A Framework for Decision Making in Career Development

(Reward, Risk, Regret | Constraint) X Time

Image Source link

I have engaged in quite a few interesting discussions recently on decision making at various stages of career development. Topics may range from jump-starting the first job to how to sustain growth at senior level. Through these on-going discussions and thinking, I found it important to construct a framework for the decision making process, especially when a lot of noise and uncertainty are present. Such a framework can be utilized in many kinds of decision making in day to day life, big or small. But today let us focus on career development primarily.

We seek answers to the questions like the following:

• Should I join an early stage start-up or a prestigious company?
• Should I spend a couple of years to obtain an advanced degree, or should I join the workforce and learn by doing?
• Which field should I better invest in my time?
• Which industry should I join?
• Should I optimize for short-term growth, or should I prioritize more for the long term vision?
• Should I do something I really like but with a lower compensation, or something less interesting but may be financially more rewarding?

And the list goes on. Without further ado, let’s begin!

Let’s assume in each decision making scenario, you have ≥2 options to choose from. The high level idea is to evaluate the outcome of each option with a framework such that they become more comparable, hence an optimal solution maybe obtained via reasoning. The framework I will explain here consists of three vital piece: Outcome, Constraint and Time. I will explain the concepts first, then we can go over a couple examples.

For each available option, we can think about the following:

Outcome: Reward, Risk and Regret

First of all, what is the direct Outcome resulted from a certain option? More specifically, what are the Reward, Risk and Regret associated with this option?

• Reward very straightforward — it is the direct positive outcome resulted from taking an option, which can take many shapes and forms: financial, academic, skill growth, social capital, impact, etc. For the same option, the value of each reward can be different to each person — early stage professionals may prefer more experience growth while seniors may favor financial or impact. Reward is usually the first thing we think about, but it’s important that we consider the totality of the rewards but not just looking at a single dimension like compensation, and also think about how rewards may change overtime, which we will discuss a bit more later.
• Risk is the direct negative outcome associated with this option — it can be microscopic or macroscopic. Do you have the necessary skills to evolve with the position? Will the company you are about to join survive competition with the emerging disruptors? Everyone has a different perception and tolerance of risk, and it’s critical to have clarity about the risk so one can have adequate preparedness should things go south. People also tend to be over optimistic about rewards and underestimate risks, so it’s a good practice to assign equal, if not more weights, to estimate risks in the mental model. Whatever can go wrong, will go wrong.
• Regret is a valuation of the opportunity loss from NOT taking a certain option, and it can be evaluated on the same scale of Reward. What have you missed in terms of monetary or growth opportunities by choosing an alternative? If the rewards and risks from two options are on par, weighing in the potential regret may often lead to tie-breaking. Regret is also the disruptor, or the X factor in this very quantitative and rational framework. It can make the final call to overrule all the other analysis and reasoning. You may be so enthusiastic about a job that doesn’t provide good financial outlook in the short-term, or a start-up doesn’t look very promising at all. But the question you may ask is: if not now, when? The seemingly irrational decision-making could be the sole factor that differentiate humans from machines.

Once an option is mapped to Reward, Risk and Regret, the next thing to do is to estimate the uncertainty aspect. Specifically, I tend ask myself the following questions:

• What is the Best-case, Worst-case, and Expected(Average) outcome of Rewards? These values also set the basis for Regret, as the values you will miss out if an alternative was chosen. Since the distribution of Rewards is not typically Normal, it’s important to understand not only the average case, but also the extremes. Joining an early stage start-up may result in lack-luster expected return, but extremely high best-case return and very poor worse-case return, whereas the reward from a mature company may have much less variability. Again, people often tend to anchor on the best-case reward from a hyper-growth company, so be more disciplined about the entire reward structure will lead to a more objective, level-headed perspective.
• What is the Worst-case, or near Worst-case outcome of Risks? Different from Reward, Risks tend to be rare (hence named risk in the first place and not the first thing to consider), so understanding the average-case risk may not be helpful. What’s more important though, is when indeed the rare event happens, will you be physically, mentally and financially prepared. A start-up’s valuation may go down to zero (90% case), a mature and lucrative industry maybe disrupted overnight (taxi), sustained prolonged working may lead to health problems (burn-out), investing time to learn a skill that may be obsolete very soon — understanding such risks and mitigations can counterbalance the thinking in Reward, especially when the rewards are comparable. Are you willing to take 10% more reward, maybe at the cost of 50% more risk? Maybe not.

All this may still sounds a bit confusing, and I will explain a bit more about these concepts later with an example. But please bear with me for now.

Constraint: How Much Can We Explore?

The second ingredient is Constraint, which determines whether an option is viable at all, or how much flexibility there is for a certain option. If certain constraints prevent us from taking an option, are there anyways to unblock? Constraints can be external or internal, and some are harder to remove or stretch than the others. An example of external constraint is the years of experience required for a certain position —it could be quite challenging for a new grad to apply for a senior level position straight out of school. For internal constraint, sometimes a minimal amount of income is required for financial liabilities, which makes joining an early stage start-up difficult since the majority of the compensation may be paid in equity. A certain skill may be needed to perform a task (such as machine learning, or managerial experience), which one may not currently have. Even it may seem impossible to take a certain option due to constraints, it’s still a good exercise to evaluate the outcomes of that option so one could be better prepared next time.

It should be noted that constraints should not be considered immutable — it’s the first line in the sand to show where the limits are. With a certain degree of efforts constraints may be stretched, at some cost. For example, even starting from a senior level for a new graduate can be very difficult, it’s not entirely impossible if one already gained enough experience at school from internships and entrepreneurial activities, and possessed superb technical capability and maturity. We all know 20 year olds can be CEOs, and I have had a stunning colleague who never even went to college. If joining a start-up may lead to oversized returns (such as IPO or acquisition is very likely on the horizon) and only moderate risks, one could live on debt to support family for a while to seek long term gain, even the current financial constraints suggest otherwise. Again, the constraints provide another orthogonal dimension to decision making, complementary to the aforementioned Outcomes.

Time Horizon: Short Term v.s. Long Term

The last piece to inspect is Time. How will the Outcomes evolve overtime? Will the reward and risk grow disproportionally such that a good choice today may become less desirable in a few years? This long-term thinking will also guide us to take a less optimal path now in hope for a much better outcome later. If every decision is based on maximizing near term outcome without long term vision, one could easily win every single battle but lose the war. An example is job hopping early on — it’s tempting to hop around every year or two to get a 10% ~ 20% raise in monetary reward. However such constant context change will come with adverse effects: shallow technical growth, lack of in-depth understanding of the industry, poor social capital, etc. A steadier growth could lead to quantum leap of impact, scope of coverage and compensation down the road, with lower long term risks and regret. One step back may lead to three step forward.

Even though long term thinking is generally more beneficial, it’s worth noting that a good balance between short term and long term is essential — long term is irrelevant if short term survival is not guaranteed. It’s tempting to invest considerable time and resources into big, massive endeavors, such as taking a pause from the current career to pursue a new degree, or for a company to over-invest in a new technology which may not generate meaningful return soon enough. Understanding the trends of Reward, Risk and Regret, as well as the evolution of Constraints, can better help strike a balance between short-term v.s. long term. return of investment.

Summary of the Framework

Here’s a brief summary before we move on to play some scenarios. For each option, we consider:

• Outcome: Return, Risk, Regret (R3). Think about the uncertainty of each dimension — best-case, worst-case, and expected(average) case.
• Constraint: What are the limits and flexibilities in each option, and how you may stretch or mitigate such constraints.
• Time: How Outcome and Constraint may evolve over time, and how to strike a balance between short-term and long-term.

Together, I hereby name it: #R3CT# (…Wrecked? :D). Once this exercise is completed, we can then evaluate each option against all other alternatives to find the best decision that maximizes Reward, minimizes Risk and Regret within a reasonable Constraint that yields best short-term and long-term outcome.

Now… let’s run through a couple of examples to see how it may work. I will start with a very simple scenario, and go through a situation which is quite typical in career progression. Let’s start with … food?

Image: thewoksoflife

Case Study 1: Should I buy Salted Duck for Dinner?

Despite seemingly trivial, I have had numerous debates with my wife whether I should buy salted duck for dinner. It’s a very traditional inexpensive delight very popular in my hometown, but to her it’s overpriced and only serves the purpose of feeding my inner gluttony. So every time I go to the grocery store and contemplating whether I should buy a package of salted duck home for dinner, I run a quick R3CT analysis in my head. Formulated as:

• Option A: purchase a half salted-duck without thinking
• Option B: proudly walk away knowing I really curbed the craving

So what to do? First of all, let’s look at the Outcome:

• Reward: I will feel great tonight at dinner, enjoying the salted duck and indulging in nostalgia about my hometown. We can cook one less dish to save some kitchen time, which may be invested elsewhere (most likely, random time-wasting on the phone). There’s a bit of variance in the reward of the taste; since the process of producing this dish is not quite consistent, sometimes the flavor is right on and other times it can be quite off. But on average I reap a good return.
• Risk: My wife may be quite upset that I made the decision without consulting her first since she always owns the shopping list. If she’s in a good mood today then my risk is lower. The highest penalty I ever had is a complete ban of buying another salted duck and no video games for a full length two weeks, while prioritizing all her tasks in our house to-do list. There could also be food poisoning but it’s very unlikely. So the worst case scenario is a delayed penalty that I have survived before.
• Regret: Regret here is a function proportional (or maybe exponential) to the time since last I had this dish. If I haven’t had it for a long time (such as a month), then giving up the opportunity to buy could generate oversized regret. On the flip side, if I just had salted duck the day before then the regret would be much smaller. If I missed when it was on sale or the last box, the regret could also be quite substantial. So in any case, regret is contextual — but generally the longer I haven’t had this dish, the higher risk I would be willing to take, and the higher perceived reward there is.

After this part is done, I look at the Constraint:

• Since I am the only person besides my toddler daughter who really craves this dish, my wife would only allow me to buy from my personal allowance, which is the most obvious constraint. Since I have a fixed budget, so sometimes I have to choose between a salted duck or new screwdriver from Home Depot. I can always attempt to stretch my allowance, but that’s a move that I typically reserve for other gadgets and the reward from the salted duck probably doesn’t justify. So if I have enough allowance especially at the beginning of the month, the constraint usually doesn’t matter too much into the decision making.

Finally, I think about Time horizon:

• Buying a salted duck today definitely benefits short-term — I know I will really enjoy it later at dinner. However since there’s a chance my wife may not be very happy, I really have to weigh in the other parameters. I have to think about how much other purchases later in the month with the remaining balance in my allowance, the bargaining power I could gather by informing my wife that I walked away from the salted duck stand and proudly restrained myself to make an irrational decision of feeding my inner crave without thinking, and how that leverage may benefit me in later negotiations. Missing salted duck one time could potentially bring in numerous long-term benefits.

You can see that the decision can be different every time based on the contextual analysis. This thinking process may seem to be an overkill for such a small decision, but once the brain is wired to function this way, it’s not actually that difficult to go through this analysis very quickly. Like I mentioned before, sometimes I simply just want to have the salted duck at dinner and the regret from not buying could be so large, that it simply overrides every other rational thinking (hence regret alone, dominated).

If you are interested in this Nanjing duck dish, it even has its Wiki link and a recipe. Oh well, sorry ducks…

Source: Crazy Engineers

Case Study 2: A Career Decision

Let’s construct a more real-life like scenario — say you have worked very hard and have two job offers on the table:

• Option A: Local startup, good outlook, very exciting work. Annual compensation is 100k base salary + 150k stock options, senior title.
• Choice B: Prestigious public firm, requires relocation, large team, less interesting work but market leader. Annual compensation is 100k base salary + 100k RSUs, entry title.

At the first glimpse both are viable options. An early stage start-up provides tremendous growth potential at a higher risk — only 1 out of 100, even 1000 startups made it with an exit. A prestigious firm offers a much larger platform and learning opportunity from experienced professionals. Job security and stability is presumably better, but it may be more difficult to drive larger impact with good visibility. The compensations are comparable; startup equity can easily multiple if things go well, but it can also reduce to paper. The RSUs from the public company may not shoot through the roof, but still compound at 10 ~ 15% every year.

Yet there are other factors to be pondered on. Will you become acclimated to the new environment after relocation? Building social connections will take time, both professionally and afterwork. How will the market landscape shift in a few years? Will today’s leader remain competitive, or got disrupted and rendered irrelevant? Will you regret in a few years that you favored stability over passion, and asked what would have happened if you chose the high risk path? On the other hand, if the start-up didn’t work out, will there be enough reserve to endure the crisis? The list can go on and on. You can see without a framework things can get murky very quickly. How can R3CT help here? Drumroll…

First of all, let’s evaluate the outcomes. Let’s use the Startup and the Firm to represent the two options, respectively.

Reward: There are several aspects of Reward. Let’s first look at the monetary reward in detail:

• Let’s assume the base salary will grow at the same average rate of 10%, to make things simple. For the Startup, without an exit, the expected compensation at year 5 is 160k, base salary only (worst-case). With a good exit, let’s say the options doubled, so 150k stock options X 5 years X 2 = 1.5Mil total realization, ~ 300k each year (average case). A best case scenario may quadruple the stock options, netting an average 600k a year. This is not common but not rare neither — an example is Square’s stock rally from 2017 to 2019. Stock appreciation could have lead to over 1MM annual compensation, if joined at the right timing. So at year 5, the startup could generate 160k-460k-760k annual compensation [worst, average, best]. Note that the average and best case are contingent on a successful exit; if an exit is not immanent then we need to forecast on a longer time horizon.
• On the other hand, the Firm’s same YoY growth will also yield 160k salary. Let’s assume the RSUs will also appreciate at 10% each year, so by year 5 on average all RSUs from year 1 will worth 1.6X. The total comp at year 5 would be 320k. We can assume a better case is appreciation of 20% annually, and this will lead to about 410k total comp. With only inflation-level appreciation (worst case), the total comp would be around 270k. Note there can be refreshers, bonuses … which will all be liquid rather than the paper money from the Startup. So we can add a 10% baseline lift, with 300k-350k-450k annual compensation [worst, average, best].

We can see that while the upside of the Startup (760k) is wildly better than the Firm (450k), the downside is also much worse (160k v.s. 300k). The average case of the Startup is better (460k v.s. 350k), but contingent on a reasonable exit. There are other factors to consider — since the RSUs are liquid, it can be divested into other stocks at higher growth rate, thus further push the baseline of the Firm and reduce the risks even further.

Without thinking this through, one maybe overly excited about the upside of the monetary reward but overlook the downsides.

The non quantitative rewards can be evaluated similarly. Since you could start at a senior title in the Startup, if the business is doing well then in five years it’s possible to elevate to senior management. Average case could be a manger of a team ~ 5 people, and worst case may be no title growth. So Senior IC-Manager-Senior Manager could be considered as the [worst case, average, best case] outcome. For the Firm, job ladder will be harder to climb. If entry level is not terminal, the worst case scenario could be involuntary churn. But let’s keep it simple by assuming it’s technically possible to remain on the same level in 5 years. Average case may be Senior IC and best case could be Techlead — Manager. So IC-Senior IC-Manger could be considered as the [worst case, average, best case] outcome. The scope of work and impact to work in the Startup and the Firm are obviously different; but at least we can somewhat compare career growth on the same scale this way.

We can continue to layout similar thinking on other rewards: domain-specific learning, technical growth, professional networks, impact, market value…but I will leave this out for you, maybe as a fun exercise.

After Mapping out Rewards, we can then move on to evaluate the Risks. You may ask some questions like the following:

• What’s the probability of a successful exit in 5 years? Obviously no exit will lead to the worst case scenario, but if you are prepared to start-over again in five years, then it’s a risk that can be absorbed. External market factors should also be considered; big firms with more cash reserves are more resilient than startups.
• Will the experience you earn in the next 5 years become increasingly more valuable, or irrelevant? The technological innovations can be very disruptive. You could work for the Firm on an older tech-stack, only to find yourself becoming less relevant with others moving onto the next best platform. On the other hand, the Startup could be very nimble and you could have exposure to the newest technologies, or you could gain experience that will be in high demand in the near future: next-gen healthcare, remote education, subscription business model, etc., hence less overall career risk.
• Can you handle stress from the Startup? I have been there, and working around the clock can be emotionally and physically draining. It’s rewarding but definitely not for everyone. The risk of health deterioration due to constant stress should not be overlooked; there’s nothing wrong seeking WLB, if that’s the most productive way both for work and life.

And many more.

The Last Piece of Outcome is Regret. The most straight forward regret is the opportunity loss: if you chose the Firm, you may be missing out 300–400k compensation in the 5th year alone! Besides monetary miss, there aren’t many chances in our lives to join a startup and work hard to make it work. If the Startup really took off, you could have a very deep regret that you missed the train to extraordinary professional and financial rewards. If the Startup did not make it to the finishing line, you may wonder what could have happened if you joined the mission. The counterfactual is unobservable and there’s always going to be some regret — you could attempt to put yourself in your future selves shoes and weigh in. As mentioned before, regret could just disrupt all the reasoning and drive to a decision by following your heart (with manageable risk, of course).

If you have read all the way here, I hope you could formulate the constraints and time on your own, which are already covered somewhat in the previous sections. While the growth in the Firm can be quite linear, the growth in the Startup can be exponential. Different weights on short term v.s. long term could lead to very different reasoning with the same circumstances.

Now with all these thinking — do you have an answer in your mind now? Either option is a good option, and you are very unlikely to regret years later, knowing you already tried to minimize it.

Final Summary

The framework may seem to be quite convoluted, but once you are used to this way of thinking, it will happen very quickly just like riding a bicycle. I find myself adopting this type of thinking quite naturally even in small daily decision making — maybe an overkill but it’s fun. It allows me to evaluate each option rationally (Reward), manage rare events (Risk), follow my heart (Regret), and explore each option (Constraint) with an evolving perspective (Time). Hope this is not a waste of your time, and happy decision making with #R3CT#!

Notes and Disclaimers

• Karthik Chandrashekar raised an interesting meta point of the decision of decision making: given one has already committed to a long term strategy, how often should re-evaluation happen? For example, you may have decided to join the Startup, but you still may want to re-evaluate in the course of the journey. There may not be any external factors to trigger an re-evaluation, so technically you can do this exercise as often as possible. It’s obvious that thinking about this daily is inefficient and wasteful since there’s probably not too much change everyday (in Bayesian thinking, the newly collected data didn’t add meaningful value). Overthinking may also lead to exhaustion and unnecessary questioning about the current course (similar to peaking may lead to false positives in standard A/B tests!). So it’s imperative to build proper discipline to only re-evaluate either by fix cadence like every six months, or meaningful event-driven (re-org, acquisition, etc.). It’s almost like there need to be a resource management, or load-balancer, to only decide to re-evaluate when only appropriate.
• After finished this writing, I came across this book from Harvard Business Review: HBR Guide to Making Better Decisions. From the cover it seems similar to my structure: analysis, risk, uncertainty. I will read this book and maybe come back and make some updates.
• Several folks suggested that I could have used a more detailed example to illustrate this framework. I felt the same way; the case studies fell a little bit short. However since aspects are quite personal, it may not resonate as well, and there are also privacy and confidentiality concerns. May be I will revisit this in the future, and update accordingly.
• From Mr Jeff Bezos congress hearing: “He convinced me to think about it for two days before making a final decision. It was a decision I made with my heart and not my head. When I’m 80 and reflecting back, I want to have minimized the number of regrets that I have in my life.”
• The opinion presented in this article is my personal perspective and not associated with my employers.