Reverse Engineering the Medium Earnings Model for Authors

Part 3 — What is the optimal read time for your stories?

In Part 1 of this series we learned about the impact and importance of followers on our earnings and we evaluated various techniques to gain them. In Part 2 of this series, we investigated the earnings impact of publishing our stories on Medium publications versus self publishing. To do so, we created a proxy for earnings, based on claps and followers, that we used to measure the "goodness" of each of the stories in our data set.

Now that we have a measure of story "goodness", our job has gotten a lot easier, and there's a lot more to learn.

In Part 3 of this series, we will use the data, and especially our new measure of goodness, to discover the impact story read time has on our earnings.

Our data set consists of roughly 500 Medium author profiles and 50,000 stories.

This story will be, not coincidentally, much shorter than the others.

"Goodness" vs Read Time

Does story length impact your earnings? It does, and the impact is profound.

First a quick review. Because earnings per story is private to the authors that write them, we need to create a proxy for earnings based on public information for each story. In Part 2, we developed a metric called "Normalized Claps" to serve as that proxy for earnings. This allows us to estimate the relative value of stories created by all authors across Medium, since claps per story is public information and presents a simple representation of "goodness".

Figure 1 below shows a plot of nClaps, our normalized claps goodness measure, versus Read Time. While most stories are less than 30 minutes in length, there are a few that are considerably higher. In one case, Read Time is well over two hours!

Figure 1: Normalized Claps versus Read Time

We can see from the data that stories over about 25 minutes in length just don't do well. Readers clearly don't stick with them.

Let's expand the scale to focus on shorter stories.

Figure 2: Normalized Claps versus Read Time for Shorter Stories

We see, in no uncertain terms, that stories 5 minutes and under do best, with stories over 7 minutes generally fairing badly.

Stories around 4 to 5 minutes in length do slightly better than shorter stories, but not by much. Stories 1 minute in length do nearly as well.

What does this tell us about Medium readers? Clearly, they like their stories short and sweet. You're clearly better off breaking up your stories into smaller chunks and publishing more frequently.

I suppose this should have been a 10 part series as opposed to a 5 part series. Next time.

Wrapping Up

The message here is clear. Keep your stories short.

And perhaps you noticed — Part 3 of this series is a lot shorter than Part 1 and Part 2. Lesson learned.