Poker Strategies to Use When Sacrificing Salary for Stocks in a Startup
I hope this post helps talented professionals in their evaluation when joining startups
A friend recently got a proposal from a young startup founder. The offer was 1.5% over 3 years of vesting period for part-time unpaid work. The startup has an early-stage prototype tested with a handful of users, very high risk.
This was just one of many examples I have come across the past five years were offers are made that can never become profitable for the future employee. Being an avid mathematician and hobbyist poker player, I thought I share how I evaluate offers from startups.
Prerequisites — what is my sacrifice?
Let’s take me as an example, as a contract cloud-native engineer/architect in London, I make 800 GBP per day (roughly 1000 USD). During a year, there are roughly 260 working days. This means that my annual salary, all benefits included, can roughly be calculated to 208 000 GBP or 260 000 USD. Few startups can pay above the £100k / $120k mark.
Hence, regardless of whichever startup I join I will need to make a substantial salary sacrifice, roughly £108k / $140k per year. Since working in a startup is more fun and rewarding, let’s write down the sacrifice to only £54k / $70k per year. That means, every year I am working for the startup, I am losing £54k / $70k compared to contracting.
Poker Theory — how to value a bet
Imagine you are playing poker, there is £11 in the pot and you got a 50% chance to win if you call your opponent. The price to call their raise is £1.
In the long run, every two times you call the bet, you will win £12 one time. Or simply, you stand to win £6 for every time you call the bet.
In poker, this is called “pot odds and expected value”. The pot odds in this example are 6:1 and the expected value is 6£. Any expected value above your bet means that in the long run, the odds are in your favor to win more than you lose.
Imagine a different example where there is only £6 in the pot and you got a 20% chance to win if you call and the price to call is £4. In this case, on average you need to call five times with £4 to win £10 one time. Your total spent for one winning hand is £20 (5 times 4£), your pot odds are 10:20 and the expected value is £2, significantly lower than the £4 investment to call. In this example, you are losing £2 every time you call your opponent.
Poker Theory applied to startup stocks
My sacrifice when joining a startup is £54k per year. That is the price of every year I spend in a startup instead of contracting. The average time for a startup to exit is 8–10 years (see article 1 at the bottom), let us say 10 for ease of calculation. This means that the price of my bet will be £540k.
The chance of startups exiting is constantly very low. The highest chance of exiting is 25%, but in seed and series A stages, it is only 3% (see article 2 at the bottom). We can assume that on average, there is a 14% chance for a startup to exit.
This is not mathematically correct, as there are many more early stage companies than late stage companies, the real chances of success are therefore below 10%.
Regardless, we will stick to 14%, or roughly 1/7 startups achieve an exit.
This means that my shares need to be valued at least £540k * 7, or £3780k (almost 4 million GDP), for this bet to have an expected value above my bet.
That is after writing down the value of my bet significantly due to the rewarding environment of startups.
Also, remember that chances are high that my shares are diluted before the exit, making it even harder for them to reach those values if the initial offering is low.
Additionally, most VCs demand clauses that guarantee them their money back first if an exit happens, meaning that if VCs invested 100 million and the exit is on 100 million, no shareholder beside the VCs will receive any money at all.
Earlier, we assumed 14% of all startups make an exit, however the share of startups who exit with a profit for the 25 first employees is even lower.
More gambling theory — bankroll
Another important number is the number of bets that you can afford. For example, consider a lottery that has 1000 000 000 000 (that is a trillion) in prize pool where the winner takes it all. Assume that there are 1000 000 000 (that is one billion) lottery tickets and that they each cost £1 to buy. This lottery has a very high expected value. For every £1 spent you stand to win £1000.
The problem is that the gambler on average needs to have £1 000 000 000 (a billion) in the bank to be able to participate enough times to hit the winning ticket. Can you afford participating enough times to give yourself a fair statistical chance of winning?
Also, if the above lottery occurs once a day and you can only buy one ticket every time you would need one billion days to participate enough times for a fair chance of winning. A lifespan of 100 years has roughly 36500 days in it, in other words, you die long before you receive a fair chance.
The bankroll theory shows that the expected value is not the only consideration to be made when taking a bet. The gambler needs to be able to bankroll enough bets to give themselves a statistical chance of winning or be prepared to loose on those terms.
Bankroll applied to startups
Most senior people in tech are 30 years or older. This means that they only got a few “startup-bets” to play. If they work till their 60s, that leaves them with three 10-year bets. That is not too many bets, meaning that if a senior technician does not receive their requested remuneration, they do well for themselves to leave the company as soon as it loses steam to maximize the number of bets they can make.
I find it off-putting when a young, promising, but fresh and inexperienced founder tries to downplay the value of my contribution as part of the negotiations. My friend who received the offer was told that the deliveries “would not require full-time participation” — by a university freshman who has never done that job. In the past, I have had a Machine Learning graduate without any work experience telling me that developing an application with a billion monthly visitors, a high load of data writes and strong consistency is “easy” and not worth much compared to ML-algorithms.
To me, it is critical that founders know that they know, and what they don’t know. Else I find it a poor investment to co-own a startup with them…