My Journey from Econ PhD to Tech — Part 4: Negotiations part 2 + Decision
Links to other parts of the series:
Google offer
Now that (1) it’s clear I have an offer from Google and (2) there’s real time constraint on both of my counter-offers, Ralph recommended me to directly share my ask with the Google recruiter to speed things up— typically, it’s only after a firm gives you Offer #1 that you share your ask with them. But in this case, since we’re under pressure, we’ll skip Offer #1, and directly go to Offer #2.
At that time, Ralph made a call with me to straighten things up — they told me that, they had a previous client who didn’t pay them anything because he was in exactly the same situation as I was — his first offer was already very high because Ralph advised him to do a preemptive ask, and that is also the offer he signed; since commission is based on the increase from the initial to the final offer according to the contract, they wanted to make sure I won’t do the same. I said, let’s do this: since the highest and only offer I have before meeting Ralph was Uber’s Offer #1, let’s use that as the base to calculate Ralph’s commission.
After that’s cleared, I went ahead to share with Google my ask on Nov 17 Tue without sharing the counter offers. I said (1) I’m asking for $300k/yr (notice that this is lower than the $325k we shared with the other two firms; this is because Google pays significantly less than other tech firms, and the ask is sensitive to the firm’s typical level of pay) (2) this is because I’m a particularly good fit with the team (which is true, because the team does auctions, which is fundamentally a econ problem, and there’s not that many econ phds who became data scientists at google) (3) Uber has given me an offer which I know is top of band (both the recruiter and hiring manager said ‘usually there’s not that much flexibility in our offers; what we did for you is really unusual and shouldn’t be taken lightly’) (4) Zillow’s offer is really attractive because of the remote work option + the pay not being affected by my physical location
To my — and Ralph’s — surprise, the next day, on Nov 18 Wed, the recruiter replied asking (1) there’s nothing the recruiter or the hiring manager can do to my comp, because it’s purely determined by the comp team specialists, (2) she asked whether my ask is based on the other counter offers and how much were they (3) she asked for screenshots of my counter offers.
Ralph is extremely surprised, because recruiters have no right to ask for screenshots of communications between the candidates and other firms, and the standard practice in the industry is that recruiters respond to verbal offers — after all, as a recruiter, you should be able to tell which offer is fake.
(Later on, I found that Google recruiters sometimes ask for screenshots when the candidates’ ask is out of band, e.g. see this post on blind, because ‘otherwise the comp team won’t give more money’. Google is the only firm I ran into that asked for screenshots. )
After consulting with Ralph, who consulted with their own comp team specialists, we decided to share both Uber’s and Zillow’s offer with Google.
After doing so on Nov 20 Fri, the recruiter replied saying she has shared those info with the comp team, and will keep me updated on any updates. I emphasized with her again (1) I have a hard deadline on Nov 25 Wed (2) those are offers I cannot risk losing (3) if they can’t give me an offer that makes sense for me financially I’ll go with the other firms
On Nov 23 Mon, I sent the Google recruiter an email to check in, but received an auto reply saying she is out of office. Given how urgent things were, I sent an email to the hiring manager explaining the situation and asking for help — but it turns out that the whole team were out of office for that whole week (it’s Thanksgiving on Thur that week)
On Nov 24 Tue, i.e. 1 day before the Uber deadline, the recruiter finally called to give my Google Offer #1 aka #2: $142k base, 15% annual targeted cash bonus, $340k new hire equity, $50k sign on paid in year 1, which brings the first year total comp to $298.3k, and average annual comp to $260.8k/yr for the first 4 years. She also mentioned that, although historically there hasn’t been data science jobs in Google’s Chicago campus, in 2021, the Search Ads team might open up some positions in Google Chicago, and I can internally transfer to it.
Final negotiations with Google
Given how urgent the matter is, Ralph immediately gave me an advising call.
They asked me what I think. I told them, I’m honestly torn, because I still don’t know which firm to choose.
They asked, ‘what’s your career goal?’ I told them that my goal is to have my own start-up in 5 years, and I see the Google offer as a chance for me to broaden my horizon, and the Zillow offer as a chance for me to make impacts (more on this later), so honestly both can be a good prep for my start-up career.
They asked me, what do you think you’re lacking the most at the moment? What do you want to learn the most in the next 5 years?
I said, product senses. I don’t want to be a technical co-founder, but rather, I want to, and like to, think about the product side of things, and that’s what I want to learn the most in my day job.
Then they asked me, which one do you think can give you that the most?
I said, probably Google, because at Zillow, I’m closer to a research scientist — they are hiring me for my domain knowledge in housing, and they have a well-defined question, ‘how to price homes?’ and would like my help solving it in a scientific way. But at Google, based on what I’ve learnt by talking to my hiring manager and teammates, I’ll be in charge of 1–2 specific products, propose ways to improve the products, and work with engineers to implement them.
At that point, it’s clear what the choice is. I was really glad that they were there to walk me through — there has always been two sides of me — part of me is interested in intellectually provoking questions, and another part of me wants to create and build things. The PhD was 4 good years of time for me to indulge in interesting intellectual questions, but the reason why I’m leaving is precisely that I know I don’t want to do only this for the rest of my life. The reason why the choice was so hard was that the Zillow offer is a perfect project for a scientist — figuring out a framework to price homes, how cool is that? But the problem is, it is not what I came to tech for.
Now that we’ve decided on the choice of firm, the rest is easy — just do an ask and sign. I sent the Google recruiter an email on the same day, i.e. Nov 24 Tue, saying, thanks for the update, but can you make the offer $10k/yr higher? If we can get there, I’ll sign.
Just after I sent that email, Ralph reached out to me to say we missed a very important piece of info: if I relocate to Google Chicago, there can be a very big pay decrease, and I need to ask about it before I sign.
On Nov 25 Wed, i.e. decision deadline day, the Google recruiter called to give me the updated offer, i.e. Google offer #3: $142k base, 15% annual targeted cash bonus, $348k new hire equity, $50k sign on paid in year 1, which brings the first year total comp to $300.3k, or the average annual comp to $262.8k/yr for the first 4 years. In other words, they gave me a total of $8k increase, in response to my ask of a $10k/yr increase, which honestly was a bit disappointing to me. And the recruiter also said, based on their historical data, Google Chicago’s pay is 17% lower than Google Bay Area’s pay, and, ironically, their offer is also 17% higher than Zillow’s (I’ve told them about Zillow’s remote option). She also emphasized that, the internal transfer is merely an option, not a promise.
After the call, I did the calculations, but Google’s offer was not 17% higher than Zillow’s — I’m sure what Google did was that they did not count Zillow’s annual targeted equity refresh, but to me, the fair comparison is to include the part that’s vested in the first 4 years:
Note: For Zillow, for the new hire equity grant, although they quoted $180k, what’s actually written in the contract is the number of RSU, and the number of RSU was calculated given the stock price at the time of the offer, i.e. $95.24, and given that at the time of the negotiation with Google, Zillow’s stock price has already reached $109.33, the RSU was worth more than $180k — that’s why I had two versions of the calculation. And on the vesting schedule of each of the equity refresh — it’s the same as how it’s written in the offer, although I do not remember why it’s not linear.
After learning about what happened, Ralph said we can’t sign as it is. I wrote back to Google’s recruiter saying, (1) It’s very unfortunate that the Chicago campus’ option is not guaranteed, because it’s a very big part of my decision and (2) Google’s offer is not 17% higher than Zillow’s, and I can’t sign it as it is. Not surprisingly, the recruiter didn’t reply further — it’s already in the afternoon of the day before Thanksgiving. I told Zillow that, unfortunately I can’t give a decision today, but I can next Mon or Tue.
After a not-so-relaxing Thanksgiving break, on Nov 30 Mon, the hiring manager finally reached out to me — he saw my email which was sent a week ago asking for his support, and said he’d be more than happy to help. I explained to him the situation, and said I’d like his help advocating for me with the recruiter. He said he’ll schedule a call with the recruiter right away.
After a few hours, however, he reached out to say, the recruiter seems to be out of office, as all her calendar that day was blocked. I thanked him and expressed how much his help meant to me, and asked if the best course of action now was to wait for the recruiter. He replied saying he will try again the next morning.
On Dec 1 Tue, the hiring manager finally got in touch with the recruiter, but he didn’t bring back good news: he said, the recruiter said it’s already a top-of-the-band comp, and it’s unlikely they’ll move further. The recruiter also later replied me confirming this. Most importantly, she shared the comp team’s calculations:
Their rationale is that, Google (like other major tech firms in the Bay Area) also offers annual equity refresh, but they don’t reveal the actual amount in the offer (because it’s purely performance based), so they are also not counting Zillow’s $35k annual equity refresh in the calculation. To me, what’s a fair comparison is to count everything written in the offer, because Zillow’s $35k is a ‘target’, meaning that if someone does well, she/he can get more too, same as the unspecified amount of equity refresh from Google. But, we can only agree to disagree.
The recruiter also said that, the base and the annual targeted cash bonus is tied to the level, and the $50k sign-on is the max possible. Also, they only match counter offers, i.e. they don’t raise beyond what’s written in the counter offer. Ralph also said, this is the highest Google Data Science offer they’ve seen. Upon knowing that, I said yes to Google, and concluded my recruiting journey.
Decision
The reason why this was such a hard decision for me is that all the opportunities are attractive yet very different:
Uber
It’s not an exaggeration to say this was my dream job before I started the job hunting process.
For those of you who are in tech, you must be surprised, because Uber as a firm isn’t doing so well, especially during the pandemic, and it had culture problems a few years ago.
But for an econ phd, things are completely different:
- Where are economists most influential and the econ skillset most valuable? At two-sided (or three-sided) marketplaces. Who are the biggest marketplace tech firms? Amazon and Uber. Yes, there are a lot others, such as Lyft, Doordash, Instacart, but for historical reasons, Economists are well regarded and have a lot of influence at Amazon and Uber, not as much in the other three firms I just mentioned. This is because, at a relatively early stage of Amazon and Uber’s development, some economists were able to enter, show the upper management their value and gain a lot of credit. As a result, economists started having influence in Amazon and Uber, and these two firms also started systematically hiring economists — they go to the AEA to hire. (To see that it’s truly path dependency: in principle Uber and Lyft should be very similar, but data scientists at Lyft tend not to be economists, whereas at Uber they do.)
- So among Amazon and Uber, why do I think Uber is better than Amazon? Based on what I’ve learnt from talking to people at both firms, at Amazon, Economists are more like consultants — product teams come to them with a specific problem, e.g. what’s the causal impact of X, and they solve that problem and write a report/present the findings. But at Uber, we are more in the driver seat, making changes in the products — there are brainstorming sessions where PM and us both attend, and come up with product ideas, and we can take the initiative to implement them. In addition to that, because Amazon hires a lot more than Uber does, the bar at Uber is also higher on average, which means the average quality is higher, which is important if you want to be surrounded by good people you can learn from.
- At Amazon, because of how big the company is, you might not end up working on a team that directly deal with two-sided marketplace problems — for example, you might be assigned to AWS, or devices, or even HR. If the whole point of joining Amazon is that I’m interested in two-sided marketplace, then that goal isn’t achieved. However, at Uber, you know you’ll be working on two-sided market, because that’s the only problem they work on. And the problem Uber’s solving is just absolutely fascinating, at least from an economist point of view.
- Because Uber is still younger and smaller, progression at Uber is still faster than bigger firms like Amazon or Google. And because of how much influence economists have, and how well regarded they are in the org, progression for economists are even faster than average. Some people worry about Uber’s layoff culture, but at some level, you can’t both ask for merit-based promotion and job security — if you want to rise fast, then there has to be layoffs to weed out the underperforming ones (the opposite is Google, where job security is very high, and promotion is very slow). As a result, if you know you’re good, then Uber’s system might be better. (At some level, this is similar to the Chicago style of econ grad school.)
So all in all, if I have my economist hat on, Uber would be the dream firm.
What are the cons? Well, the firm isn’t doing so well, and the product is relatively mature — of course, there are always things to work on, but compared to the kind of work I’ll be doing at Zillow, both Google and Uber’s projects are marginal.
Zillow
Although Zillow the firm is old — in a Silicon Valley VC’s words, it’s real estate tech 2.0 instead of 3.0 (i.e. Airbnb), but Zillow Offers is new — the founder of Zillow also founded Glassdoor and Expedia, and his idea has always been ‘empowering consumers with information’, but recently he realized just providing info is not enough to disrupt the real estate market, and decided to do so by directly buying and selling homes. …To be fair, Opendoor was the first one to do it, and Zillow probably created Zillow Offers in response to the competition (like facebook creating reels in response to ticktock…?).
In any case, Zillow Offers is a relatively young product — the first day that my hiring manager joined Zillow was the day when Zillow Offers was launched, and according to LinkedIn he has been at Zillow 2 yrs 9 mnths as of today, and he is already a manager. This is a very positive signal because:
- The product is young to an extent that I’ll be working directly with people who experienced the launch of that product. This is completely different from more mature firms/products such as Uber — my hiring manager joined after ‘Marketplace’ was launched, and not to mention Google — Display Ads has been around for at least 7 years or so.
- My hiring manager was promoted twice in 2.5 years, which is extremely fast — at Uber, the first promotion happens around 1.5 years in, and at Google, around 2–3 years in, and for the second promotion, nobody knows — it depends on whether you got lucky enough to be working on a high impact project. But because Zillow Offers is so young, there are a lot of big open questions still to be answered; and because it’s such an important part of the ‘Zillow 2.0’ vision, it has a lot of visibility among management
- I’ll be doing more foundational, rather than marginal, work. If I have to make an analogy between Uber and Zillow, then what I’ll be working on at Zillow — figuring out the framework for pricing homes, is like building the major products BYOQ (build your own quest) and CT (continuous trips) at Uber. At Uber, these major products have already been built, and what people are working on now is how to optimize/consolidate them, rather than creating more new foundational products. Before these products were built, or before Marketplace was created, pricing and incentives were created very manually, e.g. by operations teams in excel — that’s the stage Zillow Offers was at a year ago or so: at the beginning, they were using analysts to price homes manually, and they have not so long ago started to use ML to solve this problem, and even more recently started thinking about using domain knowledge in the housing market to impose structures on the problem, which is what I’ll be doing if I join
But the cons of Zillow? I’ll be working in the housing market — a specific topic, and I’ll be a Scientist, instead of working directly on products — I did ask people who work on the team about it, and my impression is that the PM and the manager shelters the Applied Scientists from the product side of things so they can focus on solving the scientific problem, which is actually not what I wanted.
Quora
Quora’s data science has always had a good reputation — to put in human language, people in tech know that Quora’s data science is good. As a result, going there means (1) you’ll have a good CV mark and (2) you learn from some of the best DS people in the industry. DS also has a lot of influence at Quora, which you can tell by how many DS there are — Quora the firm has ~200 employees, and ~15 DS — that’s a lot of DS! At Uber, I’ve heard that among a total of ~20000 employees there are ~1000 DSs, which is also unusually high, but still lower than Quora’s ratio — and the opposite is Google: with a total of ~100,000 employees, there are ~400 DS… which is a signal of how much each firm thinks DS is useful
For start-ups, a firm’s culture/the type of people at a firm is heavily influenced by the founder. Quora’s founder Adam was the technical co-founder of Facebook — people said the reason why Mark moved to CA is because of Adam. Adam is also a silver medalist of the International Olympiad of Informatics and undergrad at Caltech before co-founding Facebook — in short, he is smart, and as a result, people at Quora are also very smart — I could indeed feel that during the interviews. I’m not saying people at other firms are not smart, but the feeling I got was very different — Uber people are a lot more businessy/street smart, while people at Quora are more nerdy/book smart.
Also, a lot of DS at Quora mentioned to me that, the CEO would directly comment on your work, or cite your work in meetings, and this is guaranteed to not happen in bigger firms like Uber or Google. In other words, you have more exposure to upper management at a small/lean start-up like Quora.
The cons? (1) Personally, which is also consistent with a lot of people’s view in Silicon Valley, I’m not optimistic about the future of Quora — it’s just not very mainstream (like Reddit), and its monetization isn’t as good as Facebook — Facebook revenue per user is $10/yr, and Quora’s is more like $1/yr… indeed they’ve been doing it for less time, but the trend isn’t looking great (2) I personally am not the most interested in feed business (e.g. Facebook, Quora, Instagram, Tik Tok, etc.)
When I applied for Google, it’s more because I’m looking for a job in tech in general, and less about me excited about going to Google per se. Why? I’ve always heard that anyone who’s not an engineer is a second class citizen at Google, which makes sense — Google became Google because of its technologies, not because of its PMs or DSs — in fact, by the time Google was founded, there’s probably no such thing as ‘data science’ yet.
Also, at that time I was more interested in Economist positions, but Google’s econ team wasn’t hiring this year, at least according to EN. And their Data Scientist position? Just reading the job description makes me realize it’s not for me — ‘PhDs in quantitative fields such as statistics, bio-statistics, mathematics, physics, etc.’ — econ wasn’t even in the set! It’s pretty obvious Google DS is ruled/dominated by Stats PhDs — just like Uber DS being dominated by Econ PhDs and Quora DS being dominated by stats phd + people from good US undergrads.
Despite that, I applied, and somehow got the phone screen, which really changed my view. The interview was extremely fun — it was by far the hardest phone screen, for how non-standard it was — it reminded me of when Anthony first met me and started quizzing me with fun math/physics puzzles — it’s a similar feeling with Google’s phone screen (as I mentioned earlier, a friend of mine who went through the same process had a very standard phone screen, so this might be an accident). I was very impressed by how much effort they put into prepping the interview questions, and how interesting the questions were, and was very intrigued.
The onsite honestly wasn’t as impressive — I’d say it’s an easier version of Quora’s onsite — also basically just metric definition + experimental design/analysis problems, plus data manipulation in Python/coding, but less hard. Through it, you can also realize what the DS at Google do most of the time — most of the time is spent on metric definition, i.e. thinking about how to measure something, and help engineers design experiments and analyze experiment results. — honestly not what I’m interested in doing the most.
But after the team match came out, things changed — my team, i.e. bidding at Display Ads, is one of the very few teams at Google where DS plays the central role and where DS leads engineers. Auctions is fundamentally an econ problem, and my manager and his manager are also both economists. The DS in the team are the ones proposing improvements in the algorithm, and leading engineers to implement them. Also, the way work is structured is that, every DS owns a few products, which is what I wanted.
Decision Process
Stage 0: At the beginning I was super sure I’ll go to Uber, and after Uber’s offer came in, I was almost going to just sign it, but after learning how low the offer was, I decided to keep interviewing to get counter offers to get higher wages
Stage 1: I was honestly impressed by how interesting the Google phone interview and the Zillow onsite was, and started re-thinking. Especially after learning that I’ll be helping to build the pricing framework for Zillow Offers, I was more intrigued — it seems I’ll be having a lot more influence at Zillow than at Uber. But the problem is — I’m just not so interested in ML, and at Zillow, I’ll still be mostly doing ML, just through an economist’ lens.
Stage 2: After Google’s offer came in, I became divided — what do I want? What am I looking for? I started soul-searching, and realized that what excites me the most as my career goal is to have my own start-up. Upon learning that, I started asking around ‘Given my goal of having my own start-up, which firm should I pick?’
From there, I narrowed down to Google and Zillow, because Google’s brand name is better, has a wide range of products, and a good network of people, and at Zillow I’ll be taking on more responsibility, work on earlier stage product, and rank up faster. (The reason why Uber was out was because, if I want to keep being a tech economist, Uber would be best, but that’s not what I want anymore. Quora was out because, if I do want to join a start-up, I would join somewhere like Stripe or Robinhood or even Airbnb, which is the latest generation of successful start-ups.)
Honestly, I can see a path towards entrepreneurship whichever one I choose:
- If I join Google, I’ll use my ~2 years of time there to learn about the products at Google — Google does a wide range of things, and Google’s environment is very open which makes learning about other teams’ work very easy. I’ll also use the fact that I’m in the Bay Area to network with a bunch of start-ups and learn about the landscape: what are the start-ups that are being done, what worked and what didn’t, and what haven’t been done yet. From there, if I run into a start-up that I resonate with really well, I can quit Google and join them to work for ~2 years. And then, I can start my own business
- If I join Zillow, I can spend the next ~2 years learning about the real estate tech space — I can even get a real estate agent license, and get into the ecosystem. With the remote option, we can also speed up having kids (Anthony is in Chicago, so if I join Google we’ll definitely not be having kids until we solve the colocation problem), and have 2 kids while I work at Zillow. Then, I can join a real-estate tech start-up, and eventually start a real estate tech start-up myself — I’m honestly quite interested in real estate tech in general, so I wouldn’t be too annoyed if I spend the next 5–10 years in this field
If I can see a path towards my goal whichever I choose, what should I do? Flip a coin?
At that time, I talked to a Silicon Valley VC whom I had known since my undergrad years while he was an adjunct professor at HKUST. He shared a very insightful framework for thinking about decision problems: what you would learn if your expectations came true VS what you would learn if your expectations did not come true for option A VS option B.
This made me realize one important flaw in all of my thinking/planning above was that, it was based on a thousand layers of assumptions — What if after I joined Zillow, I very quickly learnt that real estate tech isn’t a great field for start-ups? What if after I joined Google, I realized learning about other teams’ products or networking in bay area isn’t so easy? After all, it’s all expectations and you don’t know what you don’t know.
Putting on my economist hat for a moment: In a life cycle decision problem, the agent optimizes the sum of utility over all periods in her/his life, which means the more number of periods left, the more important the term after the expectation (‘E’) sign is, and the more important information acquisition is. I’m still just at the entrance of the start-up world, and even though I was able to learn quite a bit over the past few weeks by talking to ~100 different people who are at different positions in the ecosystem, what I’ve learnt, compared to what there is still to learn, is negligible (the Chinese saying 九牛一毛 captures this so well — ‘a hair compared to 9 cows’). So really, I should accept/expect that whatever decision I’ve made at this point is sub-optimal, because of how little information I have, and what I should prioritize now is getting more info.
Thus, I decided to go to Google, because it’s the option that allows me to get information the most — of course there is assumption going into that also, but that’s the most likely compared to others. Put not so nicely, I’m ‘kicking the can (i.e. decision making) down the road’, but that’s the right decision given where I am now.
Side note: It’s completely possible that, in 2 years time, I realized that real estate tech is the best field for me to do a start-up in, and I go back to Zillow. But even if that happens, my choice today is still the right one — I made the decision to acquire more information, for me to make a more informed decision later. When evaluating/reflecting on decision making, a common mistake people make is mixing up info you have now VS then — what you should reflect on is not so much the decision, but the decision making process.
