Aditya Agarwal - Former CTO Dropbox; Partner-in-Residence @ South Park Commons
Aditya Agarwal is a Partner-in-Residence at South Park Commons - a collective of technologists, tinkerers, and entrepreneurs who have come together to freely learn, explore new ideas, and help each other launch their next venture. Based in San Francisco, SPC has over 150 members and alums who have either started their own companies or gone on to join top organizations like Google Brain, OpenAI, and Chan Zuckerberg Initiative. In early 2018, South Park Commons raised a \$50M seed-stage venture fund to support entrepreneurs both inside and outside of SPC.
“This is actually super counter-intuitive, and many of you might disagree with this. Whenever I meet a candidate at the top of the funnel, my only goal in that conversation is to do right by the candidate.”
- Aditya Agarwal
Prior to SPC, Aditya was the CTO and VP of Engineering at Dropbox. He scaled the Engineering team from 25 to 1000 and was responsible for new product development, infrastructure, and technical operations. Aditya came to Dropbox via the acquisition of Cove, a company that he co-founded. Prior to Cove, Aditya was one of Facebook’s first engineers. He helped build the first versions of key products like Search, NewsFeed and Messenger. He then became Facebook’s first director of Product Engineering, overseeing engineering for products like NewsFeed, Profile, Groups, and Events.
Aditya serves as an independent director on the board of Flipkart, India’s leading e-commerce company. He is on the advisory board of Carnegie Mellon’s School of Computer Science and on the board of trustees of the Anita Borg Institute. He is also an active investor and advisor to Silicon Valley startups. In his spare time, Aditya likes to listen to techno music, read, travel and most of all, hang out with his 2-year old. Aditya holds a bachelor's and a master's degree in Computer Science from Carnegie Mellon University.
Dan Portillo - Talent Partner @ Greylock
Dan is Talent Partner at Greylock. Previously, he was VP of Success & Engagement at Rypple, and VP of Organizational Development at Mozilla, creators of Firefox. Earlier in his career, Dan spent a decade building out successful early-stage, venture-backed consumer and enterprise companies. Dan also served as a Council member for Code2040.org, a non-profit creating opportunities for underrepresented minorities in tech.
Dan: So have you ever not promoted an engineering leader because they couldn't recruit a good team?
Aditya: So we haven't rehearsed this, clearly, 'cause that's kind of like a, that's a great question.
Dan: That wasn't on the card.
Aditya:I know, what I'd say is that the best answer to that is basically like I wish I had not promoted people who did not know how to recruit. And, the reason I say that is that, how many of you, like I'm curious, in the audience, if you don't mind raising a hand, how many of you, one of the key parts with job today, as expressed by your direct manager versus CEO, is recruiting at a ridiculous percentage of the size of your team? Great, so I was in that position myself and I did not have the, I didn't quite understand, I was having to do that. I had to basically recruit.
Let's see at Dropbox, we went from, our company got acquired in early 2012, so we went from 26 engineers in March of 2012 to 87 engineers by the end of the year, at the end of 2012. And then, we went from 87 to 170 to 260 to 550, and then close to 1,000 and I think today it's close to 1,500. The crazy thing is that you don't know how exactly to do that. Yet, that is the key expectation that everybody has of you, whether it's your direct manager or the CEO. So, the answer to your question is no, I didn't know I was having to do that.
Dan: The question was, should I have? And, I think you answered that.
Dan: So, in one of the programs that I built at Greylock was our community program where we're building networks of infrastructure leaders, data leaders, designers, and in the events, recruiting always came up. Which is just more and more, recruiting is being pushed to the edges and the managers of the individual teams are actually being asked to do more of the work. What do you believe is the role of the engineering leader in the recruiting process?
Aditya: So, I think this might be a slightly controversial take on it, but I believe that too many engineering leaders don't really fundamentally own the fact that... whether or not you hit your headcount number, whether or not you hit your recruiting number... It's your responsibility.
And, I think it's really easy to come up with a bunch of reasons. Recruiting didn't give me enough top of the funnel. Our office is not good enough for us to close for the people that we want at the bottom of the funnel. Our interview process is too hard.
But, what I'd say is that I've been surprised as to the lack of ownership generally in Silicon Valley about the fact that recruiting is such a core part, given how many of you raised your hands, of what the role is of an early-stage engineering leader and I think it's something that... I'll give you an anecdote.
I remember getting, we got acquired by Dropbox and I remember talking to Drew early on in 2012 and it's like, Drew, what do you want me to do? He's like, just go increase the size of the team. I'm like, what do you mean? He's like, just fucking 3x it! I'm like are you serious...? Because we can…
He's like yeah but make sure you keep the quality consistent right. I'm like okay... okay... this is hard but that's what we ended up doing and it was at the expense of other things in terms of our ability to do other things but you have to really own it and internalize that this is a key part of your job.
Dan: So I'd like to get like tactical a little bit so part of it is from a recruiting perspective that is both the sourcing piece of getting good people to come into the company and then there's the actual closing piece of getting them to join and there's a number of factors in there. Are there any things from the Dropbox or the Facebook experience that allowed you to cheat or get an advantage from a sourcing perspective?
Aditya: Yeah that's a great question.
This is actually super counterintuitive and many of you might disagree with this. My goal with top of the funnel in terms of you can call it sourcing and call it like the initial, the first conversation or what you're gonna call it, whenever I meet a candidate at the top of the funnel, my only goal in that conversation is to do right by the candidate.
And I take this to an extreme, which is that if in that conversation I'm like actually, Dropbox is not the right company for you, you should actually go and talk to this other company.
I have made so many of those introductions after that first conversation and my take is that… Number one, if you have discovered that like, let's assume Dropbox, productivity software is not what the person wants to work on, why would you want them to work on, like why would you want them in the company either way?
Instead, you can actually be forward. And you'd be surprised, every one of the people I've referred to other companies has sent me somebody else in their network who is a better fit for Dropbox, right.
And I think I've been surprised also by how hard a conversation like this has been with other people who are smart. Like it seems really obvious to me, right? Do good and do good by the candidate and like they'll basically pass it forward. But what I'm here to tell you is that THAT actually works. So that's actually the one biggest, I think you've told me the same thing.
Dan: I think it's important to take the long view on recruitment.
Dan: So many managers are like, well if they're not looking to move right now I don't wanna talk to them. My longest from the first time I tried to recruit someone to getting them was 15 years.
So they were like new grad outta Stanford and then when I finally closed them it was for an executive job, they're bald with like three kids, it was like a completely different point in time. But Silicon Valley and like Reader and XP will tell you that networks are incredibly valuable and so you have 20, 30 minutes to make an impression on someone that allows them to come back and say like you should talk to this person, I think having that valuable introduction gets referrals and gets you known as, like person X is trying to figure out what to do with their career, like you're the person to go talk about that because you're partial but like you care more about the outcome for the individual and not like optimizing for your own.
Aditya:I mean like my longest... what did you say yours was?
Dan: 15, 15.
Aditya: Fucking hell, mine is about 13 years, my first day of my job ever as a 21-year-old working at Oracle was Pete and I tried to convince Pete to come join me at everywhere I worked but eventually at Dropbox, 13 years later, he was like yeah, he answered the call and he was like you still actually care about me as an individual as opposed to, as opposed to just optimizing for yourself so it worked out, yeah.
Dan: So I talked to one of your former engineering managers earlier today and he said you're a beast about closing people and so they're pretty creative. You guys did some interesting things on the University side that I'd love for you to talk about and then you also got pretty creative on the compensation side, specifically when you're going after kind of staff and principal level people out of Google.
Aditya: Yeah, I'm gonna talk about the comp later but let's talk about the creativity part of it right.
Here's my hypothesis, my hypothesis is that most of us in this room are actually fairly intellectually lazy when it comes to closing candidates. And what I mean by that is that, let me give you an example, I think that'll illustrate it.
There was a candidate, incredible engineer, when Dropbox was maybe I don't know like 70, 80 engineers and he came to us. He was like listen, I'm gonna have to move out from Boston and you know, I have a family, a growing family and I just can't find a house and like a living arrangement that makes sense. And we were like... the easy reaction would've been that like, oh yeah, San Francisco is fucking expensive like nobody can afford to live there. So this is not somebody that we can close and that would've been an easy out.
And I honestly saw a lot of engineering managers at Dropbox taking that ‘out’ but as we dug into it further… what we found out was that he was not looking for a cheap housing solution in San Francisco. What it was, was that he was looking for a religious community. He went to church every Sunday and he was looking for that setup that he had in Boston on the east coast here.
So at Dropbox, we ended up actually working with him to find him a set of options in the Peninsula 'cause San Francisco is hard in that respect. That satisfied that. And I bring this up because you can't always find this particular thing that motivates people. But it would've been really easy to say that, yeah he's not going to join us because housing is too expensive, right?
So the reason I bring this up, is that as engineering leaders and hiring managers, we have a lot of outs in our scenarios we like to give ourselves, but do you actually wanna take that out? Or do you want to actually understand what the candidate wants, and to dig deep? It’s unlikely that it'll always work out, but you can do a lot more than you think, in my opinion.
Dan: And I see this with companies all the time as they rush to offer. So like do all of this time evaluating and then they don't take that extra phone call to say, “what questions do you have left? What last things do we need to answer for you? Do you wanna actually work here?”
And then that's when you learn about whether they're missing a community piece or they, you still need to do some more work with their significant other.
Aditya: Oh yeah absolutely.
Dan: Or they're tryna figure out how do they adjust if they have kids in school and want the compensation piece? I mean with one specific company, they made an offer to a candidate and then I talked to the candidate. He's like, I don't know who I'm reporting to. I don't know what job I'm supposed to do. I don't know my core responsibilities.”
I'm like there's no way in hell that guys gonna take that offer. So we had to back it all up, answer all of those questions and he eventually accepted. But if you just do it ahead of time, like you save yourself a whole bunch of thrash.
Aditya: Can I share this story?
Dan: Absolutely not.
Aditya: This is one of my favorite ones from Dropbox.
So there was this like, crazy good intern. We'll just call him with his initials… SS okay. SS was incredible throughout his three months over the summer in maybe 2014 at Dropbox and he really did not wanna go back to school. And I remember that I had to go talk to his parents and convince them, because his parents were just like, “dude, like what are you doing to my son? You're convincing him to not go back to school, you're a bad person.”
I'm like Auntie, Uncle, hold on, hold on... I will make sure he gets his degree, right. And eventually after working at Dropbox full time and he was only six months away from graduating, Drew and I kind of like forced him to go and get his degree. And well I'm Facebook friends with his mum and dad which is amazing. But what I'd say is, you can make these promises but it's really important that you do not say just for the sake of hiring the person.
You have to be true to your word because this shit will come back through. Life is small, karma gets back to you and I actually am really proud of that fact that like between Drew, Arash and I, who all of us really care about the individual, he got his degree, his mum and dad are happy and he's still a happy Dropbox employee and that just really matters to me.
So maybe what I'm saying is that, in the course of recruiting, you can find yourself making absurd promises. It's fine to make those, but if you mean them, like if you mean them, then it's all good. And you should actually stretch yourself to do the crazy things as long as you mean them. Don't say shit you're not gonna do.
Dan: I do wanna talk a little bit about comp.
Aditya: Oh sure.
Dan: So relative to most companies, Dropbox had a war chest. But like even compared to some of the larger companies, you're still a small stack at the table. So what is your approach to how you leverage compensation and do you have one strategy that you always use or how do you apply it depending on who you want?
Aditya: I'm gonna actually turn this back to you really quickly because I think you're a better at the comp stuff than I am. But my simple take is that most people, most engineers, are looking for satisfying. Basically they're looking to get to a point where they feel as though they're not getting screwed and they're being fairly treated. And then after that, it's about all the other things. And then they might use comp as a way to talk about the other things, but it's not actually comp. They're just looking to get the minimum level that they feel as though they're being treated fairly.
Dan: I agree, so I think being able to talk about comp in a way that says “here's market” and understanding that like there's fairness. I also believe that, and you tell me if you disagree, that there's always gonna be 10% of the company roughly, that delivers 90% of the value in the set of things that you're doing. And so there are situations where you should be using comp selectively in asymmetric ways to make sure that you're able to keep those people even at Dropbox.
Aditya: Oh no I strongly agree with that, I strongly agree with that. The way that I think of it, is not to basically attach preconceived notions. Let's assume you have a level one through level six in terms of like an engineering career ladder. I think it's completely fine for a person who has five years of experience to be a level six, right? Basically, have some rigor in terms of the structure that you have, but then don't have a bias in terms of how somebody can go through the levels or jump through the levels if you need to. So you need to have the rigor and structure but then also be unbiased in terms of how you move people through the levels is my take on it.
Dan: Yeah I think you need to exercise judgment and understanding that I believe 100% of having rigor so it allows you to like hire to level and make sure compensations all over the place but at the same time... like if someone carries most of the water make sure that...
Aditya: Yeah I mean like to be blunt right, like it's totally fine for somebody three years out of college to be at the top level of your engineering ladder if they're adding enough value.
Dan: What do you say to someone who's having to build their team for the first time?
Aditya: What I would say is that this is, that's a great question.
What I would say is that, I really hope that you have a good manager. And I mean that seriously. I feel as though I've learned this the hard way myself. Most of the time the constraints are given, he's like “oh you have this much money to go and grow the team” or “we're gonna double the team” and everybody just focuses on recruiting and hiring.
But more and more I'm just like the biggest limiting factor is not the growth rate of a company but like the productivity rate of a company sorry throughput rate of a company, for engineering is actually the number of good engineering leaders that they have.
So how do you bootstrap engineering managers and I guess this is why we're all here. So you have all made good career decisions. But I think that the number of good engineering managers that you have in a company is actually the highest, the biggest limiting factor that you have in terms of growth rate.
Dan: Yeah I think that's another hour which we don't have 'cause I think we're running out of time but I like this question, so does comp asymmetry reward good performers or good negotiators?
Aditya: Oh that's a good question, it depends, well I actually think it totally depends on your system.
I'm a big fan of Facebook's system, if I can talk about it for a second. Facebook is, I think it's a little bit out there. So Facebook is a seven-point scale in terms of the rating that you get of how you've done. And the seventh point, the top, (which I never got, I was there for six years I was pretty early there) is redefining excellence. That's how you've done and you basically get like probably 3-4x equity refresher than somebody else gets.
And I looked at it and I was just like boom, amazing. There are people there who are adding that much value that are getting that, I was totally cool with it. So my take is that I feel like you can negotiate and all that shit but like that's fine. It's way more interesting to me to understand what is a steady-state kind of comp within these companies right? We can also talk about on the way in but that's just in my opinion not that interesting.
Dan: I agree, I think that on the way in you should strive to be as standardized as possible.
Aditya: Exactly, yeah.
Dan: And then I think comp is a tool that most companies just make a mistake on. Every year you do a standard process or some adjustment pieces and so you kind of miss on this opportunity to reward people ahead of them asking for it because that changes the dynamic entirely. When you surprise someone, when you delight them to say like, “We see you, we see what you've done and we wanna reward you for that.”
Like that endears them to you.
Aditya: Yeah, I mean, like you said a one-hour conversation but like my take is actually that we should only give maybe a two-year grant on the way in, and then all of it should just be like based on the mutual fit, the mutual performance.
Why do we give four-year grants? In my opinion that doesn't make sense. The length of a grant should be based on how long people last at companies and most people don't last four years nowadays, right? So anyway, long story.
Dan: I think we're on the same page, so like how does a five-year person get to be level six?
Aditya: I will not give a comprehensive answer but I'll give an answer that I think is non-intuitive maybe for most managers.
One of the profiles that we struggle with at both Facebook and Dropbox is how do you really value a person who is not working on the hardest shit. But is that person who is working and basically is like always there, working on the kind of like the stuff that is not glorious, like working on the stuff that is essential, not sexy stuff, and kind of like giving their heart to it.
And invariably all of you, come performance review, the question you're asking is not like the total amount of output or the throughput. In your head you're basically thinking “what is the one seminal thing this person has done,” right?
And the person that ends up getting screwed in this whole setup is the person who has not done anything seminal but has done like a lot of small things and has been there when you want them. And they never got rewarded in the sense of performance reviews.
So if there's one thing that I would think about as you all tailor your level six, is how does that person who's always there for you, is a super hard worker, but is not working on like the hardest problems or the sexiest problems get to that level?
And I feel personally really, I know a bunch of people at Facebook who have done that and at Dropbox are incredible engineers. And I feel really strongly about making sure that we reward them. So I would encourage you to think about that as well because the easy one is to reward people that worked on hard problems but hard is not always high utility.
Dan: So the CEO of Coda used to run YouTube. He did a master's of scale class with Reid. And one of the things he did at YouTube, which completely broke comp for them was they had the engineering managers ask, “if you're rehiring the team, who do you draft first regardless of level?” and then they bonused accordingly.
So like you could be a level one, level two but you would be getting a bonus at the highest level because they focused on what was their overall impact to the team, starting to remove level from the calculation.
So there's a question about recruiting when you don't have a brand. Every company starts without a brand. When I joined Mozilla, people thought that they had to work there for free and so we had to educate people like, no, we actually will pay you if you come to work for us.
You had to recruit for Cove. You didn't have any brand, I guess it was your brand?
Aditya: It's true. You know, I'll be honest, I don't actually have a great answer for this. I've been in the lucky situation of being with companies that have had a great brand, Facebook, Dropbox, Cove, which is a brand, my wife's brand. What I'd say though is that you can still do all the basics right, right.
Like you can, you can be good people, you can have a good culture, like, you can actually attract people if you have interesting problems. Like you can do the basic things right and after that, it's not even about the brand it's just like yeah somebody cares about what you're doing, they should work with, they might work with you and if not, probably a good thing that they don't work with you, right. So I wouldn't, this is not something that I would stress out about it's like in my head is what I look at, yeah.
Dan: Well I mean I think your margin of error goes way down.
Aditya: That's it, yeah.
Dan: So there's plenty of large companies, I'm sure people have interviewed some of them, and like they take weeks to get back and you still take the offer anyways but you have to execute really well if you don't have a brand.
Aditya: Yeah don't make the mistakes, right, like be quick, be genuine, make good offers. Like don't try to like short change people and yeah exactly.
Dan: You have to sell the value like you have to, I think we had a, we'll do one more question.
Aditya: Okay sure, why not.
Dan: How do you teach engineering managers to sell the company to people that they're trying to recruit? Or what do you think they need to know to effectively sell the company?
Aditya: I would say that, okay this is a good question. If you're at the point that you are actually in the process of closing or selling, you should be super transparent about all of the stuff that is not happy in your life right now. Both yourself individually and potentially the company as well because candidates are smart enough.
Like if you say stuff that's too rosy or if you don't like actually give them the real talk, people will actually just discount everything else that you say. So my take is always just to be like super transparent. That is not to mean that you can not be optimistic, right. Like you can be optimistic about it. But just gauge it, being like hey, this is the big upside, this is also what could go wrong, this is the stuff I'm worried about and you'll be surprised as to how many people appreciate the super real talk. Yeah.
Dan: I agree and like there's like this is us. Us, this is us the team, us the company, this is you the individual, this is how you fit into us the team, us the company.
Aditya: Oh sure yeah.
Dan: And not enough people know how to think like an investor. So to say and a lot of times when I'm helping portfolio companies it is giving an investor side perspective to that of like here's all the interesting technical challenges, then understanding what it is that they care about. Like some of it maybe they just care about the technical challenges.
Some of them want to be able to know is this the company that I could spend my next X number of years for and so understanding both like what is it that someone desires. But being able to know how to understand what they value and all of the various kinds of things in the tool book about the company that allows you to kind of communicate.