00:00as a slides of loading there there is no
00:02topic that should occupy your minds more
00:05as you build your company then bring it
00:09on the team that's going to make your
00:11company successful as you move forward
00:14Hajin amin from triple byte YC alumnus
00:17are going to talk about building in
00:21today's day and age perhaps the key
00:24portion of that team the engineering
00:27team so please welcome harsh you're
00:30starting yeah harsh alright thanks for
00:39so I'm Hodge I'm one of the cofounders
00:41of trip I along with Arman previously I
00:44used to be a partner at Y Combinator and
00:46part of the inspiration of sighing
00:48triple buy it was noticing how after
00:50graduating y Combinator
00:51and raising the first investment around
00:52everyone's number one problem was hiring
00:54and specifically hiring engineers
00:56because the hardest hiring challenge so
00:59three working on to a bite which is a
01:01hiring marketplace that's used by
01:02engineers to find new places to work
01:05Armen and I have gathered lots of data
01:07on what works well when it comes to
01:10I've personally focused a lot on
01:12spending time with companies helping
01:14them think through their strategies for
01:15finding engineers and obviously getting
01:17the most out of triple buy it and Arman
01:19spent a lot of time thinking through the
01:21details of how do you evaluate an
01:23engineer how to evaluate the engineering
01:25skill and answer the question of are
01:27they a good engineer or not so we're
01:29going to share that and sort of have a
01:30divide and conquer strategy going on
01:32here the four main topics we're going to
01:34cover are where to look for engineers
01:36and when you should start thinking about
01:38using recruiters and that's what I'm
01:40going to start with then we'll talk
01:42about evaluating technical skills which
01:44are Mons gonna cover and then I'll
01:47finish up by talking through the process
01:48of making offers and closing which is
01:51getting people to actually join your
01:52team but before we start on any of that
01:55I want to just issue a warning and make
01:58sure you're well prepared for the fact
01:59that hiring really well and truly sucks
02:01it's like incredibly painful process
02:04because for many reasons which I'll
02:06describe in detail the first is it takes
02:09a lot of time like it takes a lot of
02:11time just to convince someone good to
02:13and have a conversation with you and as
02:16a founder as you know time is a very
02:18scarce resource there's bugs to fix
02:20there's sales customers to close various
02:24things going on and hiring will never
02:27feel like it's your top most urgent
02:29priority and it's very easy to
02:31procrastinate and push it back but if
02:33you do that for too long you won't scale
02:35and you won't grow your startup and
02:37someone else will come along do that and
02:39take the market to hiring involves a lot
02:42of repetitive work so actually as Tyler
02:44was giving his presentation I was
02:46talking to Jeff about this back there
02:48there's a lot of similarities between
02:49sales and hiring and actually there's a
02:51lot of similarities between sales and
02:53hiring and fundraising and buy lots of
02:56things that you do as a start-up founder
02:57large part of it is effectively selling
02:59all of the time and selling as Tyler
03:02pointed out involves a lot of repetitive
03:04work this so hiring will involve lots of
03:07messaging and revolve taking lots of
03:09coffee meetings lots of phone calls lots
03:11of interviews and most of those will
03:14result in the dead end and be a complete
03:15waste of your time but you have to keep
03:17going and finally you will get your
03:20heart broken you will inevitably end up
03:23getting rejected by people you really
03:25wanted to hire who would have been the
03:27perfect fit to help you hit your growth
03:29goals but it turns out they were never
03:31really that serious about leaving their
03:33comfortable job at a big company to join
03:36your exciting but risky startup so be
03:39prepared for all of this and as you're
03:42thinking through building a hiring
03:43process I encourage you to think about
03:46this as a funnel that you're creating
03:48that has three parts to it the top of
03:50the funnel is sourcing and that's
03:51finding people who could be good fit the
03:54second is screening that's answering the
03:56question of do you want to hire this
03:58person or not and the final part of the
04:00funnel is closing making the offer and
04:02getting the offer accepted I'm going to
04:05start by talking through some strategies
04:07for building the top of your hiring
04:08funnel these are the five places I'd
04:12recommend that you look for making
04:14engineering hiring for making engineer
04:15hires I'll talk through the pros and
04:19cons of each of these they are personal
04:20networks hiring marketplaces LinkedIn
04:23github job boards and meetups I'll talk
04:27the most out of each of these and this
04:29list is ranked or it's sorted in order
04:31of where I think you should start out
04:33focusing most of your attention and
04:35energy down to where you focus the least
04:38so we start with personal networks in my
04:42opinion personal networks are the best
04:43place to hire especially when you're
04:45early and making your first few hires by
04:47far and the reason is anytime you're
04:50deciding if you want to hire someone
04:51you're essentially asking yourself two
04:53questions one does this person have the
04:55skills that you that you need for them
04:57to do the job and two can you personally
05:00work effectively with this person when
05:03you're a big company you can mostly
05:06focus on answering just the first
05:08question because they're your large
05:09enough there's enough people there's
05:10enough teams that it's likely that
05:12someone one team somewhere will be able
05:15to work effectively with anyone but when
05:17you're small that's not true whether you
05:20can work effectively with someone or not
05:21is a big determinant of your success and
05:23if you hire the wrong person early on
05:25that could literally be fatal so when
05:28you hire someone that you've worked with
05:29previously or you hire someone that's
05:31worked with the person you trust you do
05:33risk the chances that you won't be able
05:35to work well together which is a big
05:37thing to consider early on so that
05:39probably sounds like somewhat obvious
05:41advice and yet I'm surprised by how
05:44often founders still don't really use
05:47their personal networks effectively when
05:48they're hiring and I think there's two
05:50reasons for this the first is they don't
05:53use a process to exhaustively search
05:55through everyone that could potentially
05:57hire and the second is they don't
05:59actually make the ask that's usually
06:02comes down to being afraid of being
06:04rejected by your friends it's somewhat
06:06easier actually to be rejected by a
06:08complete stranger than it is to ask your
06:10best friend to come join you and they
06:11say I'm not sure the idea is that great
06:14there's there's also like you can also
06:16worry about what happens to the
06:18friendship of the startup doesn't work
06:19out like it's the sort of more that goes
06:22into it when you're talking someone you
06:24know personally that when you're talking
06:25to someone you don't but the truth is
06:28you just have to sort of suck it up and
06:30but you won your startup to be
06:31successful hiring from people you know
06:34it's a tremendously valuable resource
06:36and you just have to make that ask so
06:38I'd recommend you follow a strict
06:41start with just making a list of every
06:44good engineer you know whether you think
06:45they're available or not like that's
06:47actually irrelevant doesn't matter if
06:49they just sold their company for a
06:50billion dollars put them on the list
06:52then ping each and every one person on
06:55that list to meet up and commit to
06:57asking them if they would join if you
06:59have a crazy you think it is however
07:00unlikely commit to making the ask if
07:03they say no they're hesitant so
07:06deviously a little B and say will you at
07:07least come by the office and see what
07:09we're working on and if the office is
07:10your apartment that's totally fine to
07:12you but just keeps pushing until you've
07:14at least shown them something that
07:16you've done keep working on convincing
07:18them if it doesn't work out if they say
07:21no and you get a definitive no thence
07:23then ask them if they were in your
07:25position who would they try and hire
07:27make a list and go out and repeat this
07:30exact same process with them and this
07:32process just never ends like I know
07:34public company startup founders know
07:36long startups pod like probably company
07:37founders who kids still do this on a
07:39daily basis like this is just like a key
07:41thing you have to embed in yourself as a
07:43start-up founder as your company does
07:46scale and grow and the tea and you start
07:48putting two team together you want to
07:50start tapping into the personal networks
07:51of your team and the way I'd recommend
07:54doing this is team events where people
07:56brainstorm potential hires and this is
07:58commonly referred to as a sourcing party
08:01the way I had to recommend going about
08:03this is get everyone together send out a
08:07shared spreadsheet and describe the role
08:11that you're hiring for so if it's an
08:12engineering role sort of described in
08:13detail like who you're looking for who
08:15are examples candidates what are skills
08:18and qualities you're you'd be excited
08:20about and then literally have everyone
08:21spend 30 to 45 minutes going through
08:24their LinkedIn or their Facebook or
08:26whatever right there and then thinking
08:28of everyone that could be a fit and
08:29putting them into the spreadsheet
08:31once that's done a trilobite I'll
08:34actually then personally follow up with
08:35anyone on that list who seems like a
08:37good candidate or not and we've made
08:38several really great hires doing this it
08:40works works really really well and you
08:42can kind of make it like a fun thing to
08:44write so we'll do it at the end of the
08:45week just before our Friday all-hands
08:48food and drink and you can also offer
08:50referral bonuses to your team to
08:52incentivize them to do this so
08:55really make sure that you use the great
08:57you're sticking to an exhaustive process
08:58you're making the ask of people you know
09:00and then as you scale tap into the
09:03personal networks of your team once you
09:07sure that you've exhausted your personal
09:10network for leads the next place I'd
09:11start looking would be hiring
09:12marketplaces hiring marketplaces are
09:16actually relatively new they've become
09:18more popular over the last few years as
09:20has become harder to hire engineers by
09:22using traditional methods like reaching
09:24out on LinkedIn or github and I'm gonna
09:26talk more about that next the way I
09:28think about hiring marketplaces is
09:30actually work a lot like dating sites
09:32the idea is there's engineers who create
09:35profiles companies that create profiles
09:37and both are advertising their best
09:39selves you message each other and you
09:42figure out if it's worth meet up in
09:43person and you know if it all works out
09:46you make a hire the dynamics the
09:49marketplaces those such that the demand
09:52for good engineering talent far exceeds
09:53supply and so typically it's the
09:56companies that are being a lot more
09:58proactive in terms of reaching out first
09:59to the candidates the candidates are
10:01getting multiple of inquiries and the
10:04candidates or the engineers are the ones
10:05that are choosing who they want to speak
10:07to and who they don't a big benefit of
10:10using a marketplace especially in the
10:12early stages is that they can help you
10:15hire very quickly because everyone most
10:17candidates are on the marketplaces are
10:19actively looking for a place to move
10:21right now it's very quick to get on a
10:23phone call with them and start pitching
10:24and if you run a good closing process
10:26you can significantly reduce the amount
10:29of time you'll spend as a founder on
10:31hiring which is obviously great the
10:33downsides though are that they tend to
10:35be quite competitive so engineers are
10:38being reached out to do by multiple
10:40companies at the same time so you'll
10:42have to be very effective at convince
10:44them to join if you want to make hires
10:45and the second is that they can be
10:47expensive so most marketplaces will work
10:50on a fee per hire basis which can be 15
10:52to 20 percent of the first year salary
10:54that's cheaper than a recruiting agency
10:57but still a significant cost if you're
10:58in early-stage startup I'm obviously
11:01biased Aircast requires a hiring
11:03marketplace but I'd say like the three
11:05main ones that come up in conversation
11:07at least when we're pitching customers
11:09be trilobite higher than Vetri you
11:12should try they're all free to use to
11:14get started so you're welcome to try I'd
11:16say provide the way we differentiate
11:18ourselves is essentially by having
11:20better candidates and we measure that by
11:22what percentage of candidates that
11:25companies interview through to provide
11:27do they make an offer to you and that
11:28tends to be twice the rate that they
11:30make through other sources and just as a
11:32general note hiring as a funnel you're
11:36optimizing your funnel so you should pay
11:37attention to what percentage of
11:39candidates are making it through each
11:40step in your funnel when you're early
11:42you won't have that many candidates so
11:44you can't be that scientific about it
11:45but start sort of capturing that data
11:48and just building that into the habits
11:50you have when you're thinking about
11:51hiring the third source that I recommend
11:55you go to and Mehcad is LinkedIn and
11:57github and these are effectively the
12:00biggest online directory of engineers in
12:02the world and most of hiring that's done
12:05at big companies is through teams of
12:07technical recruiters reaching out to
12:09engineers on LinkedIn or github finding
12:12the ones that fit sort of certain
12:13keyword criteria and sudden cold
12:15messages the reason they and they go for
12:18a very high volume approach here so
12:20technical Kreutzer may well sound over
12:22100 messages a day just in the hope that
12:24they get a few replies and a dynamic
12:26that occurred especially over the last
12:28few years is there's more technical
12:30recruiters there's more messages going
12:32out on these platforms so response rates
12:34are dropping for everyone which means
12:36for early-stage startups in particular
12:37it's going to require a lot of your time
12:39sending a lot of messages in order to
12:42get a few interested candidates so
12:45making this work for you requires in my
12:48opinion not playing to a high volume
12:51approach like a big company recruiting
12:52team word and instead spending the time
12:55actually researching the reading the
12:57details of profiles or reading through
12:59someone's Linkedin looking through their
13:01github looking at the details of the
13:03work that they've done and sending a
13:04smaller number of personalized targeted
13:07messages and emphasizing when you send
13:10the message why would someone be a good
13:11fit for your company specifically and
13:13give them clear evidence that you've
13:15read their profile and you're interested
13:17in them as an individual as opposed to
13:19sending a spam message
13:21but that doesn't mean the message has to
13:23be super long I still advise keeping it
13:25short and concise just the key fact is
13:27there's proof that you've read their
13:29profile final note here send emails
13:33instead of messages so if you sign up
13:36for LinkedIn recruiter light which is
13:38about $120 a month that will give you
13:40access to connect to fire it's a chrome
13:42plug-in that makes it really easy to
13:44pull out email addresses from anyone's
13:46LinkedIn profile and you'll consistently
13:48see much higher response rates through
13:50emails and you want messages so
13:51definitely definitely go for that okay
13:54the fourth place I try looking for
13:56engineers would be job postings or job
13:59boards these the two main ones for
14:02startup so stack overflow
14:04jobs and AngelList I haven't included I
14:08can use jobs on there because it's only
14:09available to YC companies but how can
14:13use jobs is somewhat unique and just it
14:14has a particularly high quality of
14:16engineer and I rank it second on this
14:19list actually if it were sort of
14:20stand-alone source of Engineers job
14:23board in general do suffer from a
14:25quantity of a quality problem so what's
14:27good about them is you don't have to
14:29spend a huge amount of time posting to
14:31one of them the downside though is that
14:33the time suck can come in later because
14:34most of the applicants you'll get will
14:36be vastly under qualified and you'll get
14:39a lot of applications from people who
14:40aren't even really software engineers
14:42and it will take a lot of your time
14:43reading through all of these resumes and
14:45applications to find the one or two good
14:47applicants so to maximize your return
14:51and maximize the number of good
14:52applicants you do get I'd recommend
14:54focusing on making your job listings
14:56unique and interesting and bear in mind
14:59that the majority of job descriptions on
15:00the internet are written by someone in a
15:03recruiting or marketing department
15:04that's using corporate boilerplate
15:06language that isn't going to especially
15:09using an appeal to an engineering
15:10audience right so as a start-up founder
15:12you can serve experiment to try and
15:15bring through a bit of your personality
15:17in the job listing so one thing you
15:19might try is writing the first person
15:22about the personal story for why you
15:24started the company why you're excited
15:26about the mission and making it seem
15:28like you get that sort of excitement and
15:30passion across other things you could
15:32try is there anything unique about the
15:34culture is there anything
15:35specific about the technical challenges
15:38or product challenges you're facing so
15:40if you put in more details that again
15:42will stand out because big companies
15:43tend to be very generic and vague when
15:46they're talking about what you actually
15:47get to work on they're the final source
15:50I talk about is actually just like
15:51physically in person meet ups so I don't
15:55think that these are actually going to
15:56be very effective and they're sort of a
15:58long shot like I just the numbers don't
16:00really work out like in-person meetups
16:02don't have that large of a number of
16:05people in attendance and truth is most
16:07of the time people are there for free
16:08food and drink more than trying to
16:10actively find somewhere to work so it's
16:12unlikely that you're gonna find both a
16:14really qualified candidate who's
16:16actively looking to move who's excited
16:18about your company and you also
16:20personally need to be very effective at
16:22talking to strangers and convincing them
16:24of things it all for this to work at all
16:26right so I've accrued it on there
16:28because I do know startup founders who
16:30have had success through meetups but
16:32they are few and far between if you do
16:35try this approach I'd recommend focusing
16:37on technical meetups and by that I mean
16:39the local closure programming group
16:42where people come together and meet up
16:44and bring laptops and work on things is
16:46more likely a better source of engineers
16:48than going to Dreamforce the final thing
16:52you could try here is hosting meetups at
16:54your own your own office right so you
16:57could also combine this with the
16:58personal network hiring and use as an
17:00excuse to have friends come by the
17:02office or have their friends who are
17:03engineers come by the office at a
17:05Tribble bite for example one of our
17:07engineers is a fanatical Emacs user and
17:09he hosts the Bay Area Emacs meets up at
17:11our office and it's sort of a it's a it
17:14hasn't worked for us for hiring but it
17:15is a good way to just meet and build a
17:17network of good engineers that could
17:19come invaluable at some point in the
17:20future so it's worth the definite worth
17:23considering okay now I'm going to talk
17:26about when you should think about using
17:27recruiters truth is there's there's not
17:29really any hard rule on when you should
17:31hire a technical recruiter I've seen
17:34companies of less than ten people hire
17:36one I've seen companies wait until
17:37they're 50 people on plus it's sort of I
17:41treat these are my opinions or when you
17:42should and I treat these as a rule of
17:44thumb right so first I think you
17:47absolutely should wait until you've at
17:49you're first engineer before you
17:50consider bring you on a recruiter in any
17:52capacity the reason for that is actually
17:55I think general startup advice applies
17:57here so one is just in general when
18:00you're hiring it's good for you to do
18:03the job yourself for a bit so you feel
18:05the pain and you understand it and you
18:07sort of understand the details of what
18:08makes someone good at that role at your
18:10startup in particular before you go and
18:12hire someone you'll be better able to
18:14assess them so if you feel the pain of
18:15recruiting yourself I think you have a
18:17better shot of hiring the right
18:18recruiter for you and the second the
18:23second reason I say is just it's it's
18:25like sales like it's actually good for
18:27you to go out and try and pitch and
18:28convince people to join your startup so
18:30you can just understand yourself like
18:32what message works and what doesn't
18:34because as a startup founder you are
18:35always selling and you never know when
18:37you might bump into someone who could be
18:39a really great hire and if you've
18:41already sort of practiced the pitch for
18:42convincing engineers to join you and
18:44you'll have those ready to go so I'd
18:46recommend making sure you always make
18:47your first hire before you try and
18:49delegate this to a technical recruiter
18:51second I'd expect to have a good hiring
18:54cadence so somewhere around hiring an
18:55engineer a month for the next six months
18:57or so and before you start bringing on a
18:59recruiter otherwise it's likely they'll
19:01run out of work to do a quickly and then
19:04finally as a rule firm if you're
19:06spending more than 50% of your time
19:08sourcing that's like all the things I
19:10mentioned before and doing initial phone
19:12calls and screens and getting people to
19:14come and meet you in person over 50
19:16percent is probably about the time where
19:18you want to start thinking about
19:19bringing on help because 50% is about
19:22the amount of time you want to be
19:23spending on hiring so recruiters
19:27themselves come in roughly three types
19:28you have contract recruiters who you pay
19:30by the hour and they can do anything
19:33from just : messaging on LinkedIn all
19:35the way to doing initial phone screens
19:37this in-house recruiters it's just
19:39hiring a full time technical recruiter
19:40and that works as a member of your team
19:43and the final would be agency and
19:45agencies are essentially teams of
19:49salespeople that are paying lots of
19:51engineers on LinkedIn or wherever they
19:53and then sending their resumes out to as
19:54many companies as they as possible and
19:56they tend to charge 25 to 30 percent or
19:59first year salary if you hire an
20:02maybe they sent you they do tend to be
20:04fairly high touch so they're quite
20:06involved in trying to give you
20:07information that will help you close an
20:09engineer but they're also sending that
20:10information to multiple companies at the
20:12same time so my recommendations here
20:15would be when you get to the point where
20:17you feel it's time to bring on help or
20:19bring on a recruiter start with a
20:21contract recruiter and have them focus
20:24on just the sourcing piece of it so have
20:26them just focus on reaching out to
20:28engineers on LinkedIn and github and
20:29their main deliverable for you should be
20:32filling your calendar with calls with
20:34promising candidates then you're doing
20:36the pitching and practicing or yielding
20:38the pitching and convincing when you get
20:40to a point where that becomes too much
20:41work for just you then I'd consider
20:44hiring a full-time in-house recruiter
20:45and training them to do the pitch cool
20:48so now they're both sourcing and doing
20:50the initial calls and sort of setting up
20:52the on-site interview process for you so
20:57to sum up most of startup hiring plan if
20:59I were just getting going and building
21:00an engineering team would be start by
21:02making sure you've exhausted your
21:03personal Network spend lots of time
21:05taking people out for lunch coffee
21:08making the ask to you experiment with
21:11the hiring marketplaces your mileage
21:14will vary on these depending on how
21:16effectively you can pitch your company
21:17but even if you don't make hires from
21:20them you'll get valuable experience
21:22pitching real engineers and real
21:23candidates and learning what resonates
21:25about your company to that audience
21:28third spend some amount of time doing
21:30personalized and targeted outreach to
21:32engineers on LinkedIn and github making
21:34those messages really personalized and
21:35finally treat job boards and meetups
21:38into meeting people in person as a
21:40background process you run where you're
21:43not expecting to make any highs from
21:45them but you're sort of building a
21:46general pipeline that could be valuable
21:48in the future cool that's the first part
21:53of this almonds are now going to talk
21:54about how to screen and evaluate a
21:56technical school of engineers then I'll
21:58jump back in a wrap-up with making
22:03awesome thank you hard so I'm Aiman I'm
22:08hard as co-founder of a triple byte and
22:10before this I started Socialcam with
22:13Michael Seibel and I was an early
22:15employee at twitch and I'm gonna talk
22:17about just the screening step so how to
22:20identify skilled engineers at your
22:23company so the first question here is
22:26just you know why you should believe me
22:28about any of this and one answer is that
22:32I've done a lot of interviews so since
22:35starting trilobite I've interviewed over
22:37a thousand engineers personally I but I
22:41think a better answer is that you know
22:44trilobite has a pretty special vantage
22:47point we're able to you know see how the
22:51views do see how candidates rather do
22:53you know in the interviews at multiple
22:55different companies are we can see how
22:57the same candidate performs in multiple
22:59companies and that gives us a data set
23:02that I think no one else has and you
23:06know that data is is you know where the
23:09advice I'm gonna give today comes from
23:13before I jump in I want to just go over
23:16the the basic hiring process that most
23:19companies use and this is actually
23:21pretty standardized so probably 95% of
23:24tech companies use these basic steps to
23:26screen campaigns so first is a resume
23:29screen someone applies to a company they
23:32send in a resume any recruiter looks
23:34over the resume and decides if this
23:35looks like someone who's a basically
23:37good fit then a recruiter call so this
23:41is a usually a 30-minute phone call with
23:43the recruiter I just ask about the
23:45candidate background judge culture fit
23:47and make sure they're interested in the
23:49company then a technical phone screen
23:53and so this would be between 30 minutes
23:55and an hour with an engineer usually
23:59solving a single programming problem so
24:01this is sometimes something like
24:02fizzbuzz or a little bit harder often
24:05this point done over over a synchronized
24:07text pad of some sort then this is
24:11optional sometimes I take home project
24:13so a substantial project
24:16the candidate completes on their own
24:17time and then you know sends in to the
24:20company to be evaluated then finally the
24:23on-site interview so it can they comes
24:26into the office and does between three
24:29and six one-hour sessions with engineers
24:33at the company I you know covering
24:35covering individual problems and then
24:38finally a decision meeting so you so
24:39usually the next day after the candidate
24:41has gone home everyone who interviewed
24:43the candidate and the hiring manager
24:44gets together in a room and talks about
24:47sort of their perceptions and make you
24:50know they they as a group make a hire no
24:51hire decision some stats on this
24:56companies make offers to between two and
25:00eight percent of all of the engineers
25:02who apply but interestingly exactly
25:05where in the process that drop-off
25:07happens is is very different between
25:09companies so we so we see companies
25:11where you know seven five percent of
25:14people who apply get screened out you
25:16know at the first step on a culture fit
25:18call and then we see companies where
25:20almost all applicants make it through to
25:22the final interview and that's where all
25:23this reading happens and about ninety
25:25five percent of people who are hired
25:27work out that is about five percent of
25:29technical hires result in someone being
25:30being fired within a few months and then
25:34from the candidates side what we see is
25:37a distribution of success in interviews
25:40so the the the top you know few
25:42percentage points of programmers by
25:44skill receive job offers after most of
25:47their interviews but then the bulk of
25:49programmers are somewhere in the middle
25:50where they receive you know the you know
25:53they do interviews and they receive
25:54offers after somewhere between you know
25:55fifteen to thirty percent of the
25:56interviewers they do and an interesting
25:59point is that no one passes all their
26:02interviews so there are not magical
26:04engineers who receive offers after every
26:05interview they do and yeah this gets at
26:10what I think is the the the major
26:14challenge when designing any of you
26:15process and this is inconsistency so
26:19this is noise in the interview process
26:22it's kind of the idea you know is your
26:25interview fundamentally repeatable and
26:30you know if you had to if you could
26:32somehow you know reading to view your
26:34co-workers your employees or everyone
26:37you who passed review in the last year
26:38if you could be any of those people how
26:41many of them would pass again and it's a
26:45pretty scary question but kind of
26:49interesting thing is that we've able to
26:50add some data on this that's what I did
26:53was I calculated a stat called the
26:57inter-rater reliability between all of
27:00the interviewers at all of the companies
27:02on trilobite and so what that means is
27:05this this is the statistical measure of
27:06the extent to which different
27:09interviewers tend to agree about which
27:12candidates are best and it's on a scale
27:16of zero to one where zero would be no
27:20agreement or the amount of agreement you
27:22would expect to see you know in random
27:23data from chance alone and one would be
27:26perfect agreement and what I found was
27:29an agreement of just over 0.1 so so the
27:35first point is that that's obviously
27:36much closer to to no agreement than it
27:38is to perfect agreement but for some
27:40context than that I calculated the same
27:43stat on a data set of online movie
27:47reviews and what I got was a agreement
27:52of very similar but it's actually
27:55slightly higher so it ends up that you
27:59know interviewers agree about which
28:02engineers are the best at about the same
28:04rate that Netflix viewers agree about
28:07which movies are best and that's yeah
28:11that's scary I interviews are
28:14fundamentally noisy and they are they
28:16are more noisy than the data shows them
28:17to be more noisy than most hiring
28:20managers sort of want to believe so I
28:23know the first the first question here
28:25is sort of why do invested all that
28:26infinitives are noisy why do them at all
28:27why can't we just use something like
28:31atashi engineers and yeah I think this
28:35is actually a great idea it is almost
28:38certainly much more you know if you can
28:40work with someone for a week you can
28:41almost certainly get a much
28:43sense of if they're a good employee then
28:45and then you could in in a you know if
28:46we are interview with them the problem
28:50is that most engineers don't actually
28:53want to do trial periods so we did some
28:55research on their sexual bite and it
28:57ends up that only 20% of engineers are
28:59willing to to do trial periods and
29:02there's actually some there's some
29:04adverse selection there so the you know
29:06most of the best programmers are in the
29:0880% who would prefer to do a standard
29:11technical interview because it takes
29:12less their time and is faster so I think
29:15that trial employment is an excellent
29:18option and you can offer that as an
29:19option but you know if you don't want to
29:22scare away most good programmers you do
29:23have to still run a standard traditional
29:25interview process so we're gonna talk
29:29about sort of today is you know specific
29:32pieces of advice that you can use to try
29:34to reduce the noise in a traditional
29:36interview and I'm gonna go over seven
29:38points so point one the first way to
29:42reduce the noise in an interview is to
29:45decide what skills matter for your
29:49company there are a lot of different
29:51ways that a programmer can be skilled so
29:55you know someone for example can be very
29:58productive or they can be you slow
30:00careful write great tests and and make
30:02sure they don't commit bugs you know
30:04someone can be strong in math and
30:07computer science or they could be you
30:09know deep knowledge about the internals
30:11of the Linux kernel and and you know
30:13scheduling and real-time operating
30:14systems or something and if you don't
30:17decide as a founder which of those
30:19skills matter for your company then your
30:22interviewers are going to decide that
30:23for you so they're going to come up with
30:26questions that they asked you in the
30:27interview process and they will fail
30:30people who answer poorly on those
30:31questions whether or not those are areas
30:34that that that should matter for your
30:36company and this is actually a major
30:39source of noise in interviews every you
30:43know engineer has this bias where they
30:44think the things that they know the best
30:45are the most important skills one could
30:47know an absent specific direction from
30:50above about what to look for they will
30:52fail people for areas that are not
30:53important for your company
30:56first piece of advice here is that you
30:59should go through and ask yourself these
31:02questions before you start hiring us the
31:05first is I need to clarify these are
31:06these are the axes along which we see
31:08most disagreement between interviewers
31:10the first question is you know do you
31:13need sort of fast iterative programmers
31:15or someone who's careful and rigorous do
31:18you want to look hire someone who's
31:19motivated by solving hard tech problems
31:21or someone who is motivated by building
31:24you know beautiful products for users I
31:26important that someone comes in with
31:29skill in a particular technology pick
31:31your language say or can you hide you
31:34want hire a smart person and let them
31:35learn your tech stack on the job is
31:38academic computer science or algorithm
31:41ability something which is important for
31:43you or is that an irrelevant skill and
31:45then is there any other sort of specific
31:48expertise that you need in people you're
31:52hiring it's actually so you know it's
31:54fine to answer both some of these
31:56questions you don't have to specify only
31:58a single profile of person you're hiring
31:59the important thing is to decide what
32:03matters even if that's multiple profiles
32:04um that sets you up to design the rest
32:07of your process and make sure that
32:08you're not you know failing people for
32:09being bad in relevant areas okay point
32:12number two second way you can reduce
32:14noise in interviews is to use structured
32:17interviews so to define this a free-form
32:22unstructured interview is an interview
32:23where a new viewer gets in a room with a
32:26candidate and they ask questions and
32:28they follow their intuition they they
32:31you know based on the answer the critic
32:32a that gives they ask follow-ups and
32:33they try to get a sense what the
32:35candidate feels like if that person can
32:36be a good fit at the company and at the
32:38end to make a global yes/no I don't know
32:40how a decision based on that Daddy
32:42interview and that contrasts with a
32:44structured interview where the
32:45interviewer comes into the room with a
32:47question which they're going to ask and
32:49a defined criteria they're trying to
32:52evaluate and the very interesting point
32:55here is that everyone thinks that that
32:57freeform interviewers feel better almost
33:00all interviewers prefer freeform
33:02interviews and they think they are more
33:04accurate in many candidates if you ask
33:06them actually say they prefer to be
33:08entered in that way but the
33:10things that this is completely opposite
33:11to all of the available data on this
33:13structure knickers are simply more
33:15predictive they are better at predicting
33:17success on the job and there's no excuse
33:20not to use them we should all be using
33:21structure interviews and so in the purse
33:24exams what that means um the first thing
33:26is that you have to hold the process
33:27constant so the goal of an interview is
33:31essentially to evaluate variants and
33:33candidates and you know if you do not
33:35hold the rest of your process constant
33:37between candidates you're introducing
33:39noise so there's simply no excuse for
33:41not asking every candidate for the same
33:43job the exact same set of questions I
33:46think the reason this is not more common
33:49is that the interviewers themselves tend
33:51to find this boring and you know I can
33:53say it suck it up like hey you like it
33:54you have to do that so second point on
33:58structure interview is that you want to
34:00give your interviewers defined criteria
34:03to evaluate so rather than putting in
34:05the room and saying decide if this K
34:07this person to be a good fit for the
34:08company say we care about coding
34:10productivity and you know knowledge of
34:13back-end web systems and so your goal is
34:15to ask this question the interview and
34:17grade the candidate on coding pregnat
34:19ivities and knowledge of back-end web
34:21and there's actually some great research
34:23on this it ends up that a lot of the
34:26sort of worst biases that can come up in
34:29interviews are made worse when the
34:31interviewer is trying to make a global
34:33decision so if something's making a
34:35global - this person feel like a good
34:36employee decision that's where things
34:38like you know do they look like someone
34:40got you in the past or you know their
34:42race their gender that's what those
34:44things come into play and interviewers
34:45are much better at ignoring those
34:47criteria ignoring those attributes when
34:49they're given a defined criteria to
34:51evaluate and then final point is that
34:55you want a unified decision-making on
34:57this and this applies posted to larger
34:59companies but you want to make sure that
35:00one person or one group of people is
35:02involved of all the final decisions so
35:06rather than viewing interviewers as
35:07people making decisions view them as you
35:09know people taking notes grading
35:11candidates on criteria and that all
35:13those notes and criteria go to a
35:14centralized person or a centralized
35:16group who makes the final decision and
35:18the idea is just that again the goal is
35:20consistency and it's far more easy to be
35:21consistent if one person
35:23making all the decisions point three for
35:25how to reduce noise is that it's just
35:27used better interview questions and so I
35:29have some tips on that
35:31the first is that you want to avoid
35:33questions that require a leap of insight
35:37from the candidate and rather you are
35:40questions where there is a gradual sort
35:42of ramp of steps that they can follow
35:43that lead to a solution and this is the
35:47general idea the rule of thumb I like is
35:49that you can ask yourself can this
35:51question be given away so if it's a
35:53question that has a sort of a single
35:55fact in the solution which a candidate
35:58could know in advance and perhaps be
36:00communicate them by a friend or by
36:01anything they read in Glassdoor then
36:03that means that it's probably a bad
36:05question and the example of that is
36:07there's a plastic question the question
36:09is you know imagine you're staying at
36:11the bottom of a flight of stairs and
36:13every step you can either take a single
36:15step up one stair or a double step up to
36:16stairs you know how many unique paths
36:18are there at the top of the flight of
36:19stairs and you know it ends up the
36:22solution of that is the Fibonacci
36:23kind of strangely if someone knows that
36:26obviously it's you know that's the
36:28solution if they don't know it they may
36:30well struggle then think about it and go
36:31down some some rabbit hole and so that's
36:34exactly what is the solution example of
36:38a good question is you know can you
36:40please implement the game Connect four
36:42for me so there you know it's a series
36:44of steps each one relatively
36:46straightforward but that leads to a
36:48solution and and there's nothing F
36:50summons friend could tell them in ten
36:51minutes which would give them no
36:53massively unfair advantage you know in a
36:55game Connect four so another idea here
36:58this is the kind of related but you want
37:00multi-step problems and there's tend to
37:04lead to problems I don't have insight
37:06but also candidates will often get stuck
37:09in an interview even good candidates and
37:11if your problem has multiple steps to it
37:13you can you can get them a hint you can
37:16help them out in one portion and it
37:18still have enough left for them to go on
37:20to do well and be themselves and
37:21demonstrate skill whereas if your
37:23problem is all in one one one one one
37:25sort of nugget of difficulty if they
37:26can't solve that you have to help them
37:27at that point they basically failed or
37:29done poorly on that section you want to
37:33avoid specialized knowledge
37:36so this is you know if your goal is to
37:38assess general program ability you
37:40probably want to ask questions that
37:41evolve you know things like list and
37:43hash tables and strings rather than
37:46questions that evolve you know tries or
37:48prefix trees you know this list of court
37:51you know unless you've decided that you
37:53really care about algorithm
37:54you know structure knowledge and your
37:55goal is to make sure that me knows about
37:56tries in that case it's totally fine to
37:57ask about them but if you're if you're
37:59measuring something else if you if a try
38:01the portion of the problem you know some
38:03portion of your your people your your
38:05candidates will be familiar with us
38:07others well in that will introduce noise
38:08so in general I think it's good to stick
38:10with the sort of classic most basic CS
38:13concepts another route you want you run
38:16a budget about three times the amount of
38:18time it takes you to solve a problem for
38:20the candidate so if you come up with a
38:22question and you solve it yourself it
38:23takes you ten minutes that's probably a
38:26good question for a 30 minute interview
38:28session and the reason here is just that
38:31it's it's far easier to be the
38:33interviewer that is to be the candidate
38:35it's far easier to ask the question and
38:36we tend to downplay how hard the
38:40questions actually are and we can
38:42actually research on this we went
38:43through all the questions we asked we
38:45asked a terabyte and we looked which are
38:47most correlated with candidates going on
38:49to succeed and it ends up that the most
38:51effective questions tend to be much
38:54easier than what you know our intuition
38:56going in was predicting so it ends up
38:58that most interviewers think that you
39:01know the optimal question is quite a bit
39:03harder than the optimal question
39:04actually is you wanna make sure that you
39:08four or more questions in an interview
39:10they hear is that you know each
39:13individual question carries a certain
39:14amount of noise as the person seems
39:15questioned before is it you know did
39:18they have whatever they lucky answering
39:19it if you ask more questions you'll get
39:21a a more consistent signal out and then
39:26finally just one one tip type of
39:29question we like a lot a trilobite is a
39:30question where we we give the candidate
39:32a problem and then rather than wanting
39:34them to devise a solution we tell them a
39:37solution to the problem and then their
39:38goal is to take that idea and implement
39:41that in the code and so we give them an
39:44algorithm and we see if they can
39:44implement that algorithm rather than
39:46require them to devise an algorithm
39:49a fourth way to reduce noise is to
39:54ignore credentials during your interview
39:57so by credential I mean things like I
40:01you know did someone study at a
40:03well-known school or has someone worked
40:04at a well-known company and I'm not
40:08claiming that credentials are meeting
40:10are meaningless credentials are
40:11important right so the productivity case
40:13the group of people who you know our ex
40:16new colors are indeed better engineers
40:18the group of people who have not worked
40:20at Google I'm so it's totally you know
40:22legitimate to take that fact into
40:25account when when when deciding about
40:27the hire or someone however it's not
40:30relevant to their actual programming
40:32skill so my advice here is to make sure
40:35that you're not biased by the fact that
40:37someone comes from Google when you are
40:39narrowly evaluating the programming
40:40skill so I recommend you hide that
40:43credential from the interviewers and the
40:45reason is we found that interviewers are
40:47actually pretty biased by this if they
40:49know someone has you know what were your
40:52strong credentials they're more likely
40:53to interpret the result of the coding
40:55screen as in a positive fashion or this
40:57person you know yeah they didn't know
41:00that answer but I'm sure they you know
41:01I'm sure there was a temporary slip-up
41:02and so how do pencils from your
41:04interviewers let them assess program
41:06ability and then when making the final
41:08decision in the decision meeting
41:09consider credentials and also the
41:11performance in the interview and yeah I
41:13this will help you find the programmers
41:15who are skilled and who lack traditional
41:17credentials and those are the
41:19undervalued people in the market and as
41:21a start-up if you can you know get if
41:23you're good at finding those people that
41:25is a big advantage okay point five is
41:28that you want to think about the false
41:30negative rate in your interview so a
41:33false negative is when someone fails
41:37your interview who could have gone on to
41:39do the job well and the opposite of
41:40false positive is when you have someone
41:42past if you hire someone who then goes
41:44on to do poorly and probably be fired
41:46and both false positives false negatives
41:49are very costly so if you hire a bad
41:52person if the fire of them that's
41:53terrible it'll hums the morale for the
41:54team and also also very expensive its
41:55actual money big problem but you know if
41:59you're a start-up and you're really
42:00hungry you know you're held back but not
42:02engineers if you pass up a person who
42:04could have joined your team and couldn't
42:05productive that's also very expensive
42:07and I'm making a pretty subtle point
42:10here but I think there's a bit of a
42:11cognitive bias where false positives are
42:16very we're very aware of them
42:17we're very like we hire a bad person we
42:19feel that pain for a month or a month
42:21plus best-case and we've seen comes into
42:25a bite they're generally they generally
42:26are have too much faith that the folks
42:29who fail their process must have been
42:30bad engineers and couldn't have done the
42:31job and that's empirically not the case
42:34right if you watch you know as I said
42:36earlier no engineer will pass all
42:38interviews right so a significant
42:40portion of people who fail interviews do
42:41go on to be employed very productively
42:43at other companies and so I just
42:45recommend that that folks designing any
42:47of you processes think about the false
42:49negative rate and try to give that some
42:51way in in in their calculations of
42:53setting studying hiring bars um one of
42:57my goals would apply to actually just
42:58get to a point where you can actually
42:59measure this rate because cuz no one
43:01knows what it is no one knows what the
43:02false negative rate on the interview is
43:03because to measure that you have to just
43:04hire people randomly and see how they
43:05perform that's very expensive
43:07my goal is to get to where we can do
43:09that let's see okay point six is that
43:15you want to genuinely calibrate on the
43:18maximum skill that each candidate brings
43:20rather than their average skill or their
43:22minimum skill so someone comes in an
43:24interview they're very strong in one
43:26area or the weaker and others what
43:28matters the most is what they were
43:30strongest in everyone can look stupid
43:33right everybody you know if you ask me
43:35the right question I will definitely
43:36look very stupid that's true for
43:37everyone out there and so the problem is
43:39that someone might go through an
43:40interview and do very well on some
43:42things that would be you know useful for
43:43the company but look stupid on one
43:45question and you know if that
43:47interviewer gives a blocking know that
43:49that person was stupid that's
43:50interesting noise in the process perhaps
43:52now again if someone does poorly in a
43:54crack in the area that is important for
43:56the job totally failed them but be open
43:58to the fact that everyone does look
44:00stupid sometimes and don't fail someone
44:02just because they look stupid on one
44:03portion of the interview okay a final
44:05last point here is that you want to
44:08think about the candidate experience
44:09when designing an interview you know you
44:12want to you know make sure that every
44:14candidate who goes through
44:15I likes you likes your company and this
44:20is true for a few reasons one is just
44:21you know if they enjoy the process they
44:23have a higher probability of accepting
44:24an offer you make and so this will help
44:26in the closing step but a interesting
44:28point is that it will also actually make
44:30your screening itself more accurate
44:32because stress has a big impact on
44:35performance high percentage of
44:37candidates get very stressed in
44:38interviews and underperform their actual
44:40peak ability and so some tips here to
44:43help you do stress include just letting
44:46everyone bring in their lap and work in
44:47their own environment
44:48you know what throw and throw but their
44:49own language their own tools they will
44:50be they will be able you're much more
44:51productive much less stressed and then
44:54coaching your interviewers in just some
44:57soft skills so being friendly
45:00providing breaks for the candidates and
45:03you know when they are doing poorly on a
45:05section training them and how to
45:07intercede anyway which isn't too
45:08stressful insulting to the candidate
45:10rule of thumb here that comes from the
45:13the old Jolan software blog is that you
45:16you you want every candidate no matter
45:18how they do to finish your interview
45:20wishing that they had you know wanting
45:22join your company you know even people
45:23who fail your any very poorly you want
45:25them to end your interview wanting to
45:28join you being signs about the
45:29opportunity and the final point I you
45:32want to avoid hazing so this is this is
45:34rare but results in some of the worst
45:36horror stories of interviews this is
45:37where you know I mean who takes on the
45:38role of you know a ritual of acceptance
45:42into a group you know think if this if
45:44this happens it's terrible and as a
45:46hiring manager you can just stay totally
45:47away from that okay those are my points
45:52um I want to emphasize that I'm not
45:56saying that you should lower the bar for
45:58who you hire I think if you follow this
46:01advice you will get the more accurate
46:03signal you can then set your barbecue
46:05higher RF you want on that signal i but
46:09you'll still be making at that point
46:10better hires so just to summarize here i
46:13recommend that the first step you follow
46:16is to decide what skills matter for your
46:18organization and make sure that you're
46:20screening on things that you actually
46:21care about and then design a structured
46:24interview around those skills so come up
46:27with ways to assess each skill
46:29and stretch the criteria for the
46:31interviewers so they are you know less
46:32likely to be biased by by outside
46:35factors then you want to use good
46:38interview questions you know multiple
46:40parts in the lives of insight you want
46:43to hide credentials from the technical
46:44interest because that interest is noise
46:46in the process you want to think about
46:48the false negative as well as the false
46:49to positive cost in your interviewers
46:51and you want to genuinely calibrate
46:53around the maximum skill that each
46:55candidate brings well doing all that you
46:58want to try to provide a positive
46:59experience for the candidate so those
47:03are those are my my tips so far I think
47:06this all applies to both big and small
47:08companies so I'm gonna go over a few
47:10points here that I think are specific to
47:13let's say Series A and smaller companies
47:16so one one point here is winning what if
47:18your what if you're so small you don't
47:20have the scale to standardize your
47:21process so your seed stage start off
47:24your hiring your interviewing your first
47:25few people you obviously can't run an
47:28extremely standardized process because
47:29the first two candidates have seen those
47:31questions that's a totally real point I
47:35think it still is worth trying to run
47:37instruction interviews so you know you
47:41won't have you'll still you know your
47:42play-doh pie you simply Google Doc with
47:44some tips tips written down but it still
47:46is helpful to think about what skills
47:47matter and try to design questions that
47:50assess those skills that will still
47:53reduce you know the bad bias in the
47:54interview let's say that you're in
47:57trouble sourcing so you're already
47:59staged start up and your number one
48:00problem is we know how do you get enough
48:03qualified applicants for your company so
48:05in that scenario it's fairly obvious
48:07that a false negative costs more so if
48:09you're struggling to get people to apply
48:10to your company screening out someone
48:11who's bad stress all right so here
48:13someone who could have been good is a
48:14very expensive more expensive than for a
48:16large company but the flipside is that
48:18if you're a small company hiring a bad
48:20person is also way more cost you know
48:22from the way more expensive so it's not
48:25cream if the ratio of those two costs
48:26changes so I think you still have to
48:28care a lot about both I think what you
48:31can do is be less aggressive about
48:33screen folks out early so if you're an
48:36early stage companies struggling to
48:37source candidates I recommend you are
48:41less aggressive in speeding out
48:43you're on site you know pass more folks
48:45through and accept a lower final
48:48interview you know success rate in
48:50exchange for you know better screening
48:52all the applicants let's say that you're
48:56a small start-up in use don't know what
48:57skills matter so you're not sure if you
48:59if you want hire someone who's you know
49:01very CCS focus or someone who's very you
49:03know web focused and there's no crisp
49:07answer here so we have plenty of
49:08examples of you know billion-dollar
49:10companies that have taken various routes
49:13here my personal advice from from
49:17trilobite and from Socialcam and from
49:19twitch when it was small is that yeah I
49:22believe strongly that that that that the
49:23two most important skills for the first
49:25few hires are productivity and ownership
49:27so being able to basically take a
49:29project figure out has to be built and
49:31just make that happen and I recommend
49:33applies for that at the expense
49:35potentially of code quality so I think
49:38the first few hires at a company you
49:40should accept that perhaps the writing
49:41crappy code but you know by god they're
49:43writing it quickly and they're getting
49:44stuff done and I yeah just say that was
49:49the case during the early stages of you
49:51know all through comes I've been
49:52involved with and I think that's normal
49:53and probably good so let's say that
49:56you're hiring for an area where you do
49:58yourself don't have technical expertise
50:00so you're you're a web developer and
50:02you're you need to hire an iOS engineer
50:04what do you do so what enters you can
50:06you sure you can use a stick neutral
50:07byte I will help you do that
50:08I you can call them friends did you any
50:11before you if you have them but I think
50:14a good point here is that a trick that
50:16I've used in the past is to just ask and
50:18asking this to explain things so your
50:20web developer you ask the iOS engineer
50:22question they give an answer
50:23you have no idea if it's a good answer
50:24or a bad answer what you can do is just
50:26ask them to explain you say oh that's
50:28interesting can you explain why that why
50:29that's a good idea and for the most part
50:32if someone truly is a skilled iOS
50:33engineer they should be capable of
50:35explaining their answer in terms out of
50:37web screen flow much get to understand
50:38it's not always the case sometimes
50:41someone come along who's you know good
50:43at assumption but the bad communicator
50:44and that that might fail them but this
50:45is a pretty good trick to hire outside
50:47your area of expertise
50:50okay that's basically I want to end act
50:54with an ask for you guys so this is this
50:56is a we've been developing an exercise
50:59for interviewers that we use when
51:00training new interviewers patrol bite
51:03and I want to see if I get you guys to
51:04try this and I sort of email me and let
51:07me know how it goes because I'm curious
51:08but this works more broadly and the idea
51:13is that and you know interviewers tend
51:18to sniffily underestimate how noisy
51:20interviews are they they overestimate
51:22their own ability to distinguish to
51:24engineer some bad engineers and part of
51:27becoming a better interviewer is to is
51:29to sort of become a little more humble
51:31and more aware of your limitations and
51:34so this exercise is designed to sort of
51:36highlight that and so what I want you to
51:39do is to do a mock interview with one of
51:43your co-workers so have one of your
51:45co-workers interview you where they're
51:47role-playing you know they're asking
51:48questions they're the interviewer and
51:49you're role-playing a candidate you're
51:50giving answers and I want you to tell
51:52them in advance that you're going to be
51:54giving bad answers to some of the
51:56questions you're gonna be roleplay I
51:57can't think who makes mistakes and then
52:00during the interview they ask questions
52:01I want you to half the time go ahead and
52:04roleplay a poor answer to the question
52:06and the other half the time giving your
52:08best shot and actually try to answer the
52:11best possible answer you know to that
52:12question and then after the interview is
52:14done do a debrief session where you know
52:18your coworker goes through and sort of
52:19goes over all the mistakes you made and
52:21gives you feedback and what's great here
52:24is that they don't know what answers
52:25that you gave were intentionally bad and
52:27what were you trying your best and if
52:29they MIT look if they don't highlight
52:31the mistake you made where you were
52:33trying to make a mistake they're gonna
52:34look a little bit bad so they're fully
52:36incentivized to be completely honest and
52:38rigorous when putting out what the flaws
52:39they saw in the interview and though you
52:42know the criticism they give will apply
52:43both of the points where you're
52:45intentionally making mistakes but also
52:47to your attempts to give your best
52:48possible answers this is being really
52:50interesting we've been doing this a lot
52:51a trilobite and it's great highlighting
52:55how interviewers disagree so if you do
52:58this with your critic or worker and you
53:00do it a few times reverse roles you know
53:04all well you'll highlight your area so
53:06you're both weak but you'll also
53:07highlight a bunch of areas where you
53:09have legitimately strong disagreements
53:12about what constitutes the best answer
53:13to it to it to a technical question
53:15there will almost certainly be more
53:17disagree more than you're expecting
53:18going in yeah I found this super
53:21insightful I'd love you guys could try
53:22this and I email me let me know how it
53:25goes awesome that's a thing for me
53:27hardies gonna talk about closing and
53:29then we'll both answer questions okay
53:36I'm gonna wrap up here by going through
53:37best practices and point to optimize
53:40your chances of people actually
53:41accepting the offers you make them okay
53:43I have five main pieces of advice here
53:45the first is just understand as a
53:48start-up in particular speed is a huge
53:50hiring advantage that you have over
53:52bigger companies so through two of why
53:54we work with startups and larger
53:56companies and we ourselves are surprised
53:58by how often a big company can take
54:00weeks just to deliver the actual final
54:01offer details and when someone's getting
54:04to the end of their job search process
54:05that's when there's most keen to just
54:08make a decision and move on and know
54:09what they're doing so as a start-up if
54:11you're quick at every step in your
54:13hiring process from the moment you have
54:14the first contact through to when you
54:16make the offer you increase your chances
54:18of closing a candidate so just be fast
54:21responsive at every step
54:22second Almond already mentioned like
54:25train your interviewers to have a good
54:28bedside manner with interviewees and so
54:30we ask candidates or engineers on
54:32trigger by what were some of the the
54:35reasons you had bad experiences when you
54:36went and interviewed in person at
54:38companies and all of the reply centered
54:41around their experience with the
54:42interviewers and the top two complaints
54:44were interviewers not being familiar
54:46with the technical questions they were
54:48asking and second the interviewer being
54:50determined to show how smart they were
54:52specifically smarter than the
54:53interviewee so if someone has a bad
54:56experience interviewing with you they
54:58are highly unlikely to accept your offer
55:00three be prepared to talk in some detail
55:04about your company culture because
55:07almost every candidate is going to ask
55:08you that question what I like to work
55:10here what's your company a culture like
55:11and I think we've seen a particular
55:14uptick in this question
55:16over the past year anything frankly a
55:19uber and just a lot of the focus on the
55:22culture in the technology industry in
55:23Silicon Valley is making people
55:25especially individuals think more about
55:28the kind of culture they want to work in
55:29so two specific bits of advice here I
55:32definitely be prepared to answer
55:34questions about how you think about
55:35diversity if your teams already diverse
55:37or not I think you should have some
55:39thought around how you plan on
55:41incorporating that into your hiring
55:42practices going forward and second when
55:46you're describing your culture
55:48be aware that almost every company like
55:51a significant majority essentially use
55:52three adjectives I open transparent and
55:54collaborative and we've seen this on
55:56trim why because we see them carrying
55:57their profiles right and so that can be
56:00an effective strategy because those all
56:02three of those things are good
56:04aspirational qualities right like no one
56:06would not want to work at an open
56:08transparent collaborative place and but
56:10it's not very good for differentiating
56:12yourself right so one strategy you could
56:15take when you're talking about your
56:16culture is take more risk and the
56:19example the risk might be talk about the
56:21trade-offs in your culture so if you are
56:23if you believe yourself to be an open
56:26culture talk about the trade-off there
56:28right like openness means like being
56:31honest about things and being honest
56:33about things can be difficult and
56:34uncomfortable for people so if you're
56:35willing to kind of go in that direction
56:37you may a lien ate some candidates but
56:40you may also really increase your
56:42closing rate on the candidates are most
56:44excited about you forth get your team
56:47and investors involved in the process so
56:50once you made an offer make sure your
56:52team reaches out afterwards offers to
56:54meet up with the candidate again answer
56:56more questions and generally be involved
56:58in their decision making process and do
57:01the same with investors and in
57:02particular pick any investors you have
57:04who would be particularly good at
57:05closing that candidate so if they're on
57:08the fence about whether to leave a big
57:09company for your startup do you have an
57:11investor that successfully left a big
57:13company and joined a successful startup
57:15and finally present full and transparent
57:20offers with all the details the
57:22candidate needs to understand how much
57:24they're being compensated
57:25this sounds against probably somewhat
57:27obvious but I've been surprised by
57:30companies engage in this sort of
57:32hypothetical battle where instead of
57:34just telling in candidate here's how
57:36much salary and equity you get so posing
57:39questions like how excited are you about
57:41us if we made you an offer do you think
57:43you join and this always creates a bad
57:46awkward candidate experience it's also
57:48unfair because the truth is no matter
57:50how powerful you powerful your mission
57:52is compensation is a big factor in where
57:54someone's going to work and so asking
57:56them to commit before giving them
57:58details it just puts them in an
58:00uncomfortable position and decreases the
58:02chances that they would accept your
58:04offer to join so when you do make the
58:06offer provide full details this slide
58:09has I take the minimum of what you
58:11should put on an offer letter salary is
58:13fairly self-explanatory with equity make
58:17sure that when you present the details
58:19of an equity offer so the number of
58:21stock options exercise price etc that
58:24you're also checking into a Sikandar how
58:26comfortable are they with this right if
58:28they're if they're a senior engineer and
58:29they've worked at 6 or 7 different
58:31startups they probably understand this
58:33you don't have to go into too much
58:34detail if they've spent most of their
58:37life working at Intel they may not be
58:40familiar with how startup equity works
58:41right so ask them how comfortable they
58:43are don't overwhelm them when you
58:45present the offer details if they're not
58:47just send them follow-up resources and
58:49that they can read on their own time and
58:51also this sounds somewhat obvious but
58:54make sure you yourself as founders
58:55really understand how stock options work
58:57because people will ask you questions
58:58about it and you actually don't end up
59:01spending much time on this because you
59:03yourself have restricted stock and
59:05fundraising doesn't usually include
59:07options but you should understand even
59:09down to level all I'd say tax
59:10implications of different types of
59:11options a final thing I'm doing close on
59:16here is a common thing that comes up
59:18with startups is how do you compete for
59:19engineers against the tech giants like
59:21Google and Facebook in particular right
59:23so Google and Facebook obviously can
59:26offer much larger liquid compensation
59:28packages than a start-up can and frankly
59:31like the public tech company stocks have
59:33been doing great over the last few years
59:34so it's becoming more challenging to
59:36compete against them still we routinely
59:39see startups do it and win out against
59:41much larger companies offering more
59:43think there are four pictures they make
59:45that have been quite compelling first is
59:48emphasized learning the the case I'd
59:51make here is that you learn the most or
59:54you learn the fastest when you're given
59:55real decision-making it's a
59:56responsibility which means if you make a
59:59mistake like bad happens and
01:00:01there's very very little chance
01:00:03practically no chance that that's going
01:00:05to happen to a big company right like
01:00:07big companies have checks and balances
01:00:08in place to make sure that no one can
01:00:10create too much damage in particular for
01:00:13new hires right so if you tell someone
01:00:15that hey like we just don't have that
01:00:17luxury we're growing quickly we've got
01:00:19to get everything shipped yesterday
01:00:20you're gonna get thrown straight into
01:00:22making a real decision making authority
01:00:24and if you screw up that's going to be
01:00:26bad for the company and that's the way
01:00:27that you learn and build real experience
01:00:29so emphasize that too I talked about
01:00:33career progression and one bottom two
01:00:36macro comment I'd make is if you look at
01:00:38executives at public technology
01:00:40companies they tend to be you're often
01:00:43can be much younger than you see at the
01:00:45counterparts that like public deactive
01:00:48or like a public utilities company right
01:00:50and the the reason for that is you can
01:00:53grow up and be like chief product
01:00:55officer Facebook at 38 or however old
01:00:57the chief product officer is because you
01:00:59join the startup when it was really
01:01:01early and you grew along with it right
01:01:03and the tech industry is really the only
01:01:05place that happens like it's more of a
01:01:07meritocracy you don't have to pay your
01:01:0920 years dues at a company before they
01:01:12consider you for a senior role you can
01:01:15rise up as quickly as a start-up grows
01:01:16so emphasize that third talk about
01:01:20opportunity cost and in particular the
01:01:23way I'd frame this or talk about this
01:01:24would be at any given moment in time
01:01:27there's always going to be a basket of
01:01:29big safe technology companies to go work
01:01:32at the names might change sort of every
01:01:34five years or so but the experience of
01:01:36working at one is largely fungible
01:01:38whereas the experience of working at a
01:01:40start-up varies wildly right so I would
01:01:43emphasize the fact that your startup
01:01:45right now the team you have the
01:01:46opportunity is a unique one that's a
01:01:48uniquely good fit for that candidate and
01:01:50that they won't get again and if it
01:01:52doesn't work out they can always go back
01:01:54to generic big company
01:01:57and final point this one specific to
01:01:59hiring more junior candidates talk about
01:02:02mentorship so one thing that we see is
01:02:04especially for new grads people started
01:02:05off college is that they're worried they
01:02:07won't get the right level of mentorship
01:02:09they won't learn how to best practices
01:02:11work if they join a start-up and maybe
01:02:13they should go work at Google for a
01:02:14couple of years get that under their
01:02:15belt and then join a startup I mean
01:02:18first decide if that's actually true or
01:02:20not right like if you're if you yourself
01:02:21have only junior engineers on your team
01:02:23and frankly you just own one to mentor
01:02:25or get someone up to speed and then
01:02:27don't make that case and maybe don't
01:02:29hire someone like that and but if you
01:02:31have experienced engineers on your team
01:02:33emphasize that and emphasize the fact
01:02:35that they will get to learn alongside
01:02:36experienced engineers who have worked at
01:02:38places and learn best practices etc etc
01:02:41so that's everything I think we're gonna
01:02:45open up for questions I will ask the
01:02:48questions about the topics we've covered
01:02:49you wanna come up like background things
01:03:16are things that rule out the candidate
01:03:22background background checks frankly
01:03:25most startups I know don't do them
01:03:26they're especially easy to do now though
01:03:28because you can integrate Checker in
01:03:30with your payroll service so sure but
01:03:32that's not something I would like focus
01:03:34on I focus more on like if someone can
01:03:36be a good fit or not yeah
01:03:54that's a hard question so the majority
01:03:57of computer science engineers are
01:03:58probably not gonna be a to a pic Euler
01:04:00good job of doing the work you want that
01:04:05said this sounds like a case where you
01:04:06could probably just evaluate past work
01:04:08this person has done so you could
01:04:10probably ask this person to take to give
01:04:11you a portfolio or other examples of
01:04:13work they've done and if that is at a
01:04:14level and of a type that you're happy
01:04:16with hiring them would probably work
01:04:31yeah probably so the thing we focus on
01:04:35is background blind hiring so we focus
01:04:37very narrowly on just direct skills
01:04:38assessment rather than looking at other
01:04:40things that that correlate in our
01:04:41predictive so you know personality type
01:04:43is real people vary in their personality
01:04:45so that stuff matters for performing
01:04:47jobs I think it's more research on like
01:04:49the big five rather than big five
01:04:51percent of traits rather than - Briggs
01:04:53all relevant but I would try to keep it
01:04:56pretty separate from technical
01:04:57assessment so it's assess technical
01:04:58attributes yeah and then you know
01:05:01separately try to assess you know
01:05:02culture for it friendliness soft skills
01:05:03and then combine that all and they try
01:05:05to make a global decision yep sorry back
01:05:09in the sunglasses
01:05:31that's the extreme version and I said
01:05:33you probably can't so you okay so the
01:05:36question is if you're a non-technical
01:05:39founder how can you assess an applicant
01:05:42technical skills and I think the answer
01:05:44in fortune is that you probably can't so
01:05:47you can fall back in the third device I
01:05:48try to look at past work but I think
01:05:51this is why it's very helpful to have a
01:05:53technical co-founder is because you are
01:05:55going to be at a fundamentalist
01:05:56advantage there yeah I okay so the
01:06:11question is how do you find a technical
01:06:12co-founder and you figure out they're
01:06:14good or not I think sort of Arlen's
01:06:17advice basically applies here they the
01:06:19most thing is the even if you use a
01:06:20friend or some that you know who is an
01:06:22engineer to talk to them and figure out
01:06:24they're good engineer or not otherwise
01:06:26there's there's not really any other
01:06:27options yep yeah honey I'm not precise
01:06:33with my pointing yes but I think this is
01:06:54largely focused I'm sorry moving the
01:06:56question the question the question is
01:06:59other ways to work with the universities
01:07:00and colleges to generate like a pipeline
01:07:03of talent that you could hire from that
01:07:04accurate I'd say yes there certainly are
01:07:09on you know they exist like career fairs
01:07:11and sponsored events at colleges et
01:07:13cetera et cetera but I think those most
01:07:15there's probably work better for larger
01:07:16companies that can fit themselves around
01:07:18the college graduation deadlines and
01:07:21that schedule for start-up is kind of
01:07:23tough because usually you want someone
01:07:25that's available right now to hire as
01:07:26opposed to waiting for them to graduate
01:07:29so I probably wouldn't recommend
01:07:30investing too much time in that as a
01:07:32start-up and just to clarify big
01:07:34companies at this point make offers a
01:07:36full year in advance with large signing
01:07:38bonuses you know you deliver it a year
01:07:40in advance so it's very hard for
01:07:41startups to compete with college hire
01:07:59yeah the question is to clarify unlike
01:08:03my last point about role playing a bad
01:08:05interview so the key point here is that
01:08:06you're gonna roleplay need to do in the
01:08:08cat you know the co-worker who's playing
01:08:10the candidate who's getting the answers
01:08:11they're going to intentionally make some
01:08:14mistakes and what that does is it frees
01:08:17up their coworker to be critical so
01:08:19normally if you say if you do a mock
01:08:21interview between two co-workers the
01:08:22problem is that you know in the end
01:08:24everyone's gonna pull their punches
01:08:25they're not gonna honestly give their
01:08:28opinion of the performance that that
01:08:29their co-worker gave it just doesn't
01:08:31quite work due to all kinds of social
01:08:32pressure if you tell them I'm gonna be
01:08:35intentionally making some bad I'm saying
01:08:37I'm gonna tell you things and mistakes
01:08:39that frees them up where they're taking
01:08:41they're then free to point out the flaws
01:08:43in your interview but only free to they
01:08:45have to because if you say I'm making
01:08:47mistakes and you ask them what's your
01:08:49opinion of not even the results if they
01:08:51don't point out a mistake you made that
01:08:52means you got something past them right
01:08:54so it flips the social pressure and
01:08:56creates pressure where there then
01:08:56incentivize to be critical which is
01:08:59extremely useful because you know as
01:09:01experienced engineers like weird like me
01:09:03it's been a while your senior engineer
01:09:05on a team after small company it's been
01:09:07a while since you've had someone really
01:09:09brutally critique how good you are at
01:09:12answering your own questions and that
01:09:14you know what will happen is that they
01:09:16will point out legitimate flaws you'll
01:09:18be trying to give your best answer and
01:09:19they will point out things you said that
01:09:21are wrong or things that could be better
01:09:22and that it's very humbling it's very
01:09:26humbling and I recommend that they do
01:09:28that's I think this is exercise you can
01:09:29give yourself that experience of being
01:09:31humble and having flaws and even your
01:09:33own performance pointed out in the
01:09:40middle yes got it
01:09:43you talked about a 95% plus success rate
01:09:48on one of the slides and I have given
01:09:53the the fact that the interviewers tend
01:09:57not to have a very high rate of
01:09:59reliability it seems time to believe
01:10:01that there's a 95% and also doesn't
01:10:04match my experience that 95% of the
01:10:07hires successful so when you think is
01:10:12the real success rate where you know
01:10:16you're after hire something's still
01:10:17there and they're pop before yep that's
01:10:21gonna be so this comes from some so that
01:10:24number I 5 comes from some surveys you
01:10:25get upon companies okay sorry so the
01:10:29question was I my slides I said that
01:10:32that there is that tires are successful
01:10:34ninety-five percent of time but that
01:10:36doesn't mesh with you know the
01:10:39experience of being a hiring manager
01:10:40height people many hire are not not
01:10:43great so the question is what's the
01:10:44actual rate at which you know candidates
01:10:46end up being top performers and yeah
01:10:49questions totally right so that that
01:10:51number comes from some surveys we did
01:10:53where we asked companies of all their
01:10:55hires what percentage got fired and we
01:10:57actually asked web stands our top
01:10:58performers and that five times the fire
01:11:01rate so five percent actually get fired
01:11:03you know an additional 30% are people
01:11:06who stick on but are not particularly
01:11:07great employees and then about 10
01:11:10percent are in the bucket that cut that
01:11:11companies said were there their best
01:11:13hires and top performers thank you