I'm working on putting together a piece on something like this. I've got a running theory that the response to one of your stories can be broken down into pretty simple categories based on time.

The first few days are the response from your immediate audience (usually). So, the publication you submit to, your subscribers, and so on.

This drops off quickly once that pool of people are exhausted. Also, there's usually a very big drop in response rate after your piece is no longer visible on the "front" of a publication's page.

After that, most of what you see is from Medium's distribution algorithm. And it's a product of two things:

1) How many people Medium shows it to

2) Your click-through rate.

There will be big changes at this point if Medium decides to show it to a new group of people (i.e. they decide your piece will go out via their email digests, etc..), or stop showing it.

And they'll show it more if you have a higher click-through rate.

Seems like a no-brainer to me.

I started exploring this idea a while back because of the people who whine endlessly about how distribution killed their views.

Like... seriously?

How could they not be aware that the initial burst of enthusiasm had nothing to do with Medium's algorithm at all? And that it had to die out, and quickly, because your initial pool of followers is finite, and can only read your article once?

Anyhow, good article, sir.