Data-Driven Tips on Writing Medium Essays

Five-Minute Reads, Four-Finger Salutes

X-axis: minutes it takes to read my essays; Y-axis: Read ratio, in percent.

I often say that when you can measure what you are speaking about, and express it in numbers, you know something about it; but when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meagre and unsatisfactory kind…

— Lord Kelvin, Popular Lectures and Addresses, vol. 1, “Electrical Units of Measurement”, 1883–05–03

I’ve been writing essays on Medium intermittently for the past two years, ever since my graduate-school friend and prolific writer suggested that one of my long Facebook analyses deserved a wider audience on Medium. “What’s that?” I recall asking.

I’ve since published 13 essays on a wide range of topics (from higher education to football to weather to racism) in those two years. Thirteen essays is a whole lot less than the output of many people on Medium — but more than quite a few, too. Three of my essays have been picked up for broader audiences and have been read by more than 2,000 people — which is puny compared to the big guys, but more readership than some. I’m in that big in-between.

As a scientist, I’m a data nerd (although because I’m a born journalist from a journalism family, not as nerdy as most of them — again, somewhere in-between). I started noticing some trends in the “stats” data about my essays. I thought I’d share them here for those who are just starting out on Medium, or are just curious. Note that these insights are different from the usual writing advice you’ll see, because this is evidence-based rather than some sage opinion on how to write that killer essay.

Keep It To a 5-Minute Read

I took the stats from Medium and put them into an Excel spreadsheet, and then did some correlations. What jumped out is the statistically significant (p < 0.005) anticorrelation between the length of a read and the read ratio (reads/views). Longer read = fewer actual reads vs. views. Which is no more brilliant than saying, “tl;dr.” But I’ve got actual data here, to put some specificity on just how long the “tl” in “tl;dr” actually is.

See the graphic at the top. The diamond symbols show the results for each of my 13 essays. Your eyes will tell you that the shortest reads (at left) are also the essays with the highest read ratios. I did a regression analysis to come up with a quantitative formula for it. The formula predicts that a really, really short read will approach a 70% read ratio. A 5-minute read? Between 50–60%; for my most widely read essays, which have been 3–6 minute reads, the ratio is usually very close to 50%.

You lose readership with each minute of a read, and it’s noticeable even for a 6-minute vs. 5-minute read in some cases. My longest-form essays, two 16-minute investigations, hover around a 40% read ratio. So, the moral of this story: if you want half the people who view your essay to actually read it, I strongly suggest aiming for a 5-minute read.

How do you do this, if your normal attempts at writing essays lead to 10- or 20-minute first drafts? The best advice I’ve ever read or heard comes from the famed writer Norman Maclean’s Presbyterian pastor-father-teacher, immortalized in the film A River Runs Through It. That advice is: “half as long.” And, “again, half as long.” , recorded in something less than high definition.

If you get really good at writing and editing, you will find that virtually every first draft can be shortened by 50% without much loss of vividness, and with much increase in readability. That’s the ratio the Rev. Maclean indicated, and it’s also the ratio I have found in my own writing as a columnist for newspapers and magazines outside of my Medium experiences. (If you can shorten by a total of 75% as the good Reverend required in a third draft, then you can become the next Norman Maclean!)

The Rhythm of the Views: The Four-Finger Salute

One other pattern I have recognized is what I’ll call the “four-finger salute of the reader.” Most of the essays I’ve written, regardless of whether they are viewed by thousands or by 100 people or fewer, have about a 4-day cycle. It looks a little like your right hand held with the palm facing the screen, except that the ring finger and pinky have been in an industrial accident.

Day 1, like the index finger, gets off to a good start. Day 2 is the middle finger, sometimes the peak, especially if the essay is “curated” by Medium or otherwise gets wider attention. Day 3 views are roughly “half as much” as Day 2. And Day 4 views are “again, half as much” as Day 3… maybe a little more than that. From there, it’s a fairly small tail, unless you post another story that gets new eyes on the next-most-recent story.

For example, here are the views of my most recent essay, about racism at the University of Florida commencement. The May 12 views aren’t in yet, obviously, since I’m writing this at 1:20 am on 5/12, but I expect the number of views to be about “half as much” as the 536 views on 5/11… a small fraction of the overall views.

Views for from May 8–11, 2018.

For comparison and only a little contrast, here are the views for one of my least-read essays, a 16-minute piece (now you know why it’s least-read!) on the railroad giant CSX:

Views for from April 22–25, 2018.

Mathematical modelers can geek out on curve fits or logistical models that explain the tapering viewership; fine, but I don’t think most writers will care about those finer points. The takeaway point is that the reader cycle on Medium that I am experiencing in 2018, the analog of the news cycle on cable and network news, is about four days. That’s much longer than the 24-hour-or-less news cycle, but it also means that you don’t have a week to cogitate if you want your essay to be heard by a lot of people on a “breaking news” topic. Then again, maybe Medium isn’t always the right place for that “first rough draft of history.”

To extend that four-day cycle, you probably have to get on the treadmill of writing an essay every 4–7 days or so. I have a day job, so this isn’t in the cards for me. Also, one of my essential tenets of writing is that I do it only when the inspiration hits. I’d rather write less, but make my words and essays mean more, than be driven to endless verbal diarrhea. I have this luxury because it’s not my day job, I know. But I kind of shudder at the writing contests that reward large word counts, and a culture that seems to encourage unceasing productivity. Like the Rev. Maclean, I think “half as long” applies to essay lists as well as to essays themselves. Speak when you have something to say; otherwise, don’t. This is rather countercultural among writers, I know, but, hey, this is my essay.

Final Points

Five-minute reads and four-finger salutes! I have no idea if this is a general experience for all Medium writers. If your results vary, please comment! I’m a relative newbie here. There is much better qualitative advice on writing to be found elsewhere. The data here give you, circa 2018, quantitative insights that may help guide you as you write, and as you look at the viewership results, for Medium.

P.S. This is a 6-minute essay, so I should have edited it down by 1 minute.

A geography professor and meteorologist at UGA in Athens, GA. I write about news, sports, weather, climate, education, journalism, religion, poetry, the South.