On Passion

2025.10.05

CXX

[B2B SaaS; Voltaire in Middle School; Applied Epistemology; Empirical Wins; The Coolest Toys; Simplicity; Competition is For Losers; So, what do I Do?]

Thesis: It's easier than you think to be the best in the world at something you're passionate about.

[B2B SaaS]

When people ask me what I do, I often say 'B2B SaaS.'

I actually think this is a bad answer.

It doesn't really capture what I actually care about doing, and just how much I actually care about doing it.

I work on what to me is one of the most interesting problems--efficiently modeling the world in an actionable way.

I’ve thought about this in one way or another since I was an awkward middle schooler just trying to understand the world.

And now, by building BirdDog, I get paid to help sales teams do it at scale. And, I get to leverage really, really interesting technology and methods.

The best part is, somehow, when you descope it just enough, we're doing it better than anyone else is.

[Voltaire in Middle School]

I spent much of my childhood trying to find truth in books. I wasn't very social or normal. I really liked to read, though.

By 13 or 14, I had made it on to Voltaire and Camus and Nietzsche and Marcus Aurelius. Not exactly the authors you'd expect a 7th or 8th grader to carry around school, but they were exactly the kind I found to be supremely interesting.

What I really liked about them is that they threw out everything we were supposed to take for granted as true and then reconstructed a world view from the ground up.

This is exactly what I've been trying to do since I was that awkward 13 year old boy—whether it was trying to express truth in the poetry and fiction I wrote, or parsing through noise to find the actual meaningful information when running a hedge fund, or writing this blog.

Now, everyday, I get to send out BirdDog’s code to sail the sea of information and bring back the most bountiful data for our customers.

[Applied Epistemology]

Epistemology is the study of knowledge and knowing things. I consider BirdDog to be applied epistemology.

There's a lot of things sales teams would benefit from knowing. A lot of it is not so easy to get.

BirdDog's job is two fold:

1. Help sales people to figure out what data set is actually useful

2. Find that data at scale

Both of these jobs are very exciting.

To do the first part right, you have to understand why a company buys from your client. This is more art and less science.

You also need to build a product that protects the client from his/herself—it’s dangerous to claim you can answer very opinionated or impossible to discern questions, like “Does this company want to buy my product?”.

Avoiding this trap is harder than it sounds, because if you can convince the client you can actually pull it off, you can make a lot of money! But, assuming you can’t actually do it while saying you can, it is not sustainable and won’t work in the long run.

This also means that at the moment, what BirdDog does starts out closer to consulting. That’s okay, though, because we’re continuing to discover what parts we should actually automate.

[Empirical Wins]

When you take the responsibility of helping your client model the world, there's always a fear that you lead them astray. However, it is really satisfying to have moments that show you that the model of the world you built with them is accurate.

Last week, a customers gave us 2800 accounts. When he saw the way the signals we built with him stack ranked the accounts, he was pleasantly surprised to see that the single highest account of 2800 accounts was a strong fit he was actively trying to break into!

On top of that, one of the signals showed him that his point of contact at the company was appointed to interim CEO 2 weeks ago. This is a very important fact that he had missed on his own, but our little applied epistemology engine picked up on it no problem!

It is very satisfying to see that our model of the world is getting something right.

[The Coolest Toys]

To build out BirdDog, I really like how I get to put to work a lot of different math that I've learned about in school, on youtube, in papers, in conversations, or in textbooks. I have been able to...

• Fine tune transformers

• Make really powerful linear regression models

• Use Bayes’ Theorem

• Quantize models

It's one thing to hear about these ‘academically,’ but it's a whole other thing to actually implement them in the wild. I don't know if there is a better way to come to understand math and technology.

[Simplicity]

I said that the second thing BirdDog needs to do is to find a user defined data set at scale. The beautiful part about this is that nearly every other powerful feature we have is a direct product of doing this one thing well…

• Alerts are just updates to the data set

• Ranking accounts is simply scoring them based off of the data set

• Finding new accounts is predicting what the score of an arbitrary company will be

• Finding the best signals is just looking at which ones were more common before a deal was won

This is also super convenient because it means that as we improve the quality of the core functionality, then everything else downstream of that (which is everything that is relevant) gets a lot better, too. This plays well with my fixation on efficiency; the benefits of making the core functionality more efficient are multiplicative across the whole system!

[Competition is For Losers]

Lat week, we won a deal because our signal hit rate was 15x better than a competitor. In a lot of ways, we won by not actually competing with them.

There is a Peter Thiel quote you’ve undoubtedly heard me parrot:

I don’t take this quote to mean is not that you shouldn't try to 'win'; rather, you shouldn't try to win by playing the same game that everyone else is.

This is actually way more exciting than it sounds, because it means you can win by so much you’re not even in the same league as anyone else. It pairs nicely with a Naval quote:

BirdDog is far from the best at modeling the world, broadly. But, at this exact moment, I firmly believe that BirdDog is the best in the world at defining and tracking company level natural language signals for sales teams.

This sounds like a bold claim until you realize how narrow of a definition this is in practice. In all honesty, I cannot think of a single other company that specializes in this.

• 6Sense / DemandBase / Bombora sell 'intent', which is trying to solve a similar problem in a completely different way

• CommonRoom / Unify are primarily workflow and automation tools

• ZoomInfo / Seamless / Apollo are contact data providers with bolt on features

These companies do something else, or many something elses. This means there are a lot of things they do that BirdDog doesn’t (and won’t!) do:

• Sell third party search intent data

• Automate emails or sequences

• Research people & find contact data

• Build an auto dialer

• Automate apps like LinkedIn

We're not competing with any companies on any of these axises, nor do we want to. And, in all honesty, for that reason, they're not really competing with us!

The closest thing you can get to us is Clay, but that amounts to an expensive "build it yourself” option, and it's signal feature is nowhere near as comprehensive as ours.

Last week, I closed a firm I'm quite excited to work with. The founder put in 180 startups and 4 signals per startup into Clay, which gave Clay 720 data points to look for. Only 7 came back! Yeah, sub 1% hit rate… .97% to be precise.

We randomly selected 83 of those accounts to put into BirdDog. ONE of our signals came back with 25 hits. Together, the four that he put in came back with 51 hits, with 332 total shots on net, for a hit rate of 15.36%.

Scaled up, this is 15x difference (in the unlikely event that we didn’t find a single signal on the other 100 accounts, it would still be 7x difference!).

This is less to brag and more to inspire: BirdDog is a two man, bootstrapped startup that has been around for 14 months; Clay is an 8 year old, $3B company that has raised hundreds of millions of dollars and has teams of engineers.

We have no business out performing them in any regards. And for dozens of other use cases, Clay would beat us more than 15x over (we don’t even do 10% of what they do!).

But, since we've so narrowly scoped the definition of what we do and myopically focus on it, it's not really so hard to severely outperform many other companies on that one thing.

All we care about is our little applied epistemology corner of the world: natural language company level signals for sales teams. And, we care about it a lot.

We’re not necessarily passionate about ‘B2B SaaS,’ but we are passionate about modeling the world accurately. We’ve built up a little fenced in pen in which we can dominate by not competing.

So, until somebody decides they care about this as much as we do, we’ll win every single time. (if somebody else like this already does exist, they haven’t reached the scale where we run into them in deals regularly; even if they grow large fast, we still think the world is big enough for both of us)

If you enjoyed the read, please subscribe for more—I’m here every week!

[So, what do I Do?]

I told you that B2B SaaS was not the best answer when people ask me what I do.

So, what is the best answer? I don't know if there is one, but here are some other options -

• Applied Epistemology

• Profitable Philosophy

• Building a truth machine

• Using math to close deals

• Finding needles in haystacks

Really, the most honest answer is that I'm just doing the same thing I've been doing since I was 13…

Live Deeply,