Career in Tech

Why Databricks?

Last month I accepted a role at Databricks. I’m writing this post explaining my decision mostly for me to look back on and hold myself accountable to, but perhaps it will also help anyone else going through a similar decision.

I’ve decided to be candid about the decision as that is the only way this post will be helpful for others. I suspect the ideas will be relevant for multiple groups of people: those in the exact situation (deciding between Google and Databricks), those in similar situations (people considering Databricks or people deciding between another large tech company and a smaller tech company), and even those people making general career decisions.

I mainly want this current blog post to be about Databricks and the reasons behind my deciding to join them, but it is impossible to consider that decision completely in a vacuum. Every decision represents passing by some other opportunity, and in this case, the other decision to join Databricks was also explicitly a decision to leave Google and to not pursue hundreds of other great companies. I might touch on those aspects some, but I want the primary focus to be my reflections on how Databricks performed in my evaluation and what, specifically, I’m hoping to gain while working at Databricks.

During this blog post I’ll be talking some about my job rubric. If you aren’t yet familiar with that, here is a blog post that goes over it that might be helpful to review first.

Overview on the Decision

At a high level, this was a very tough decision, as my job at Google was great and I was setup for continued success there. I had been on a career fast track, had just recently been given a lot more scope, had great team of PMs reporting to me and was an expert in a critical domain with lots of trust built up around the company.

So why change companies?

I think there is a cost to staying in one place and my default position is that periodic movement is a good thing. Perhaps for related reasons, the standard practice in my industry is to award stock grants in four year chunks and so I had decided long ago, that after four years at Google, I would take a moment to pause, look around and see what other opportunities might exist. It was likely I wouldn’t act on that decision, but at least by looking, I could make an informed decision and avoid stagnation.

I began that process in early 2022 by making a shortlist of companies I might want to consider talking to. Before I had gotten the chance to do anything else, the Chief Product Officer of one of those companies reached out to me to see if I wanted to talk about a potential role. When the head of your function at one of your top prospects reaches out to you, the answer to the ‘would you like to talk?’ question is always ‘yes’, even if the timing is really bad. If the timing is decent, that is a bonus.

My process was a bit backward, but eventually resulted in some formal interviews, which I looked at as a chance to learn about the company equally as much as they wanted to learn about me. A few short conversations later (I joke a bit, the Databricks interview process was probably the hardest I’ve ever been through) I had an offer. This is where the hard thinking came in – it took me about a month to decide, which I hope says a lot about the quality of both options and how much I valued working with all of the people on my former team. Alas, you already know the result, so I’ll get right down to the reasoning.

Requirements that Databricks met (and other companies do too)

I wouldn’t consider giving up a great role at Google if the other opportunity didn’t meet certain table stakes. Databricks met the following (as did Google and as do a select few other companies)

  1. Great People That I can Learn From – I’m all but convinced that if you could make career decisions on only one attribute, it should be that you are working with great people that you learn from. Every year I read about a successful person and the one thing they all have in common is that they are surrounded by great people. No person is an island. Even if you had to pick between a project that seemed better and a group of people that seemed better – I think the people aspect would be more important. In the short term, sometimes it ends up being better to work on a project that goes well, even if you learn less, but in an longer-term period, I’ve found that the short term costs of project failure are more than made up for with lessons learned and relationships formed, if you were working with great people. Databricks has great people. Google has great people. You won’t catch me working at a company that doesn’t have great people if I have any choice in the matter.
  2. Growth Potential – I like to be a part of things where the best days are ahead of us and there are natural tailwinds pushing us forward. The world of cloud is one of the most transformative movements currently happening and I believe that will remain true for the next decade or so. Amongst that, I believe the subset of it that focuses on data insight is the most exciting and impactful part. So a company focused on cloud-powered data insight basically set the growth potential bar in my mind.
  3. Simple Business Model – I’m a student of Charlie Munger and Warren Buffett. They preach a message of simple business models that make sense and it has worked quite well for them. Along with that, I really like it when a business model is clearly beneficial for all parties involved, because that is the only way you can scale long term. If one party receives too much in the transactions, the others will either go out of business or revolt. The key to sustained success is letting everyone win some – that is such a simple concept, that is also beautiful and, unfortunately, rare. Databricks checked this box for me – we build technology that lets people make use of data that they have and we charge them a bit of money for that. So long as that data has value and the technology helps unlock it, everyone wins. I believe strongly in the value of data (look at all of the cool things we did at Google with data – real time traffic alerts, spam filters, search, etc.) and I believe strongly in software as a scalable delivery method.
  4. Technologist-led company – there is a distinct difference when the leaders of a company are technologists (engineers, PMs, data scientists, etc.) vs strict business people (sales, marketing, operations, etc.) I’m not alone in this thought, Marc Andreessen was recently quoted giving the advice ‘Find the smartest technologist in the company and make them CEO’. He believes the lack of willingness to do this is why traditional companies keep getting disrupted by technology players. I’ve found that being technologist-led to be both more enjoyable and more fruitful. As I look back on the companies I most respect – most of them were founded by a nerd – a software nerd, a hardware nerd, an accounting nerd, etc. The last two companies I worked for (Hearsay Social and Google) were both founded by Stanford graduates so I felt comfortable with the idea of moving to a company founded by PhDs and post-docs from another bay area university (UC Berkeley in this case). When you have a technologist-led company, many things happen, you get more nerf guns, you have fewer suits, but you also end up prioritizing a certain intellectual, first principles, truth seeking philosophy, which I really enjoy and have found to be highly correlated with success.
  5. High Importance of My Project – I’ve found that I most enjoy when I’m working on a project that is between #3 and #10 on the company’s list of priorities. The front projects have too many cooks in the kitchen. Projects lower than #10 don’t get enough attention or funding. The nice thing about Billing is it always tends to be right in that sweet spot. Getting billing right is never the company’s top priority, and for that reason it is never overfunded. But getting billing wrong always causes alarm and pushes it right back into the top 10 (assuming the company has revenue, which if not, we likely failed another test).
  6. Low Risk – I’m not at a point in my life where I want to take a lot of career and income risk, and if I did want to do that, I would likely just start a company and make sure I was setup to capture the full reward for that risk. I’m not at a point where I want to prioritize founding a company though because of the duration, intensity and commitment that requires and the fact that I’m balancing it with another long, intense commitment in having four young kids. I had decided that my next career step was likely going to be one of two things: working as a manager at a big company (50k+ employees) or leading an area at a pre-IPO or recent IPO company (2-8k employees). Those felt like two stages that fit my risk-reward profile, where founding a company and being VP of Product at a smaller company (100-1k employees) didn’t feel like the best way to balance income, learning opportunities and growth potential. Tech has generally been a fairly low-risk industry for the last few decades, though the last six months have seen some layoffs. When I was making my decision the strong companies with solid fundamentals were still hiring, and most of the layoffs were concentrated on companies that had hired faster than their business had grown or who had been too leveraged. Databricks is among a set of companies that seems fairly stable due to having a solid product, many customers and enough cash in the bank to weather a decent storm. Looking towards a slightly longer future, my thought was that the risk of movement was generally pretty low because even if things don’t work out at one company, there are others that would be eager to hire folks from there. I’m not burning any bridges by joining Databricks and even if the company were to fail, there will be a lot of people willing to give me a role elsewhere based on the work I’ve done in the past. So this decision, even though slightly riskier than staying at Google, is pretty low risk overall.

Unique things about Databricks

Now that the table stakes were met, I wanted to think about things that were fairly unique to Databricks

  1. Data – I love data and that is core to Databricks’ mission. It’s even in the name! Google also has a very data-driven culture and many great data products, but at this point in time, the company has grown so much that it no longer feels like a company that is exclusively betting on data in the same way Databricks. There was something very exciting for me about working for a company that is so deeply invested in a topic I really enjoy. I was looking back at my twitter history recently and noticed that I nerded out when I first heard the term ‘Data Lakehouse’ – a term Databricks helped coin and make popular – even though that happened a long time before I had ever considered working at Databricks.
  2. Building for a championship – In 2019 the Tampa Bay Buccaneers were the 22nd ranked team (of 32) in the NFL. The next year they won the Super Bowl. How did they turn things around that quickly? They signed Tom Brady (the best football player of all time), declared their intent to win and then signed a half dozen other great players that wanted to be a part of that. Sometimes winning can be a self-fulfilling prophesy. I have an undeveloped theory that if you just watch the flow of A players in tech, you can spot the next companies that will be generationally-transformative winners. Bell, HP, Microsoft, Apple, Amazon, Google, Uber, etc. When I did my research, I noticed a lot of A players from my extended network going to Databricks and that was a strong signal for me. It doesn’t guarantee a victory, but not having A players almost always ensures failure. You simply can not win without A players and A players want to be at a place that is winning. These truths mean that for better or for worse, the narrative of a company winning often fulfills itself given a few other solid fundamentals and management that doesn’t blow the opportunity.
  3. Unified mission – Aligning the incentives of thousands of people is a very difficult thing to do. At large companies, individuals can often end up more motivated by personal success (a promotion, raise, closing a particular deal, stroking their ego, etc.) than shared success. One thing I really liked about being at a startup when I worked at one a few years ago is that the incentives are so much better aligned between employees. Being captain of a sinking ship doesn’t do anyone any good. Databricks feels like it is still at that stage where everyone is focused on helping customers to succeed, so revenue will grow. Part of that is that it is pre-IPO, so there isn’t liquidity for employee stock, meaning financial success for individuals requires a certain amount of success for the company. I think that can be a very motivating and unifying feature of a company that stood out to me about Databricks.
  4. Pre-IPO but Potential to IPO – On that note, being pre-IPO was a positive signal for me. Many people view an IPO as a finish line, which is the wrong way to look at it. Going public is often a time to celebrate, but the reality is that it offers a bit more liquidity with a lot more continual reporting and regulatory effort. The fact that Databricks isn’t public right now is an advantage as it gives it a few more strategic options that public markets might not tolerate as much as the founders and VCs would. I have also never been through an IPO at a company I was working at, so the idea that Databricks was possibly close to one was somewhat exciting, though personally I would prefer if that were 2-4 years from now rather than ASAP. Currently Databricks is listed as one of the ten most valuable private companies, so this puts it in a prime place to IPO if/when leadership decides to do so. That means liquidity isn’t a pipe dream as it is for some early-stage startups, it could become very real quite quickly.
  5. Founder-led – in my experience, founder-led companies are a better bet. I don’t know all of the reasons for this, but it seems they are able to take bigger bets and be bolder in their approach. This is often rewarded by their ability to quickly find new innovative levers for success or quickly fail and move on (which is just as good sometimes). I’m not sure why hired-CEOs can’t do that as effectively, perhaps it has to do with how a founder’s net-worth is tied to the company value, perhaps it has to do with their voting power, perhaps it has to do with an emotional connection to something they created. Whatever the reason, there is a difference, the Harvard Business Review did the hard work of getting data to prove that. If you look around at the tech industry right now, most of the big tech companies are no longer founder-led. Microsoft is on its 3rd generation of CEO, Amazon recently promoted a new CEO, Google’s founders have stepped out of day to day activities, Apple’s founders have moved on or passed away. (Update Jan 2023: Netflix’s founder recently stepped out of the co-CEO role too so I’ve removed them from the later list). There are a few good founder-led tech companies these days; Facebook, Stripe, Oracle,, and SpaceX to name a few, but many of those failed one of my earlier requirements (technologist led, simple business model, setup for growth, low risk, etc.)
  6. Bias for action – In his book David and Goliath, Malcom Gladwell talks about how sometimes the small can defeat the giant because their weakness forces them to find a new strength. This is sometimes described as the innovator’s dilemma, which similarly says that when you’re successful you have less incentive to disrupt yourself, but the up-and-coming player has every reason to try anything because they have less to lose and more to gain. I think Microsoft, Google and Amazon are all great companies that are having a lot of success in the cloud space, but the fact is those are the 3rd through 5th largest companies in the world and because of that, there is a lot less to gain and lot more to lose. Databricks has an advantage in that they have everything to gain and a lot less to lose. I’m not sure if Databricks will be the next Google, if it can grow 25x and surpass a $1T valuation, but I am pretty darned sure that Google isn’t going to 25x in the next decade (inflation excluded), because that would put its valuation at more than that of the entire S&P 500.
  7. “It feels like Google 10-15 years ago” – This phrase came up for me a few times. I’m not fully sure if it is organic or if the recruiting team has helped fuel the prevalence of the phrase, but I am sure that it helped convince me to join Databricks. One of the few regrets I have in my career is that I didn’t set myself up to work for Google in the early 2000s. There aren’t many other companies I feel that way about. Simply put, Google was one of the best places to work, transformed the world and was financially rewarding. Usually you get one or two of those things, all three is rare. During my time at Google I got to work with a lot of people who did work for Google in the 2000s and most of them shared about ways the company had slowly changed. This is natural, a company of 200k that is growing at 20% YoY is not going to be exactly the same as a company of 5k that is growing at 100% YoY, but it means that the excitement of working at Google in 2022, while great, doesn’t quite replicate the excitement of working at Google in the 2000’s. More than a few of those early Googlers made their way to Databricks though and I kept hearing from them that the culture felt familiar to them. I’m excited about that for many reasons, I think it means it will be a great place to work, I hope it means we will be able to transform the world and if I’m really lucky, it might also be financially rewarding for me.


As I mentioned above, the decision to join Databricks was really difficult for me, because I was giving up a lot. I want to call out a few items that were strict tradeoffs and how I thought about them. I suspect this section will be the most valuable to people considering leaving a large company for a smaller one.

  1. Change vs continuation – There are good reasons to stay at a company/team/role and others why change is good. This decision primarily came down to a tradeoff around that. After four years I knew how Google worked and I was rarely out of my comfort zone, this meant that I wasn’t learning as much, but it also meant I could get a lot done with a little bit of effort. Going to a new company meant I would have to learn a lot really quickly and realistically I wouldn’t be adding much value for the first 6-12 months. I’ve adopted a rule that helped guide my decision here – you should try and steer your career so you’re always in a job that is 1/3 aligned with your existing strengths, 1/3 areas of your competence that are valuable to the company and 1/3 an area of strategic growth for you. If you continually do that, you will never end up stagnant, but also never end up in a role where you are completely in over your head and at risk of getting fired for incompetence. Ultimately I felt I could learn more at Databricks in a few strategic areas – the space (data analytics + AI), other cloud providers (Amazon and Microsoft) and a different way of PMing.
  2. Run a great team or build a great team – I was managing a team of 6 PMs at Google and we had an awesome engineering org we worked very closely with. There was a lot of hard work ahead of me, but I knew who the key people that were going to help me do it were. We were a part of a larger org of ~25 PMs and 250 engineers where I was one of the leaders, but was not the leader of the function or the definer of the culture. At Databricks, I’m the only Billing PM, so I get to define the team and help set the culture of the organization with my ENG and TPM counterparts. There are pros and cons to each approach and really neither was a bad option here, but this was a clear tradeoff for me. In the end it was really hard to pass up the opportunity to build out my function from scratch. I’ll probably only get 2-3 chances to do something like that in my career, where as there will be many chances for me to lead great teams.
  3. Two in the hand vs one in hand and three in the bush – Google and Databricks presented me with a very clear tradeoff decision between something great and guaranteed vs the chance at something potentially better. I was on a career fast track at Google and well payed. If I just hung out there and didn’t do anything stupid I would likely be a Director very soon and on a VP path, with a yearly income that surpasses what I had ever thought possible. I was responsible for the billing engine that processed ~$25B a year in revenue. That is two in the hand – all-but-guaranteed success. Databricks represents one in the hand and three in the bush. I accepted a role that in many ways feels like a step down, because I believe there is potential for it to be even greater as the company grows. The current scope is smaller, the whole company only has $1B yearly revenue (as of the last public statement) and I will take home less than half the money over the next year than I did in the last year. But… if the company grows, I could be in the right place to be in an even bigger role and my stock package could end up being worth more in four years than my Google package would be. What I had to decide on is the expected value of the Databricks role based on the potential outcomes and the likelihood of them. In the end, I think there is a strong enough chance of success and enough upside for me if it happens, that I was willing to take the risk.

What I Hope to Achieve at Databricks

Now that you know why I chose Databricks, I want you to join my in the process of truth seeking. You can help hold me accountable to some goals I’m setting for myself and keep me honest as I evaluate things I hope will happen to me. You don’t have to do much, just check back in here every year or so when I write a reflection on my year and how I’m progressing on my goals. If you’re feeling really ambitious, you can leave a comment of encouragement in the blog post or call me out on some error in my thinking.

  1. Learn a lot – I said above one of the key reasons I joined Databricks was that I thought it would be a better place for me to learn a lot. I want to make sure that stays true. I expect I’ll learn a lot about recent developments in the data/analytics space, the other two big cloud providers (Amazon and Microsoft), new ways to be a PM, and different parts of the billing stack. I also expect I’ll learn a lot of things that I can’t predict right now, and I’m excited about that. My action here is to keep learning and whenever I feel too comfortable, to ask for more work, so that I have to learn more.
  2. Build a great billing team – A big part of me coming to Databricks was the chance to define the culture and scope of the billing team. I want to hold myself to that. If I’m not focusing on that or, if the team quality is no better than the average industry team, I haven’t really taken advantage of the opportunity I had here. I want to help build a team that feels special and has a reputation for greatness within the company and perhaps even outside of it.
  3. Build deep relationships with great people – I said that the people I get to work with was a key factor in any job decision I make. To get the most out of that, I need to really form relationships and not just surface-level working ones. The metric of success here is that I add people to my list of ‘former co-workers I stay in touch with’ and that the caliber of that list continues to grow. I can achieve that by focusing on continually meeting new people. That includes people I work with and folks I don’t work closely with. Leaning into this looks like spending time doing social activities, striking up conversations with strangers and generally being friendly. The difference between a work event and a work event where you are intentionally social and outgoing could be a new contact and a new ally at work. I usually do pretty well at meeting people, but it does take some intentional effort, so I am explicitly stating it here to help hold myself to that.
  4. Get to be a part of an IPO – I’d like to get to pop the champagne and celebrate. In some way it feels a bit disingenuous since I joined the company nine years into its story. But, every day that goes by, I’ll have been here for a larger percent of that story. I’d like to experience that period where people actually know the name of the company you work at for the first time and realize you were a part of that success. Hearsay never got quite that far and Google was a success long before I got there. My timing in joining Databricks sets me up to potentially be able to experience that, whether it is on the IPO day or a few years after when folks see the chart continuing to go up and to the right. My action item here is really just to be patient and put in as much hard work as I can. The more work I do pre-IPO, the better that celebration will feel.
  5. Hit my FIRE number – I have been prioritizing saving for retirement for a while and was on track to be there in ~10 years. With Databricks, that number is up in the air a bit more due to the fact that the job pays less per year but comes with a stock package that might be worth more one day. That is the financial risk I took when joining and if it pays off, it will likely mean I hit my retirement number even sooner. (I will most likely not actually retire, but I’ll just have the freedom to know that I don’t have to work for money if I’m not excited about any projects, or that I can take on a low paying job that I am really interested in without having to make lifestyle changes). There is some tiny chance that the stock goes up enough in the next four years that I hit my FIRE number while on my initial four year grant, but what is more likely is we’ll find out whether I’m slightly behind or slightly ahead of current schedule. My action item here is to help the company and our customers succeed. The more happy customers we have, the more revenue we will have and the higher the stock price will be in the long run.
  6. Continual high growth – I expect Databricks to remain in the top tier of cloud companies in terms of yearly growth. Currently that means >50% YoY but that average will go up and down periodically based on macroeconomic factors. If that stops being true for a long period of time, that will be a signal to me that change is needed, either in the company strategy, the company leadership or my involvement in the company. I’m, of course, not expecting that growth to stop, but I think it is important to call out the importance of that in my decision to join and thus make it a continuous factor in my decision to stay. I think this one metrics actually does a nice job covering a lot of my decision factors – technologist-led, founder-led, bias for action, etc. If any of these factors disappears, the growth rate will likely show it (and if not, maybe its a sign my decision criteria was wrong and needs to be adjusted).

How Has It Been So Far?

So far, so good. I’m drinking from the fire hose, learning a lot and getting thrown a bit into the deep end, but that is what I signed up for. The company has a very strong culture that is slightly different than any I’ve been a part of before, so it is taking some adjustment for me to get used to it and get up to full speed. I’m excited about that culture though and the learning opportunity it presents.

One particular area that I’ve gotten to spend some extra time on is product strategy, which has been fairly unstructured in my past, but has a lot of rigor here. Databricks has formed its culture and practices by looking at the best aspects of other great companies and strategy is one place where I feel they borrowed more from Amazon and Microsoft than from Google. That is probably appropriate as Google isn’t strong on strategy – their success has been less about making amazing strategic decisions and more about doing impossible things and finding a decent way to monetize them.

I’ve also gotten to meet a few longtime Bricksters and to connect with some former colleagues from Google, some that I knew and some that I never met, but know of their work. If this trajectory continues for four years, I think I’ll have a lot of positive things to say in my reflection.