Open letter to founders: entrepreneurship is not physics
younglingfeynman.com
younglingfeynman.com
> Because the world is so complex, you’ll often have an answer faster by just building some prototype of the hypothesis you want to test.
> If it catches on, you can always scale later.
I think one of the coolest real-world examples of this in practice is how Facebook runs thousands of mini-experiments with different users. Facebook is known for having one of the best data science departments in the world and when they want to change or add a feature of the website they do it for a small subset of users and see what the response is like. If the users seem to like it they can repeat the experiment with a larger subset of users. If the users don't like the change there's no big risk because only a few users were exposed.
Contrast this to how Snapchat completely changed their UI for nearly every user all at once and got a ton of negative feedback for it. That was a complete UI overhaul which is different than FB and their hundreds of mini-changes but the idea is similar.
The problem with facebook's approach is that you can get stuck in local optima and it might be hard to break out. That is why Facebook has done some big changes in the past as well, not just incremental improvements
Difficult to do though in huge companies. Which is why most disruptive innovation ends up being acquired by those companies rather than created by them.
On way to address that issue seems to be to allow a separate arm to essentially act as a startup, preferably far away from the main company.
You may think it should better the world by challenging obtuse viewpoints and spreading truth but that s asking a lot for a marketing platform. You wouldnt ask that of JCDecaux panels on your bus stop, right?
Paragraph 1: Topic paragraph, introduce topic and question.
Paragraph 2: Summarize argument.
Paragraphs 3-n: Topic/assertion sentence. Support assertion/argument. Last sentence may summarize support. These paragraphs should follow the overall structure but sometimes you will need more than one paragraph to support something you suggested in the second paragraph.
Paragraph n+1: Review conclusion and some salient corollaries.
Paragraphs n+m: Discussion of implications of conclusion.
I read through paragraph after paragraph and I have no idea what your argument was, where it was going. If you have lots of single-sentence paragraphs, you're probably frustrating and confusing your reader.
Check out this exchange: https://www.reddit.com/r/juststart/comments/jl02ba/useless_i...
They force the entrepreneur to put some level of deep thought into the entire endeavor and create some sort of initial hypothesis.
Even if it proves to be entirely incorrect in future, it provides a very good starting point, that you can keep iterating upon as you learn more.
Most of the times you don't know what you don't know and the business plans and financial models force you to think the entire thing through and start getting a feel for what you know and what all you perhaps don't know.
Edit: to be very clear, I’m not suggesting that one develops an old style 50 page business plan. I’m simply suggesting that one must have at least a basic understating of the business, the need, the opportunity, the revenue model and so on. It’s a business after all.
It is context dependent though. Very different if you're building Snapchat vs. SpaceX
Edit: before you spend time trying this: it does not have a proper printing stylesheet either.
https://i.imgur.com/mAP7Uqe.png
Edit: judging by the CSS, it seems actually intended. It just does not work well with my screen or reading habits then.
(For what it's worth, here is what would work better for me. Headlines reduced to 3vm fontsize, paragraph margin-bottom to 0.4em to see more content at once. Continuously having to scroll interrupts reading quite heavily for me. The changes look like this:
https://i.imgur.com/8JBo91t.png
Our mileages may vary ;)
If the plan is wrong, an underlying assumption is incorrect.
They all show hockey-stick growth up and to the right. Not many businesses succeed by that measure. And the goal of a lot of business plans is not to plan a business, but to plan to get funding. Big difference.
And a lot of businesses pivot after realizing they were totally wrong.
> If the plan is wrong, an underlying assumption is incorrect.
Incorrect, or made up. Nobody has perfect information beforehand.
Both Juicero and WeWork were known to be business failures from the earliest days, by people who didn't even have to read any financial reports.
Red Bull is kind of a sleep replacement/stimulant, which a lot of people needed pre-corona. Americans somehow were willing to sacrifice their sleep for their job/commute.
Dyson vacuums - that I can't explain. Maybe their high prices signal them as better quality. Maybe somebody can explain that?
Red Bull offered an alternative to coffee and the market was lightly validated in rural Thailand. You had Krating Daeng in 1975 but it didn't really take off until they started sponsoring fighting. (It had a bit of a status problem.)
Nespresso, in addition to what you said about essentially changing the price architecture, also enhanced simplicity. A simple pod allows you to have a good cup of coffee quickly.
I think that Dyson was appealing because of the same enhanced simplicity as Nespresso. You didn't need bags anymore. I also think there's some cognitive thing happening where people derived pleasure from seeing their mess by hovered up inside the machine. And like you said, status could also play a role here. They've created the perception of being an Apple of boring household appliances.
> I need merely two variables [...] resulting in a simple yet accurate model.
It's accurate-enough, but not accurate accurate. This is why things like quantum gravity are being researched. Maybe I'm looking too deeply into this analogy, but I think it's a really bad one.
> In highly complex systems such as business, things like butterfly effects can cause massive distortions rendering our models flawed.
Most businesses fail due to a lack of product-market fit. This has nothing to do with chaos theory or initial conditions. In fact, I'd say if you have product-market fit, the market tends to be very lenient of initial conditions.
> In order to perfectly predict the weather, you’d need incredible amounts of data and an equally overwhelming number of data points and then somehow synthesize that into an accurate prediction.
I'm pretty sure that we're still not sure if you can even actually predict the weather -- like, ever (we aren't sure if Navier–Stokes is smoothly solvable in 3D). All these physics analogies are just shallow and inaccurate.
> It’s easy for the takeaway to be human behavior is difficult, business is complex, all data is meaningless, we can’t apply a scientific process at all, let’s wing it. And while there are people who lean heavily on intuition (Gary Vaynerchuk comes to mind) I believe if that’s the takeaway, the pendulum has swung too far out of whack.
I'm not sure why this is "out of whack" -- to me, this seems like a perfectly valid strategy. As I get older, I value intuition more and more. There's a reason Warren Buffet's moniker is the "Oracle of Omaha."
> lay out.
I've been told that in this thread. I'll try to improve. Thank you for letting me know.
> Not accurate accurate
True. But it would be overkill to use complex equations for 1 dimensional motion. The degree of error is negligible in that context.
> markets tend to be lenient of initial conditions post PMF
That's a very fascinating point, you bring up. But I'm not sure how one can separate initial conditions from PMF. PMF takes initial market conditions into account by definition. At least, if we're going by Don Valentine or Andy Rachleff's definition.
> shallow, inaccurate analogy
I'm not sure how to respond to that.
> not sure why this is "out of whack"
Perhaps I didn't express myself well enough. I was referring to fully negating any utility data might have and solely relying on intuition. My point was that there's a middle ground but it's further toward intuition than most people think in our data driven world.
I'm sorry you didn't enjoy the essay. I do appreciate you reading it as well as the time you took to give me feedback. I'll try to use it in order to improve future work.
Thank you.
I don't agree. Initial conditions are minutiae that, on the surface, don't seem to matter -- hence the canonical example of a butterfly flapping its wings cascading into a hurricane a few weeks later. Product market fit is not that. If "Facebook" would've been called "Smilebook" (which is an initial condition), do you really think it would've failed?
In fact, worrying about your domain name, and what your LLC is going to be called, and what you logo should look like are all rookie mistakes when founding a business.
It's too easy, connecting the dots looking back. I believe we have to be careful with hindsight bias.
> In fact, worrying about your domain name, and what your LLC is going to be called, and what you logo should look like are all rookie mistakes when founding a business.
I agree. But that's not my point. My point was, avoid trying to think your way to PMF, it's too complex. Just try a bunch of stuff and iterate based on real world feedback.
I liked the part where they also talk about "asymmetrical risk" and creating experiments where you may risk 10% of your capital, but if that thing works, it will double your business. I think this applies to all aspects of entrepreneurship, not just building the product. For example, I write about acquisition channels that consistently work for founders [1] and found the most successful ones place "small bets" that can potentially have "big results". Like mentioning their idea in a FB community where they could get 1000s of potentially interested people. Time invested: 10 minutes. Results: 1000+ potential users. Not bad.
Nassim Taleb also talks about this "asymmetrical risk" concept [2] and a lot of his "Antifragile" book was about it. I'm really curious who originally came with this concept (it seems to go by different names, the OP calls it "asymmetrical risk", Taleb calls it "convexity), and it seems the first person who came up with it was someone from the finance industry decades ago.
[1] https://firstpayingusers.com
[2] https://www.advisorperspectives.com/articles/2013/07/16/nass...
As for who came up with asymmetrical risk, great question. I'm not sure where I heard it first? I think it might have been Vinod Khosla of Khosla Ventures.
Now that said some thing can work to insure success or make it more likely.
Take the author here. His firm has 'feynman' in their name. On HN who from what I can tell worships Feynman (I didn't even know who he was prior to HN) that will work. In another place it would have the same impact of using 'warby' in 'warby' parker. Meaning good but not stop the presses. (It's clever but some might also say it's annoying to use a cheap trick like that not the same either (it's all nuance) as calling your bagel place 'einstein bagels' (there is one or was one).
I have been 'studying' (you could say) both successful and failed companies for probably longer than most HN readers have been around. Best I will say it's hard work, luck, and of course timing as the major factors. Separately they don't teach any of that when I was in of those really good business schools that people would die to go to.
It's interesting that you accuse my company of not taking survivorship bias into account yet here you are falling victim to retrofitting.
Ok so why do you have 'Feynman' in your company name then? And I didn't say it was for getting upvotes either. I implied (or meant to imply) that it was as a way to get attention by inventing some kind of fake association. The fact that it didn't do anything before does not mean it has no value (as a test).
> yet here you are falling victim to retrofitting
Explain what you mean by 'retrofitting' in this context I don't understand that meaning here.
You are familiar with the use of celebrities who associate with products. The idea is that someone is more likely to pay attention and consider if they think a celebrity is involved or associated than a random person. Does not mean it will always work or that you can succeed without it. It means in many situations it will get you more than w/o it as always 'it all depends'.
Paul Newman's salad dressing came out in 1982. By Paul having his name on it he got instant recognition in the press and in supermarkets. Doesn't mean people would buy it if it wasn't good but more likely to take a chance as a result of that name a shopper would stop and say 'hmm what's this'.
Maybe this is what makes the difference between statisticians trying to predict the election and the method used by "Trafalgar Group" who try to understand each states regulations and specifics to make rule based predictions, instead of extrapolating from a big pool of data. In two days we will know who was right.
[1] Pollster who predicted Trump's 2016 win makes 2020 prediction: https://www.youtube.com/watch?v=0BOhRCeoMzM