THE SECRETS OF WRITING ON MEDIUM, PART 11
Did Medium’s Programmers Add a Gaslighting Function or Does the Algorithm Have Feelings?
A lighthearted look at how Medium can drive you crazy.
Wednesdays on Medium make me feel as if Christmas has become a weekly holiday.
And it’s not because of the joy Silicon Valley Santa brings to us.
It’s because he keeps sending me lumps of coal when I’m nice and presents when I’m naughty.
In today’s Medium Partner Program earnings, I saw that the highest paid article had far fewer views, readers and fans than many other articles.
Here are the weekly and lifetime earnings of my four Medium-related articles since late September:
Something feels wrong about the picture above.
First of all, the overall stats of each article didn’t correspond to the rate of compensation.
As you can see below, “Meet the Medium 11” earned over twice as much money as “You Earned THAT Much,” but had fewer reads and fans, and only four more views.
Also, it earned almost 75% the amount of “The Ultimate Guide to Clapping” even though it had only one-third of the reads and claps and half the views.
That’s nothing new, as I pointed out in the last part of this series.
But there was something still bothering me.
I remembered that “Meet the Medium 11” hardly had any views or read over the last ten days, so I looked up that article’s stats.
How could this article make so much money in the week of October 6–13 when the graph is almost flat?
It’s hard to read the exact numbers in the graphic above, so I will tell you that the article above got exactly ONE read (and a clap) from one internal view on Medium for the second week of this MPP pay cycle.
If we assume each pay cycle has one part corresponding to each week and the single fan clapped only for my article that week, a full 1/4 share of the monthly subscription fee would still be $1.25.
How does a single clap from one lonely fan translate into earning $3.48?
Should I hope for ZERO views on this story and expect to pay next month’s electricity bill?
Out of curiosity, I reviewed the individual stats for “the Ultimate Guide to Clapping.”
It was the most popular of the four articles, and I made it my profile feature article around the beginning of October.
I tried this technique to see if people looking at my profile would read the featured article.
As you can see, after the initial surge of interest, it was ignored during the last week of September.
In the second week of the pay cycle, starting October 6, there were 61 internal views and seven new fans.
Finally, I looked at the individual stats for “You Earned THAT Much.”
The article generated all of its stats between October 11th and 12th, the last two days of the second week.
Are you telling me one fan for one article is worth more than 26 fans for another article?
In case you were wondering, 25 out of the 26 fans of this story have paid subscriptions.
That’s when I had my epiphany — the Algorithm has feelings!
None of my stories were curated, in spite of their detailed citations and outside scientific references.
The only possible explanation for their differing rates of success had to be the curation bot that screen articles before they are sent to the Corgis (a.k.a. Medium’s curation staff).
Now a new correlation emerges —headline tone toward Medium and compensation.
If we look at the relative strength of value per fan (VPF), these articles are ranked by their favorability toward Medium:
- Meet the Medium 11: 46.0¢ VPF
- The Ultimate Guide to Clapping: 19.8¢ VPF
- You Earned THAT Much: 13.9¢ VPF
- From Newbie to Addict: 6.1¢ VPF
The first headline drips with brown-nosed goodness.
It does what every other self-help and life hacking article does, oozing with positivity about writing on the platform and, by extension, Medium.
The second headline focuses on one of Medium’s features and presents a positive tone on how subscribers should use the program, so it’s more like a user guide than a cheerleader.
The third headline does not supply a dollar amount (no-no #1) and ends with a question mark (no-no #2).
Either it is a positive clickbait article or one that questions the fairness of the paywall.
I added the fourth headline as a control group.
While the headline is negative, the article focused on scientific research of the gamification of business and social media and other effects that increase the addictive quality of Medium.
I wrote it as a public service for Medium beginners before they get sucked into the addictive effects of playing the Medium game.
Isn’t it strange that the most carefully researched and scientific article on a subject of cultural relevance (social media and gamification) — the very thing that Medium’s curators claim to value— had less than 1/7 the VPF of the article that acted like a Medium cheerleader?
I’m sorry I hurt your feelings, Ms. Algorithm.
Because I don’t want to consider the alternative explanation for these bizarre results.
