What Percentage of Investing is Data Driven and How Much is Going with Your Gut?


1. Financial data is easy to find but hard to make sense of

If you’re interested in a company’s financial data, you can easily find them in the annual and quarterly earnings reports. But, even if you have them, how can make sense of them? If you use a short time frame (eg. past 4 quarters), earnings can seem erratic. If you use a long time frame (eg. Last 20 years), earnings can look too good to be true since inflation makes earnings look greater than what they are in reality.

Then how do you make sense of the financial data?

If you understand numbers well, you’ll find most things regress to the mean. Simply, this means if a company is having a bad season, it is likely to pick up in the next season as long there appears to be a linear growth in the last few months or years. Likewise, if you see if a company is having a good season, it is not unreasonable to expect that a few of the next seasons may dip a bit.

2. Should share price graph really be used to make data driven decisions?

Benjamin Graham calls the share market a voting machine in the short run and a weighing machine in the long run. With this reasoning that means if we limit our investing scope to the last 6 months, we are seeing how people are voting in the last 6 months for how much they think a company is valued based on the share price.

This type of short term data has more to do with psychological gameplay than anything else. But, if you’re up for a psychological challenge then looking at share price graph data is useful.

The way to approach this is rather than making predictions based on lines alone, you’re better off understanding why the stock market trends are moving up and down and use your understanding of human behaviour to make a choice.

For instance, I was once following this company for several months and noticed one day it’s price jumped by 10%. The company’s investor relations website gave an unofficial announcement of quarterly revenue growth. I believed it since I already had a strong impression of the company’s fundamentals. But, the price kept going up and down and I wanted to maximise my purchases. So what did I do? I can’t tell human behaviour of other buyers from trendlines of the stock market but I did understand my own behaviour. I felt compelled that this was a good company, so I bought it at market price.

3. How useful are alternative data sources?

Alternative data sources are sources of data that isn’t traditionally used to make financial investing decisions. For example, credit card data, geographic data or web analytics. Usually, alternative data is supposed to give the competitive edge but this hasn’t been necessarily proven.

So, should you still use alternative data?

You definitely should consider it when available.

For instance, I once had my eye on this company. It specialised in selling metal detectors and as a matter of fact 66% of its profits came from metal detectors. But, how would I know this number is believable? I simply checked Amazon on the products they sold. How much they cost and how well buyers rated the items. Each of their items sold has over 1000 ratings and usually a 4.5 out of 5 star rating. Simple statistics tells us that a sample size of 500 ratings is usually a good representation of a larger population (eg. 1 million people) if the population is completely random. These simple summary statistics may not be convincing enough by themselves but when you find enough of them and they tend have a positive outlook, then they become convincing.