Loading post
A while back when VC jobs are still in vogue, I got into an interview and the first question is "How do you distinguish signal from noise and increase signal-to-noise ratio?" By the way, the more I look into recent VC fundings, the more I don't consider today's VC investing market as particularly strong in signal. Their investment theses are no different than those of growth investors, following crowd wisdom without judgement.
You can't trade stocks well in New York, or build a good software in San Francisco. This is largely a hypothesis but I do feel that way from personal observations, and I try to infer some logics behind this thought.
I get better performance on stocks after moving away from New York. At the same time, I do hear fewer and fewer people talking about the shiny companies. As the financial capital, New York is like a bubble where even your carpenter hear about people pitching stocks. There is nothing wrong with the enthusiasm surrounding investment. However, the more information you receive, the harder it becomes for you to determine which information is signal, and which is noise.
When you get away from the center of action, two things begin to happen. You realize that the winning investment idea you have is universal to other parts of the country/world. You also receive less information, but if you actively seek out, the information you receive is of higher magnitude and lower frequency. The lack of information diffusion whittles out the weaker noises, making it easy to pick out the signal.
The same idea can apply to software as well. When everyone is building a software, it is hard to know whether the direction you are going is correct. When the local-specificity is reduced, you begin to realize whether, the software idea is beneficial to people half-way across the globe, or not.
On several occasions, I have debated investment approaches with others. The philosophy behind a good company is ubiquitously similar, a good leadership team, a strong care of customers, and a simple mechanism of deliverying value to the society through its product. How to evaluate a good company is irrelevant to its industry, or the price of its stock.
The real business opportunities follow a hidden markov model, where the participants in the market only observes the observable output sequence. The above values are the latent process, the driving force behind a good product, or a good stock price, the TRUTH. Unforutnately the observed sequence of real-life events is likely polluted with randomness, creating noise, and the business opportunities in the real world is non-repeatable, so I cannot demonstrate causal relationship to you in a controlled experiment, or to calculate an exact Maximum Likelihood Estimation (MLE).
Tying to our earlier discussion on noise and applying scientific notion in a less scientific setting, the less noise you observe, the more likely your Maximum Likelihood Estimation (MLE) of truth in your mind will be a closer fit to the real truth. The closer your MLE fit, the better your prediction on future observable events.
The result should be close to inductive reasoning with some common-sense (which is rare in this world). Just by looking at the executive team and what they say about the business, you could have easily guessed how well Costco performs, and how terrible Adobe is and will continue to be. Similarly, when your friend keeps telling you about how great his/her perks from Amex is, and tries to get you on his/her referral, you could have invested in Amex's stock.
It is obvious how easily the MLE will overfit for noise, if we chase the crowd at the immediate moment. Costco spent decades to build a reputation, Adobe's product line decline has been at least a decade in the making, etc. A philosophy cultivates a world view -> World view leads to a certain belief -> The belief in certain values leads to business decisions -> The business decisions shows up on the quarterly reports. This process takes months, if not years or decades to complete.
As a start point, form a hypothesis of your value and philosophy, and observe how well your decisions fare. If you are right, the same hypothesis will be proven in multiple subjects/industries, across any time horizon. If you are wrong, at least you know where to adjust your value, and try again.