC
followthis.Soforthesecondhalfoftoday'slecture,DustinMoskovitzisgoingtotakeoverandtalk
aboutwhytostartastartup.Weweresosurprisedattheamountofattentionthisclassgot,thatwe
wantedtomakesurewespentalotoftimeonthewhy.

Thefirstofthefourareas:agreatidea.It'sbecomepopularinrecentyearstosaythattheideadoesn't
matter.Infact,it'suncooltospendalotoftimethinkingabouttheideaforastartup.You'rejust
supposedtostart,throwstuffatthewall,seewhatsticks,andnotevenspendanytimethinkingabout
ifitwillbevaluableifitworks.

Andpivotsaresupposedtobegreat,themorepivotsthebetter.Sothisisn'ttotallywrong,thingsdo
evolveinwaysyoucan'ttotallypredict.Andthere'salimittohowmuchyoucanfigureoutwithout
actuallygettingaproductinthehandsoftheusers.Andgreatexecutionisatleasttentimesas
importantandahundredtimesharderthanagreatidea.

Butthependulumhasswungwayoutofwhack.Abadideaisstillbadandthepivot-happyworldwe're
intodayfeelssuboptimal.Greatexecutiontowardsaterribleideawillgetyounowhere.Thereare
exceptions,ofcourse,butmostgreatcompaniesstartwithagreatidea,notapivot.

Ifyoulookatsuccessfulpivots,theyalmostalwaysareapivotintosomethingthefoundersthemselves
wanted,notarandommadeupidea.AirbnbhappenedbecauseBrianCheskycouldn'tpayhisrent,but
hehadsomeextraspace.Ingeneralthoughifyoulookatthetrackrecordofpivots,theydon'tbecome
bigcompanies.Imyselfusedtobelieveideasdidn'tmatterthatmuch,butI'mverysurethat'swrong
now.

Thedefinitionoftheidea,aswetalkaboutit,isverybroad.Itincludesthesizeandthegrowthofthe
market,thegrowthstrategyforthecompany,thedefensibilitystrategy,andsoon.Whenyou're
evaluatinganidea,youneedtothinkthroughallthesethings,notjusttheproduct.Ifitworksout,
you'regoingtobeworkingonthisfortenyearssoit'sworthsomerealupfronttimetothinkthrough
theupfrontvalueandthedefensibilityofthebusiness.Eventhoughplansthemselvesareworthless,
theexerciseofplanningisreallyvaluableandtotallymissinginmoststartupstoday.

Long-termthinkingissorareanywhere,butespeciallyinstartups.Thereisahugeadvantageifyoudo
it.Rememberthattheideawillexpandandbecomemoreambitiousasyougo.Youcertainlydon'tneed
tohaveeverythingfiguredoutinyourpathtoworlddomination,butyoureallywantanicekernelto
startwith.Youwantsomethingthatcandevelopininterestingways.

Asyou'rethinkingthroughideas,anotherthingweseethatfoundersgetwrongallthetimeisthat
somedayyouneedtobuildabusinessthatisdifficulttoreplicate.Thisisanimportantpartofagood
idea.

Iwanttomakethispointagainbecauseitissoimportant:theideashouldcomefirstandthestartup
shouldcomesecond.Waittostartastartupuntilyoucomeupwithanideayoufeelcompelledto
explore.Thisisalsothewaytochoosebetweenideas.Ifyouhaveseveralideas,workontheonethat
youthinkaboutmostoftenwhenyou'renottryingtothinkaboutwork.Whatwehearagainandagain
fromfoundersisthattheywishtheyhadwaiteduntiltheycameupwithanideatheyreallyloved.

Anotherwayoflookingatthisisthatthebestcompaniesarealmostalwaysmissionoriented.It's
difficulttogettheamountoffocusthatlargecompaniesneedunlessthecompanyfeelslikeithasan
importantmission.Andit'susuallyreallyhardtogetthatwithoutagreatfoundingidea.Arelated
advantageofmissionorientedideasisthatyouyourselfwillbededicatedtothem.Ittakesyearsand
years,usuallyadecade,tobuildagreatstartup.Ifyoudon'tloveandbelieveinwhatyou'rebuilding,
you'relikelytogiveupatsomepointalongtheway.There'snowayIknowoftogetthroughthepain
ofastartupwithoutthebeliefthatthemissionreallymatters.Alotoffounders,especiallystudents,
believethattheirstartupswillonlytaketwotothreeyearsandthenafterthatthey'llworkonwhat
they'rereallypassionateabout.Thatalmostneverworks.Goodstartupsusuallytaketenyears.

Athirdadvantageofmissionorientedcompaniesisthatpeopleoutsidethecompanyaremorewilling
tohelpyou.You'llgetmoresupportonahard,importantproject,thanaderivativeone.Whenitcomes
tostartingastartup,it'seasiertofoundahardstartupthananeasystartup.Thisisoneofthose
counter-intuitivethingsthattakespeoplealongtimetounderstand.It'sdifficulttooverstatehow
importantbeingmissiondrivenis,soIwanttostateitonelasttime:derivativecompanies,companies
thatcopyanexistingideawithveryfewnewinsights,don'texcitepeopleandtheydon'tcompelthe
teamstoworkhardenoughtobesuccessful.

PaulGrahamisgoingtotalkabouthowtogetstartupideasnextweek.It'ssomethingthatalotof
foundersstrugglewith,butit'ssomethingIbelieveyoucangetbetteratwithpracticeandit'sdefinitely
worthtryingtogetbetterat.

Thehardestpartaboutcomingupwithgreatideas,isthatthebestideasoftenlookterribleatthe
beginning.Thethirteenthsearchengine,andwithoutallthefeaturesofawebportal?Mostpeople
thoughtthatwaspointless.Searchwasdone,andanyways,itdidn'tmatterthatmuch.Portalswere
wherethevaluewasat.Thetenthsocialnetwork,andlimitedonlytocollegestudentswithnomoney?
Alsoterrible.MySpacehaswonandwhowantscollegestudentsascustomers?Orawaytostayon
strangers'couches.Thatjustsoundsterribleallaround.

Theseallsoundedreallybadbuttheyturnedouttobegood.Iftheysoundedreallygood,therewould
betoomanypeopleworkingonthem.AsPeterThielisgoingtodiscussinthefifthclass,youwantan
ideathatturnsintoamonopoly.Butyoucan'tgetamonopolyrightaway.Youhavetofindasmall
marketinwhichyoucangetamonopolyandthenquicklyexpand.Thisiswhysomegreatstartupideas
lookreallybadatthebeginning.It'sgoodifyoucansaysomethinglike,"Today,onlythissmallsubset
ofusersaregoingtousemyproduct,butI'mgoingtogetallofthem,andinthefuture,almost
everyoneisgoingtousemyproduct."

Hereisthethemethatisgoingtocomeupalot:youneedconvictioninyourownbeliefsanda
willingnesstoignoreothers'naysaying.Thehardpartisthatthisisaveryfineline.There'srighton
onesideofit,andcrazyontheother.Butkeepinmindthatifyoudocomeupwithagreatidea,most
peoplearegoingtothinkit'sbad.Youshouldbehappyaboutthat,itmeanstheywon'tcompetewith
you.

That'swhyit'salsonotdangeroustotellpeopleyouridea.Thetrulygoodideasdon'tsoundlikethey're
worthstealing.Youwantanideawhereyoucansay,"Iknowitsoundslikeabadidea,buthere's
specificallywhyit'sactuallyagreatone."Youwanttosoundcrazy,butyouwanttoactuallyberight.
Andyouwantanideathatnotmanyotherpeopleareworkingon.Andit'sokayifitdoesn'tsoundbig
atfirst.

Acommonmistakeamongfounders,especiallyfirsttimefounders,isthattheythinkthefirstversionof
theirproduct-thefirstversionoftheiridea-needstosoundreallybig.Butitdoesn't.Itneedstotake
overasmallspecificmarketandexpandfromthere.That'showmostgreatcompaniesgetstarted.
Unpopularbutrightiswhatyou'regoingfor.Youwantsomethingthatsoundslikeabadidea,butisa
goodidea.

Youalsoreallywanttotakethetimetothinkabouthowthemarketisgoingtoevolve.Youneeda
marketthat'sgoingtobebigin10years.Mostinvestorsareobsessedwiththemarketsizetoday,and
theydon'tthinkatallabouthowthemarketisgoingtoevolve.

Infact,Ithinkthisisoneofthebiggestsystemicmistakesthatinvestorsmake.Theythinkaboutthe
growthofthestart-upitself,theydon'tthinkaboutthegrowthofthemarket.Icaremuchmoreabout
thegrowthrateofthemarketthanitscurrentsize,andIalsocareifthere'sanyreasonit'sgoingtotop
out.Youshouldthinkaboutthis.Iprefertoinvestinacompanythat'sgoingafterasmall,butrapidly
growingmarket,thanabig,butslow-growingmarket.

Oneofthebigadvantagesofthesesortsofmarkets-thesesmaller,rapidlygrowingmarkets-isthat
customersareusuallyprettydesperateforasolution,andthey'llputupwithanimperfect,butrapidly
improvingproduct.Abigadvantageofbeingastudent-oneofthetwobiggestadvantages-isthatyou
probablyhavebetterintuitionaboutwhichmarketsarelikelytostartgrowingrapidlythanolderpeople
do.Anotherthingthatstudentsusuallydon'tunderstand,orittakesawhile,[isthat]youcannotcreate
amarketthatdoesnotexist.Youcanbasicallychangeeverythinginastart-upbutthemarket,soyou
shouldactuallydosomethinkingtobesure-orbeassureasyoucanbe-thatthemarketyou'regoing
afterisgoingtogrowandbethere.

Therearealotofdifferentwaystotalkabouttherightkindofmarket.Forexample,surfingsomeone
else'swave,steppingintoanupelevator,orbeingpartofamovement,butallofthisisjustawayof
sayingthatyouwantamarketthat'sgoingtogrowreallyquickly.Itmayseemsmalltoday,itmaybe
smalltoday,butyouknow-andotherpeopledon't-thatit'sgoingtogrowreallyfast.

Sothinkaboutwherethisishappeningintheworld.Youneedthissortoftailwindtomakeastartup
successful.

Theexcitingthingisthethereareprobablymoreofthesetailwindsnowtheneverbefore.AsMarc
Andreessensays,softwareiseatingtheworld.Itsjusteverywhere,therearesomanygreatideasout
there.Youjusthavetopickone,andfindonethatyoureallycareabout.

Anotherversionofthis,thatgetsdowntothesameidea,isSequoia'sfamousquestion:Whynow?Why
isthistheperfecttimeforthisparticularidea,andtostartthisparticularcompany.Whycouldn'titbe
donetwoyearsago,andwhywilltwoyearsinthefuturebetwolate?Forthemostsuccessfulstartups
we'vebeeninvolvedwith,they'veallhadagreatideaandagreatanswertothisquestion.Andifyou
don'tyoushouldbeatleastsomewhatsuspiciousaboutit.

Ingeneral,itsbestifyou'rebuildingsomethingthatyouyourselfneed.You'llunderstanditmuch
betterthanifyouhavetounderstanditbytalkingtoacustomertobuildtheveryfirstversion.Ifyou
don'tneedityourself,andyou'rebuildingsomethingsomeoneelseneeds,realizethatyou'reatabig
disadvantage,andgetveryveryclosetoyourcustomers.Trytoworkintheiroffice,ifyoucan,andif
not,talktothemmultipletimesaday.

Anothersomewhatcounterintuitivethingaboutgoodstartupideasisthatthey'realmostalwaysvery
easytoexplainandveryeasytounderstand.Ifittakesmorethenasentencetoexplainwhatyou're
doing,that'salmostalwaysasignthatitstoocomplicated.Itshouldbeaclearlyarticulatedvisionwith
asmallnumberofwords.Andthebestideasareusuallyverydifferentfromexistingcompanies,
[either]inoneimportantway,likeGooglebeingasearchenginethatworkedjustreallywell,andnone
oftheotherstuffoftheportals,ortotallynew,likeSpaceX.Anycompanythat'sacloneofsomething
else,thatalreadyexists,withsomesmallormadeupdifferentiatorlikeX,beautifuldesign,orYfor
peoplethatlikeredwineinsteadthatusuallyfails.

SoasImentioned,oneofthegreatthingsaboutbeingastudentisthatyou'vegotaverygood
perspectiveonnewtechnology.Andlearningtohavegoodideastakesawhile,sostartworkingonthat
rightnow.That'sonethingwehearfrompeopleallthetime,thattheywishtheyhaddonemoreofasa
student.

Theotherismeetingpotentialcofounders.Youhavenoideahowgoodofanenvironmentyou'rein
rightnow,formeetingpeopleyoucanstartacompanywithdowntheroad.Andtheonethingthatwe
alwaystellcollegestudentsisthatmoreimportantthenanyparticularstartupisgettingtoknow
potentialcofounders.

SoIwanttofinishthissectionofmytalkwithaquotefrom50Cent.Thisisfromwhenhewasasked
aboutVitaminWater.Iwon'treadit,it'supthere,butit'sabouttheimportanceofthinkingaboutwhat
customerswant,andthinkingaboutthedemandsofthemarket.Mostpeopledon'tdothismost
studentsespeciallydon'tdothis.Ifyoucanjustdothisonething,ifyoucanjustlearntothinkabout
themarketfirst,you'llhaveabigleguponmostpeoplestartingstartups.Andthisisprobablythething
weseewrongwithYCombinatorappsmostfrequently,isthatpeoplehavenotthoughtaboutthe
marketfirst,andwhatpeoplewantfirst.

Soforthenextsection,I'mgoingtotalkaboutbuildingagreatproduct.Andhere,again,I'mgoingto
useaverybroaddefinitionofproduct.Itincludescustomersupport,thecopyyouwriteexplainingthe
product,anythinginvolvedinyourcustomer'sinteractioninwhatyoubuiltforthem.

Tobuildareallygreatcompany,youfirsthavetoturnagreatideaintoagreatproduct.Thisisreally
hard,butitscruciallyimportant,andfortunatelyitsprettyfun.Althoughgreatproductsarealwaysnew
totheworld,anditshardtogiveyouadviceaboutwhattobuild,thereareenoughcommonalitiesthat
wecangiveyoualotofadviceabouthowtobuildit.

Oneofthemostimportanttasksforafounderistomakesurethatthecompanybuildsagreatproduct.
Untilyoubuildagreatproduct,nothingelsematters.Whenreallysuccessfulstartupfounderstellthe
storyoftheirearlydaysitsalmostalwayssittinginfrontofthecomputerworkingontheirproduct,or
talkingtotheircustomers.That'sprettymuchallthetime.Theydoverylittleelse,andyoushouldbe
veryskepticalifyourtimeallocationismuchdifferent.Mostotherproblemsthatfoundersaretryingto
solve,raisingmoney,gettingmorepress,hiring,businessdevelopment,etcetera,theseare
significantlyeasierwhenyouhaveagreatproduct.Itsreallyimportanttotakecareofthatfirst.Step
oneistobuildsomethingthatuserslove.AtYC,wetellfounderstoworkontheirproduct,talkto
users,exercise,eatandsleep,andverylittleelse.AlltheotherstuffIjustmentionedPR,conferences,
recruitingadvisers,doingpartnershipsyoushouldignoreallofthat,andjustbuildaproductandget
itasgoodaspossiblebytalkingtoyourusers.

Yourjobistobuildsomethingthatuserslove.Veryfewcompaniesthatgoontobesupersuccessful
gettherewithoutfirstdoingthis.Alotofgood-on-paperstartupsfailbecausetheymerelymake
somethingthatpeoplelike.Makingsomethingthatpeoplewant,butonlyamediumamount,isagreat
waytofail,andnotunderstandwhyyou'refailing.Sothesearethetwojobs

SomethingthatwesayatYCalotisthatitsbettertobuildsomethingthatasmallnumberofusers
love,thenalargenumberofuserslike.Ofcourse,itwouldbebesttobuildsomethingthatasmall
numberofuserslove,butopportunitiestodothatforv1arerare,andthey'reusuallynotavailableto
startups.Soinpracticeyouendupchoosingthegrayoftheorange.Youmakesomethingthatalotof
userslikealittlebit,orsomethingthatasmallnumberofuserslovealot.Thisisaveryimportant
pieceofadvice.Buildsomethingthatasmallnumberofuserslove.Itismucheasiertoexpandfrom
somethingthatsmallnumberofpeoplelove,tosomethingthatalotofpeoplelove,thenfrom
somethingthatalotofpeopleliketoalotofpeoplelove.Ifyougetright,youcangetalotofother
thingswrong.Ifyoudon'tgetthisright,youcangeteverythingelseright,andyou'llprobablystillfail.
Sowhenyoustartonthestartup,thisistheonlythingyouneedtocareaboutuntilitsworking.

[Audiencemember]:Canyougooverthatslideagain?

Soyouhaveachoiceinastartup.Thebestthingofallworldsistobuildaproductthatalotofpeople
reallylove.Inpractice,youcan'tusuallydothat,becauseifthere'sanopportunitylikethat,Googleor
Facebookwilldoit.Sothere'slikealimittotheareaunderthecurve,ofwhatyoucanbuild.Soyoucan
buildsomethingthatalargenumberofuserslikealittlebit,orasmallnumberofuserslovealot.So
likethetotalamountofloveisthesame,itsjustaquestionofhowitsdistributed.[audiencelaughter]
Andthere'slikethislawofconversationofhowmuchhappinessyoucanputintheworld,withthefirst
productofastartup.

Andsostartupsalwaysstruggle,withwhichofthosetwotheyshouldgo.Andtheyseemequal,right?
Becausetheareaunderthecurveisthesame.Butwe'veseenthistimeandagain,thatthey'renot.And
thatit'ssomucheasiertoexpand,onceyou'vegotsomethingthatsomepeoplelove,youcanexpand
thatintosomethingthatalotofotherpeoplelove.Butifyoustartwithambivalence,orweak
enthusiasm,andtrytoexpandthat,you'llnevergetuptoalotofpeoplelovingit.Sotheadviceis:find
asmallgroupofusers,andmakethemlovewhatyou'redoing

Onewaythatyouknowwhenthisisworking,isthatyou'llgetgrowthbywordofmouth.Ifyouget
somethingpeoplelove,peoplewilltelltheirfriendsaboutit.Thisworksforconsumerproductand
enterpriseproductsaswell.Whenpeoplereallylovesomething,they'lltelltheirfriendsaboutit,and
you'llseeorganicgrowth.

Ifyoufindyourselftalkingabouthowit'sokaythatyou'renotgrowingbecausethere'sabig
partnershipthat'sgoingtocomesaveyouorsomethinglikethatitsalmostalwaysasignofreal
trouble.Salesandmarketingarereallyimportant,andwe'regoingtohavetwoclassesonthemlater.A
greatproductisthesecrettolongtermgrowthhacking.Youshouldgetthatrightbeforeanythingelse.
Itdoesn'tgeteasiertoputoffmakingagreatproduct.Ifyoutrytobuildagrowthmachinebeforeyou
haveaproductthatsomepeoplereallylove,you'realmostcertainlygoingtowasteyourtime.Breakout
companiesalmostalwayshaveaproductthat'ssogood,itgrowsbywordofmouth.Overthelongrun,
greatproductwin.Don'tworryaboutyourcompetitorsraisingalotofmoney,orwhattheymightdoin
thefuture.Theyprobablyaren'tverygoodanyway.Veryfewstartupsdiefromcompetition.Mostdie
becausetheythemselvesfailtomakesomethinguserslove,theyspendtheirtimeonotherthings.So
worryaboutthisaboutallelse.

Anotherpieceofadvicetomakesomethingthatuserslove:startwithsomethingsimple.Itsmuch
mucheasiertomakeagreatproductifyouhavesomethingsimple.Evenifyoureventualplansare
supercomplex,andhopefullytheyare,youcanalmostalwaysstartwithasmallersubsetoftheproblem
thenyouthinkisthesmallest,anditshardtobuildagreatproduct,soyouwanttostartwithaslittle
surfaceareaaspossible.Thinkaboutthereallysuccessfulcompanies,andwhattheystartedwith,think
aboutproductsyoureallylove.They'regenerallyincrediblysimpletouse,andespeciallytogetstarted
using.ThefirstversionofFacebookwasalmostcomicallysimple.ThefirstversionofGooglewasjusta
webpagewithatextboxandtwobuttons;butitreturnedthebestresults,andthat'swhyuserslovedit.
TheiPhoneisfarsimplertousethenanysmartphonethatevercamebeforeit,anditwasthefirstone
usersreallyloved.

Anotherreasonthatsimple'sgoodisbecauseitforcesyoutodoonethingextremelywellandyouhave
todothattomakesomethingthatpeoplelove.

Thewordfanaticalcomesupagainandagainwhenyoulistentosuccessfulfounderstalkabouthow
theythinkabouttheirproduct.Founderstalkaboutbeingfanaticalinhowtheycareaboutthequality
ofthesmalldetails.Fanaticalingettingthecopythattheyusetoexplaintheproductjustright.and
fanaticalinthewaythattheythinkaboutcustomersupport.Infact,onethingthatcorrelateswith
successamongtheYCcompaniesisthefoundersthathookupPagerdutytotheirticketingsystem,so
thateveniftheuseremailsinthemiddleofthenightwhenthefounder'sasleep,theystillgeta
responsewithinanhour.Companiesactuallydothisintheearlydays.Theirfoundersfeelphysicalpain
whentheproductsucksandtheywanttowakeupandfixit.Theydon'tshipcrap,andiftheydo,they
fixitveryveryquickly.Anditdefinitelytakessomeleveloffanaticismtobuildgreatproducts.

Youneedsomeuserstohelpwiththefeedbackcycle,butthewayyoushouldgetthoseusersis
manuallyyoushouldgorecruitthembyhand.Don'tdothingslikebuyGoogleadsintheearlydays,
togetinitialusers.Youdon'tneedverymany,youjustneedonesthatwillgiveyoufeedbackeveryday,
andeventuallyloveyourproduct.SoinsteadoftryingtogetthemonGoogleAdwords,justthefew
people,intheworld,thatwouldbegoodusers.Recruitthembyhand.

BenSilbermann,wheneveryonethoughtPinterestwasajoke,recruitedtheinitialPinterestusersby
chattingupstrangersincoffeeshops.Hereallydid,hejustwalkedaroundPaloAltoandsaid"Willyou
pleaseusemyproduct?"HealsousedtorunaroundtheApplestoreinPaloAlto,andhewouldlikeset
allthebrowserstothePinteresthomepagerealquick,beforetheycaughthimandkickedhimout,
(laughter)andsothatwhenpeoplewalkedintheywerelike"Oh,what'sthis?".Thisisanimportant
exampleofdoingthingsthatdon'tscale.Ifyouhaven'treadPaulGraham'sessayonthattopic,you
definitelyshould.

Sogetusersmanuallyandrememberthatthegoalistogetasmallgroupofthemtoloveyou.
Understandthatgroupextremelywell,getextremelyclosetothem.Listentothemandyou'llalmost
alwaysfindoutthatthey'reverywillingtogiveyoufeedback.Evenifyou'rebuildingtheproductfor
yourself,listentooutsideusers,andthey'lltellyouhowtomakeaproductthey'llpayfor.Dowhatever
youneedtomakethemloveyou,andmakethemknowwhatyou'redoing.Becausethey'llalsobethe
advocatesthathelpyougetyournextusers.

Youwanttobuildanengineinthecompanythattransformsfeedbackfromusersintoproduct
decisions.Thengetitbackinfromoftheusersandrepeat.Askthemwhatthelikeanddon'tlike,and
watchthemuseit.Askthemwhatthey'dpayfor.Askthemifthey'dbereallybummedifyourcompany
wentaway.Askthemwhatwouldmakethemrecommendtheproducttotheirfriends,andaskthemif
they'drecommendedittoanyyet.

Youshouldmakethisfeedbackloopastightaspossible.Ifyourproductgets10percentbetterevery
week,thatcompoundsreallyquickly.Oneoftheadvantagesofsoftwarestartupsisjusthowshortyou
canmakethefeedbackloop.Itcanbemeasuredinhours,andthebestcompaniesusuallyhavethe
tightestfeedbackloop.Youshouldtrytokeepthisgoingforallofyourcompany'slife,butitsreally
importantintheearlydays.

Thegoodnewsisthatallthisisdoable.Itshard,ittakesalotofeffort,butthere'snomagic.Theplan
isatleastisstraightforward,andyouwilleventuallygettoagreatproduct.

Greatfoundersdon'tputanyonebetweenthemselvesandtheirusers.Thefoundersofthesecompanies
dothingslikesalesandcustomersupportthemselvesintheearlydays.Itscriticaltogetthisloop
embeddedintheculture.Infact,aspecificproblemwealwaysseewithStanfordstartups,forsome
reason,isthatthestudentstrytohiresalesandcustomersupportpeoplerightaway,andyou'vegotto
dothisyourself,itstheonlyway.

Youreallyneedtousemetricstokeepyourselfhonestonthis.Itreallyistruethatthecompanywill
buildwhatevertheCEOdecidestomeasure.Ifyou'rebuildinganInternetservice,ignorethingslike
totalregistrationsdon'ttalkaboutthem,don'tletanyoneinthecompanytalkaboutthemandlook
atgrowthandactiveusers,activitylevels,cohortretention,revenue,netpromoterscores,thesethings
thatmatter.Andthenbebrutallyhonestifthey'renotgoingintherightdirection.Startupsliveon
growth,itstheindicatorofagreatproduct.

Sothisaboutwrapsuptheoverviewonbuildingagreatproduct.Iwanttoemphasizeagain,thatifyou
don'tgetthisright,nothingelsewetalkaboutintheclasswillmatter.Youcanbasicallyignore
everythingelseintheclassuntilthisisworkingwell.Onthepositiveside,thisisoneofthemostfun
partsofbuildingastartup.

SoI'mgoingtopausehere,we'llpickbackupwiththerestofthisonThursday,andnowDustinis
goingtotalkaboutwhyyoushouldstartastartup.Thankyouforcoming,Dustin.

WhyToStartAStartup
Butyeah,Samaskedmetotalkaboutwhyyoushouldstartastartup.There'sabunchofcommon
reasonsthatpeoplehave,thatIhearallthetimeforwhyyoumightstartastartup.Itsimportantto
knowwhatreasonisyours,becausesomeofthemonlymakesenseincertaincontexts,someofthem
willactually,like,leadyouastray.YoumayhavebeenmisleadbythewaythatHollywoodorthepress
likestoromanticizeentrepreneurship,soIwanttotrytoilluminatesomeofthosepotentialfallacies,so
youguyscanmakethedecisioninaclearway.AndthenI'lltalkaboutthereasonIlikebestforactually
startingastartup,itsveryrelatedtoalotofwhatSamjusttalkedabout.Butsurprisingly,Idon'tthink
itsthemostcommonreason.Usuallypeoplehaveoneoftheseotherreasons,or,youknow,theyjust
wanttostartacompanyforthesakeofstartingacompany.

Sothe4commonreasons,justtoenumeratethem,areit'sglamorous,you'llgettobetheboss,you'll
haveflexibility,especiallyoveryourschedule,andyou'llhavethechancetohavebiggerimpactand
makemoremoneythenyoumightbyjoiningalaterstagecompany.

Soyouguysareprobablyprettyfamiliarthisconcept,whenIwrotetheMediumpost,whichalotof
youguysreadayearago,Ifeltlikethestoryinthepresswasalittlemoreunbalanced,
entrepreneurshipgotromanticizedquiteabit.ThemovieTheSocialNetworkcameout,ithadalotof
likebadaspectsofwhatitliketobeanentrepreneur,butmainlyitpaintedthispictureoflike,there'sa
lotofpartyingandyoujustkindofmovefromlikeonebrilliantinsighttoanotherbrilliantinsight,and
reallymadeitseemlikethisreallycoolthingtodo.

AndIthinktherealityisjustnotquitesoglamorous,there'sanuglysidetobeinganentrepreneur,and
moreimportantly,whatyou'reactuallyspendingyourtimeonisjustalotofhardwork.Sam
mentionedthis,butyourbasicallyjustsittingatyourdesk,headsdown,focused,answeringcustomer
supportemails,doingsales,figuringouthardengineeringproblems.Soitsreallyimportantthatyougo
inwitheyeswideopen.Andthenitsalsoquitestressful.Thishasbeenapopulartopicinthepress
lately:TheEconomistactuallyranastoryjustlastweekcalled"Entreupeneursanonymous",andshows
afounderlikehidingunderhisdesk,talkingaboutfounderdepression.Sothisisaveryrealthing.Let's
bereal,ifyoustartacompanyitsgoingtobeextremelyhard.

Whyisitsostressful?Soacouplereasons.Oneisyou'vegotalotofresponsibility.Peopleinany
careerhaveafearoffailure,itskindofjustlikeadominantpartofthepartofthepsychology.But
whenyou'reanentrepreneur,youhavefearoffailureonbehalfofyourselfandallofthepeoplewho
decidedtofollowyou.Sothat'sreallystressful.Insomecasespeoplearedependingonyoufortheir
livelihood,evenwhenthat'snottrue,they'vedecidedtodevotethebestyearsoftheirlifetofollowing
you.Soyou'reresponsiblefortheopportunitycostoftheirtime.You'realwaysoncall,ifsomething
comesupmaybenotalwaysat3inthemorning,butforsomestartupsthat'struebutifsomething
importantcomesup,you'regoingtodealwithit.That'skindatheendofthestory,doesn'tmatterif
you'reonvacation,doesn'tmatterifitstheweekend,you'vegottoalwaysbeontheballandbeina
placementallywhereyou'repreparedtodealwiththosethings.Asortofspecialexampleofthiskindof
stressisfundraising.

SoascenefromTheSocialNetwork.Thisisuspartyingandworkingatthesametimesomebody's
sprayingchampagneeverywhereTheSocialNetworkspendsalotoftimepaintingthesescenes.
Mark'snotinthescene,theotherthingtheyspendalltheirtimeonispaintinghimouttobeahuge
jerk.

ThisisanactualscenefromPaloAlto,hespentalotoftimeatthisdesk,headdownandfocused.Mark
wasstillkindaajerksometimes,butinthismorelikefunlovableway,andnotinasociopathic,scorned
loverway.Sothisisjusthimsignalinghisintentiontojustbefocusedandkeepworking,notbesocial.

Sothenthere'sthescenedemonstratingtheinsightmoment,it'skindoflikeoutofABeautifulMind,
theyliterallystolethatscene.Sotheyliketopaintthatsceneandjumptothesemomentsfromother
moments,withpartyinginbetween.Butreallywewerejustatthattablethewholetime.Soifyou
comparethisphoto,Markisintheexactsamepositionbuthe'swearingdifferentclothes,sothisis
definitelyadifferentday.That'swhatit'sactuallylikeinperson.Ijustcoveredthisbullet;thisisthe
EconomistarticleIwastalkingaboutasecondago.

Soanotherformofstressisunwantedmediaattention.Sopartofitbeingglamorousisyougetsome
positivemediaattentionsometimes,it'snicetobeonthecoverofTimeandtobethePersonofthe
Year.It'smaybealittlelessnicetobeonthecoverofPeoplewithoneofyourweddingphotos.It
dependsonwhoyouare,Ireallyhateit,butwhenValleywaganalyzesyourlectureandtearsyouapart,
youdon'twantthat,youdefinitelydon'twantthat.Nobodywantsthat.

OnethingIalmostneverhearpeopletalkaboutisyou'remuchmorecommitted.Soifyou'reata
startupandit'sverystressfulandthingsarenotgoingwell,you'reunhappy,youcanjustleave.Fora
founder,youcanleave,butit'sveryuncoolandprettymuchablackeyefortherestofyourcareer.And
soyoureallyarecommittedfortenyearsifit'sgoingwellandprobablymorelikefiveyearsifit'snot
goingwell.Sothreeyearstofigureoutit'snotgoingwellandthenifyoufindanicelandingforyour
company,anothertwoyearsattheacquiringcompany.Ifyouleavebeforethat,againit'snotonly
goingtoharmyourselffinanciallybutit'sgoingtoharmallyouremployees.Soifyou'reluckyandyou
haveabadstartupidea,youfailquickly,butmostofthetimeit'snotlikethat.

Ishouldsay,I'vehadalotofthisstressinmyownlife,especiallyintheearlyyearsofFacebook,Igot
reallyunhealthy,Iwasn'texercising,Ihadalotofanxietyactuallythroughoutmyback,likealmost
everysixmonths,whenIwastwenty-oneortwenty-two,whichisprettycrazy.Soifyoudostarta
company,beawarethatyou'regoingtodealwiththis.You'regoingtohavetoactuallymanagethis,it's
oneofyourcoreresponsibilities.BenHorowitzlikestosaythenumberoneroleofaCEOismanaging
yourownpsychology,it'sabsolutelytrue,makesureyoudoit.

Anotherreason,especiallyifyou'rehadanotherjobatanothercompany,youstarttodevelopthis
narrative,likethepeoplerunningthiscompanyareidiots,they'remakingallthesedecisionsand
spendingalltheirtimeinthesestupidways,I'mgonnastartacompanyandI'mgoingtodoitbetter.
I'mgoingtosetalltherules.

Soundsgood,makesalotofsense.Ifyou'vereadmymediapost,you'llknowwhat'scoming,I'llgive
youguysasecondtoreadthisquote:

PeoplehavethisvisionofbeingtheCEOofacompanytheystartedandbeingontopofthepyramid.
Somepeoplearemotivatedbythat,butthatsnotatallwhatitslike.

Whatitsreallylike:everyoneelseisyourbossallofyouremployees,customers,partners,users,
mediaareyourboss.Iveneverhadmorebossesandneededtoaccountformorepeopletoday.

ThelifeofmostCEOsisreportingtoeveryoneelse,atleastthatswhatitfeelsliketomeandmost
CEOsIknow.Ifyouwanttoexercisepowerandauthorityoverpeople,jointhemilitaryorgointo
politics.Dontbeanentrepreneur.

-PhilLibin
Thisreallyresonateswithme.Onethingtopointoutisthattherealityofthesedecisionisnuanced.
Thepeopleyouthoughtwereidiotsprobablyweren'tidiots,theyjusthadareallydifficultdecisionin
frontofthemandpeoplepullingtheminmultipledirections.SothemostcommonthingIhaveto
spendmytimeonandmyenergyonasaCEOisdealingwiththeproblemsthatotherpeopleare
bringingtome,theotherprioritiesthatpeoplecreate,andit'susuallyintheformofaconflict.People
wanttogoindifferentdirectionsorcustomerswantdifferentthings.AndImighthavemyown
opinionsonthat,butthegameI'mplayingiswhodoIdisappointtheleastandjusttryingtonavigate
allthesedifficultsituations.

Andevenonadaytodaybasis,ImightcomeinonMondayandhaveallthesegrandplansforhowI'm
goingtoimprovethecompany.Butifanimportantemployeeisthreateningtoquit,that'smynumber
onepriority.That'swhatI'mspendingmytimeon.

AsubsetofYou'retheBossisyouhaveflexibility,youhavecontroloveryourownschedule.Thisisa
reallyattractiveidea.Sohere'sthereality:

Ifyou'regoingtobeanentrepreneur,youwillactuallygetsomeflextimetobehonest.You'llbeableto
workany24hoursadayyouwant!

-PhilLibin
Thistrulyresonateswithmeaswell.Someofthereasonsforthisagain,you'realwaysoncall.Somaybe
youdon'tintendtoworkallpartsoftheday,butyoudon'tcontrolwhichones.

You'rearolemodelofthecompany,andthisissuperimportant.Soyoumighthavesomegoodweeks
andyoumighthavesomebadweeks,someweekswhenyou'relowenergyandyoumightwanttotakea
coupledaysoff.That'sreallybadifyou'reanentrepreneur.Yourteamwillreallysignaloffofwhat
you'rebringingtothetable.Soifyoutakeyourfootoffthegas,sowillthey.

You'realwaysworkinganyways.Ifyou'rereallypassionateaboutanidea,it'sgoingtopullyoutowards
it.Ifyou'reworkingwithgreatinvestors,you'reworkingwithgreatpartners,they'regoingtobe
workingreallyhard,they'regoingtowantyoutobeworkingreallyhard.

Somecompaniesliketotellthestoryaboutyoucanhaveyourcakeandeatittoo,youcanhavelike4
daysworkweeksmaybe,ifyou'reTimFerrismaybeyoucanhavea12hoursworkweek.It'sareally
attractiveideaanditdoesworkinaparticularinstancewhichisifyouwannaactuallyhaveasmall
businesstogoafterineachmarketthenyouareasmallbusinessentrepreneur,thatmakeslittlesense
butassoonasyougetpastlike2or3peopleyoureallyneedtostepitupandbefull-timecommitted.

You'llmakemoremoneyandhavemoreimpact
Thisisthebigone,theoneIhearthemostespeciallylikecandidatesapplyingtoa[?],theytellme
"YouknowI'dreallyliketoworkformuchsmallercompaniesorstartmyownbecausethenIhavea
muchbiggersliceofthepieorhavemuchmoreimpactonhowthatcompanydoesandI'llhavemore
equitysoI'llmakemoremoneyaswell".Solet'sexaminewhenthismightbetrue.

I'llexplainthesetables.They'realittlecomplexbutlet'sfocusontheleftfirst.Thesearejust
explainingDropboxandFacebook,thesearetheircurrentvaluationsandthisishowmuchmoneyyou
mightmakeasemployeenumber100comingintothesecompaniesespeciallyifyou'relikean
experienced,relativelyexperiencedengineer,youhavelike5yearsofindustryexperience,you'repretty
likelytohaveanofferthat'saround10basepoints.IfyoujoinedDropboxcoupleyearsagotheupside
you'vealreadylockedinisabout$10Mandthere'splentymoregrowthfromthere.Ifyoujoined
Facebookacoupleyearsintoitsexistenceyou'vealreadymadearound$200M,thisisahugenumber
andevenifyoujoinedFacebookasemployeenumber1000,soyoujoinedlike2009,youstillmake
$20M,that'sagiantnumberandthat'showyoushouldbebenchmarkingwhenyou'rethinkingabout
whatyoumightmakeasanentrepreneur.

Movingovertothetableontheright,thesearetwotheoreticalcompaniesyoumightstart."Uberfor
PetSitting",prettygoodideaifyou'rereallywellsuitedtothisyoumighthaveareallygoodshotat
buildinga$100Mcompanyandyourshareofthatcompanyislikelytobearound10%;thatcertainly
fluctuatesalot,somefoundershavemorethanthis,somefoundershavealotless,butaftermultiple
roundsofdilution,multipleroundsofoptionpoolcreationyou'reprettylikelytoendupabouthere.If
youhavemorethanthisI'drecommendSam'spostonequitysplitbetweenfoundersandemployees,
youshouldbeprobablygivingoutmore.

Sobasicallyifyou'reextremelyconfidentinbuildinga$100M,whichisabigask,itshouldgowithout
sayingthatyoushouldhavealotmoreconfidenceonFacebookin2009orDropboxin2014thatyou
mightforastartupthatdoesn'tevenexistyet,thenthisisworthdoing.Ifyouhavea$100Mideaand
you'reprettyconfidentyoucanexecuteitI'dconsiderthat.

Ifyouthinkyou'retherightentrepreneurtobuild"UberforSpaceTravel",that'sareallyhugeidea,
$2Bidea,you'reactuallygonnahaveaprettygoodreturnforthat,youshoulddefinitelydothat,thisis
alsothevalueonlyafter4yearsandthisideaprobablyhaslegs,definitelygoafterthat,ifyou're
thinkingofbuildingthatyouprobablyshouldn'tevenbeinthisclassrightnow,justgobuildthat
company.

Sowhyisthisfinancialrewardandimpact?Ireallythinkthatfinancialrewardisverystrongly
correlatedwiththeimpactwehaveontheworld,ifyoudon'tbelievethatlet'stalkthroughsome
specificexamplesandnotthinkabouttheequityatall.

Sowhymightjoiningalatestagecompanyactuallymighthavealotofimpact,yougetthisforce
multiplier:theyhaveanexistingmassofuserbase,ifit'sFacebookit'sabillionusers,ifit'sGoogleit's
abillionusers,theyhaveexistinginfrastructuresyougettobuildon,that'salsoincreasinglytruefora
newstartuplikeAWSandalltheseawesomeindependentserviceproviders,butyouusuallygetsome
micro-proprietarytechnologyandtheymaintainitforyou,it'saprettygreatplacetostart.Andyouget
toworkwithateam,it'llhelpyouleverageyourideasintosomethinggreat.

Socouplespecificexamples,BretTaylorcameintoGoogleasaroundemployeenumber1500andhe
inventedGoogleMaps,that'saproductyouguysprobablyuseeveryday,Iusedittogethereandit's
usedbyhundredsofmillionsofpeoplearoundtheworld.Hedidn'tneedtostartacompanytodothat,
hehappenedtogetabigfinancialreward,butthepointisyetagainmassiveimpact.

MycofounderJustinRosensteinjoinedGooglealittlelaterafterBrett,hewasaPMthereandjustasa
sideprojectheendedupprototypingachatwhichusedtobeastand-aloneapp,integratedinGmail
likeyouseeintheupperrightthereandbeforehedidthatlikeyoucouldn'teventhinkyoucouldchat
overAjaxorchatinthebrowseratallandhejustkindademonstrateditandshowedittohisteamand
madeithappen.Thisisprobablyaproductmostofyouusealmosteveryday.

Perhapsevenmoreimpressively,shortlyafterthatJustinleftandbecameemployeearound250at
FacebookandheledahackatonprojectalongwithpeoplelikeAndrewBosworthand[?]tocreatethe
Likebutton,thisisoneofthemostpopularelementsanywhereontheweb,totallychangedhowpeople
useitandthenagaindidn'tneedtostartacompanytodoitandalmostcertainlywouldhavefailedif
hehadtriedbecausehereallyneededthedistributionofFacebooktomakeitwork.

Soimportanttokeepinmindthecontextforwhatkindofcompanyyou'retryingtostartandlike
whereyouwillactuallybeabletomakeithappen.

Sowhat'sthebestreason?
Samalreadytalkedaboutthisalittlebit,butbasicallyyoucan'tnotdoit.You'resuperpassionateabout
thisidea,you'retherightpersontodoit,you'vegottamakeithappen.Sohowdoesthisbreakdown?

Thisisawordplay,youcan'tnotdoitintwoways.Oneisyou'resopassionateaboutitthatyouhaveto
doitandyou'regoingtodoitanyways.Thisisreallyimportantbecauseyou'llneedthatpassiontoget
throughallofthosehardpartsofbeinganentrepreneurthatwetalkedaboutearlier.You'llalsoneedit
toeffectivelyrecruit,candidatescansmellwhenyoudon'thavepassionandthereareenough
entrepreneursouttherethatdohavepassionsotheymayaswellworkforoneofthose!Sothisistable
stakesforbeinganentrepreneur.Yoursubconsciouscanalsotellwhenyoudon'thavepassionandthat
canbeahugeproblem.

Theotherwaytointerpretthisistheworldneedsyoutodoit.Thisisvalidationthattheideais
important,thatit'sgoingtomaketheworldbetter,sotheworldneedsit.Ifit'snotsomethingthe
worldneeds,godosomethingtheworldneeds.Yourtimeisreallyvaluable,thereareplentyofgood
ideasoutthere,maybeit'snotyourown,maybeit'satanexistingcompany,butyoumayaswellwork
onsomethingthat'sgoingtobegood.

Thesecondwaytointerpretthisisthattheworldneedsyoutodoit.You'reactuallywellsuitedforthis
probleminsomeway.Ifthisisn'ttrue,itmaybeasignthatyourtimeisbetterspentsomewhereelse.
Butbestcasescenarioifthisisn'ttrue,yououtcompetetheteamforwhichitistrueandit'sa
suboptimaloutcomefortheworldandthatdoesn'tfeelverygood.

SodrawingthisbacktomyownexperienceatAsana,JustinandIwerereluctantentrepreneursbefore
wefoundedAsana,wewereworkingatFacebookandwewereworkingonagreatproblem.Wewould
basicallyworkalldaylongonournormalprojectsandthenatnightwewouldkeepworkingonthis
internaltaskmanagerthatwasusedinternallyatthecompanyanditwasjustbecausewewereso
passionateabouttheidea,itwassoclearlyvaluablethatwecouldn'tdoanythingelse.

Andatsomepointwehadtohavethehardconversationofokaywhatdoesitmeanifwedon'tactually
startthiscompany.WecouldseetheimpactitwashavingatFacebook,wewereconvinceditwas
valuabletotheworld.Wewerealsoconvincednooneelsewasgoingtobuildit,theproblemhadbeen
aroundalongtimeandwejustkeptseeingincrementalsolutionstoitandsowebelievedifwedidn't
comeoutwiththesolutionwethoughtwasbest,therewouldbealotofvalueleftonthetable.We
couldn'tstopworkingonitandliterallytheideawasbeatingitselfoutofourchestsandforcingitself
outintotheworld.AndIthinkthat'sreallythefeelingyoushouldbelookingforwhenyoustarta
company,that'showyouknowyouhavetherightidea.

I'llgoaheadandstopthere.I'llputsomerecommendedbooksuphere.