People subscribe to your blog based on expected future value, not based on perceived past value. Unfortunately there are many scenarios where every single one of your past posts could have high perceived value, but the expected future value is relatively low. For example, if each blog post is about a completely different topic. There are very very few writers who can get away with this, and most of the ones who do are only able to do so because the writing follows a predictable pattern in some other way, e.g. Malcolm Gladwell. The sad truth is that how often you update your blog is usually a better predictor of the number of readers you'll get than the value of the actual writing.
In the end it's maybe 25% about having good ideas and helping people and 75% about putting on a show.
This is very unfortunate, but also true when it comes to pure popularity on the web. However, one would hope that people starting blogs are not doing it just for the sake of popularity. The grandparent's case is a great example of how writing with the intention of creating value with original content can be very rewarding in a way that would be impossible by just "putting on a show".
Since then it's been a really wild and rewarding ride. I've had a chance to consult on stories for reporters at the Wall Street Journal and posts featured in the Financial Times, NYTimes, WSJ. I've even been contacted by some of my heroes (really great investors who've been written about in books like The Big Short).
Now, in a couple of weeks I'm being flown out to interview with a firm that I never dreamed of applying to. All because they read some of my posts and enjoyed them.
The key to me is just to start writing. Write all the time and strive to continuously improve on it. I think that if you produce good content, people will come. Most of my most popular posts are things I've worked really hard on. Some examples include me reading 5 years worth of message board posts to track down everything that person X said or visiting the library to dig up a physical copy of a report from the 1980s to scan and then post. If you create value, readers are going to keep coming.