On Modeling the World

2025.08.24

CXIV

[From Finance Bro to Tech Bro, Day Trading & Day Drinking, Dorm Room Hedge Fund, ARPU & Oil & Pills, Sellers are Just Analysts, No Earthly Way of Knowing]

Thesis: From finance to B2B SaaS for sales people, modeling the world has been a surprising through line for me.

[From Finance Bro to Tech Bro]

When I started operating in the sales space in late 2023/early 2024, both my cofounder Jack and I didn't really know anything about business to business (B2B) sales at all. Rather, we were both what you might call finance bros.

Jack was an analyst for a real estate company, I had just got done managing some people's money, and together we had just failed at commercializing a software for retail investors.

So, how did we get from there to running a B2B SaaS business for sales people, one that they actually seem to like?

Really, on my side, what we're doing is just another iteration of what I've been trying to do for as long as I can remember--modeling the world.

To me, modeling the world is how do you take something as complex & nuanced as the physical world where things happen and turn it into some simplified version that still contains truth & value, often in the form of predictive power.

A really good entry point to understanding the journey is my degenerate gambling in high school.

[Day Trading & Day Drinking]

During my junior & senior years in high school, I started trading stocks. I did some very naive swing trading of Tesla in between classes at school.

I also worked the front desk of a landscaping place, and whenever there was down time me and my boss would spend too much time on the WallStreetBets subreddit & ride pump and dumps that we didn't know were pump & dumps.*

As I was entering college, I started studying options more and liked the math involved with them. I was playing with some strategies meant to long volatility. I made some money on those, until I lost almost all of the profits with one mistake and had a wake up call to the amount of risk I was actually taking.

This was where I was really starting to understand & learn how you can model reality within finance. I studied options so much because I was obsessed with the fact that you could express any belief with them, not just whether the stock was going to go up or down.

I also had some fun wins--I remember sitting on my laptop in my first semester english class & 2x'ing a volatility hedged bet against BeyondMeat after it's lockup was over.

And I had some fun losses--another day, two friends & I were sitting in a private study room around lunch, drinking something that wasn't water, and streaming an FTC meeting that was going to decide the fate of a satellite company we all had positions in. Needless to say, the FTC did not decide in favor of our trade!

The point is, at the very start, in a lot of regards, I really was a degenerate gambler.

At some point in there, through some mix of persistence and curiosity and repetitions, I started to learn how these instruments worked and how you could use them to model the world.

But, I suppose all that separates anyone from not knowing anything to being good at something is persistence, curiosity, and intentional repetitions. (The latter may very well just be the product of the first two qualities)

*I actually think I still own shares of Movie Pass & this Chinese company called Yangtze River Co.

[Dorm Room Hedge Fund]

The second half of my freshman year in college, I read both the 4 Hour Work Week by Tim Ferriss & The Blackswan by Taleb, two books that have a lot more in common that is obvious at face value.*

Both talk about the power of going into a line of business with an asymmetry that can positively impact you.

If you are selling books, for instance, you have a baked in leverage--after you do the labor to write the book, you did the labor. Now, if you sell 10 books or 100 books or 10,000 or 100,000 or however many Hormozi sold last week, you don't have to do 10 or 100 or 10,000 times the labor... the book is already written, and you own the license to it! You can continue to sell it forever.

As I now understand, good software is like this, too. We don't have to rewrite BirdDog every time we sell a license to use it, even if there is marginal cost related to training & delivery.

At the time I read those books, given that a friend and I had started to figure out some trading strategies that worked, the natural place for us to find this sort of asymmetry was within finance.

The thought was the marginal cost of managing more money was low, and that we would be able to gain asymmetry quickly--we would not need to do more research or more trading for every dollar of AUM (assets under management) we added, as our strategy would scale with little extra effort up to a certain point.

While this decision was made with and an under appreciation for the challenges that would present themselves to two 19 year olds trying to get a hedge fund to a sustainable scale in a reasonable amount of time, the die was cast--I was to become a finance bro.

*Funnily enough, Taleb & Tim Ferriss are apparently good friends.

[ARPU & Oil & Pills]

While it took a couple of years to go from dorm room resolution to having a legal entity that was managing money, we eventually got there. And, as we started fleshing out our strategy more and more, my interest in modeling returns with options expanded to include modeling companies.

Countless books have been written on the subject of analyzing & modeling companies; this little section is nothing more than attempt at touching on the difficulty in picking which level of granularity you want to model the company at as well as the variance in the sorts of data you would want to track from company to company.

When modeling a company, you could in theory get infinitely granular--as an example, for each facility a firm owns, you could figure out how many people work there and what kind of paint is on the walls.*

Typically, though, if you get that granular in your research, you'll expend too much time and energy looking at things that don't matter. You might spend a lifetime trying to understand a single large corporation & get nowhere!

On the other hand, you can't zoom out too far. It's very easy to find a big data set on any company's financial from which you can build a standard issue model. The trouble with this is that these numbers usually aren't enough to really have a good idea of what's going on.

For starters, the same number can have a different meaning in different contexts--the level of debt you'd expect to see with a utilities company is very different than what you'd expect to see on a software company that just ipo'd.

Further, for any given industry, there's usually some collection of data & variables that are particularly relevant for firms in that industry. Those variables often come out before the company reports earnings and can be used to help predicate the earnings!

For an oil company, as an example, you'd be remiss to not look at the number of oil barrels coming out of the Permian Basin. Or, in the context of pharma companies, you'd want to know how many prescriptions are being written per month for the drugs that company produces (yes, this is often an available data set). In the context of software companies, you might want to know a lot about average revenue per user (ARPU) and where that number is trending.

Since these factors are so varied yet so valuable, the really interesting thing to me became how to efficiently find & interpret this an arbitrary useful data set for any arbitrary company or collection of companies. Ie, given what data points go into properly pricing an oil company, could we possibly filter a giant stream of information, like the internet, and extract the right data points wherever we could find them, whether it be in a financial report or on a government website?

Well, we started building out some tech to see if we could.

*The grand catch is that both of those things become exceedingly important if the firm gets a class action lawsuit for having lead paint on some of it's walls; if you had that above information, you would likely be able to make a much better estimate of how much money the firm will lose from that law suit & decide if the market was over or under reacting.

[Sellers are Just Analysts]

Technically, we never made much headwind on the problem in the context of finance. We did, however, realize that sellers have a lot of problems that are solved by a very similar solution. And, that solution doesn't need to be quite as complex or granular as one would need to be for a finance person to use it.

When we caught wind of this, I left the fund & Jack & I pivoted to working with sellers.

We spent December 2023 and the first two months of 2024 talking to (literally) hundreds of sellers, learning as much about the space as rapidly as we could, and sending them data until we found what they actually cared about. March, April, and May involved each of us spending over 40 hours a week manually delivering research & data to some sales teams we contracted with.

After getting over some misalignment with some other parties in June & July, Jack & I started BirdDog at the end of July 2024.

Bringing it full circle and explaining BirdDog in pure technical terms, we let sales people define their dream dataset on a company and our system goes and fills out as much of that dataset as possible.

In other words, we're still solving the same problem that I thought was interesting in finance.

And, so far, a lot of sales teams are really loving it.

If you enjoyed this post, click subscribe below—I’ve been here every week for the last 114 weeks.

[No Earthly Way of Knowing]

Looking back, the most interesting thing about this story is the way that it flows.

I've been interested in modeling things for a very, very long time. A journal entry from when I was 13 has notes about me trying to understand the crowd dynamics & flow of students at the end of the lunch hour when we would congregate in the hallway. A fascinating part about options for me was modeling returns under 'all' conditions and building a 'foolproof' trade. Then, modeling companies was interesting to me in a way that has been relevant in both finance and sales.

In every case, I’ve just been trying to build out something that simplifies the world by capturing only the relevant & useful facts.

Even if from an outside perspective this journey doesn't obviously move with any rhyme or reason, it certainly makes sense to me in retrospect.

I’m not sure if I broadly believe that ‘everything happens for a reason.’ I do think, however, if you follow your curiosity, you will be surprised by where it takes you.

Live Deeply,