An Analysis of 10k 👏Claps on Medium

This is a follow up to our last post:

Our redesign of Medium’s Claps…and why they may not have chosen to do it this way.

Over 1600 of you showed your support by applauding more than 25,000 times. Thank you.

Since clapping is a new metric, I wanted to dig into the numbers a little. Here is the clap distribution from 1 to 50, of 700 people (making up ~10,000 claps) who applauded taken from our last post. For the raw data, see here.

Key Insights

Below are some insights we derived from the data. Keep in mind the sample size is rather small (700), and the post topic could have influenced more clapping. I’ll also try to explain what these findings could mean for iterating and improving claps.

The most common number of claps given is 1 (~17% of total)

This was an interesting discovery, but made sense as we thought about it some more. Perhaps readers are still used to the former Recommend 💚 metric. To them this metric may mean a boolean way to show approval. We expect this behavior may change over time as people get more used to clapping.

Do we ever clap once in real life? 🤔 Animation by bene.

What’s noteworthy is of the total 13,000 claps we received, the applause of 117 readers, which represents 17% of everyone who reacted, only amounted to <1% of claps. The question is, which group shows more overall appreciation to the writer: 3 people who clap 50 times each or 117 people who clap 1 time each? The current implementation of claps favors the 3, even though that may not be the case.

This insight can be useful in understanding if Medium’s current 50-clap maximum is appropriate. Or if clapping should go up by increments of 10 to give more weight to the “single clappers”. I’m interested to see if Medium adjusts these numbers in the future.

People like to give 1, 2, 3, 5, 10, and 50 Claps

At first we suspected increments of 5 claps to be the most common, which is true until the tip off point of 10. Turns out the common points are 1, 2, 3, 5, 10, maybe 20, and 50, with a decreasing trend for the numbers in between. The number of people who give between 30–49 claps is small. This might be because after a certain point, people are more inclined to give the full 50 claps.

The Medium team has access to way more clap data than we do and so I’m curious if this this trend persists across all posts. Seeing this data makes me wonder about the following:

1. Should claps simplify to five options: 1, 5, 10, 20, and 50?
2. Should these options be some variation of ‘good, great, awesome, amazing, and outstanding’ with a number value assigned to each?

The answers can very simply be no and no. Just questions to consider.

Claps decrease as the number goes higher. The exception is 50 claps.

This is probably because of the way claps is currently designed. Naturally, giving more claps takes more time and so there is a decreasing trend as claps go up (more time commitment). We broke our numbers down into increments of 5, which presented the following graph.

The downwards trend is obvious except for 46–50, which is due to the people giving +50 claps.

Clapping from Medium Members vs Non-Members

Maybe this behavior is different for a locked article (member-only), since claps influence how much Partners get paid out. If not already done, there’s an opportunity to present to Partners a Clap Dashboard which can show information like the graph we made.

For Medium Members, there’s an inverse relationship between the number of claps they give and the $ value per clap for Partners. That is, the value of claps dilute as more are given out. Given that information, Members might actually end up applauding less per post to control the value of their claps. This actually presents a paradox: clapping less shows more support to writers because they receive a bigger portion of the $5 membership fee pie. I’d love for Medium to chime in here if my interpretation is wrong.

More Insights?

There’s likely more to be learned from a deeper analysis of our data. We’re curious to see what insights you guys can come up with.

TomYum is a Toronto-based Digital Product Design and Branding Company. Get in touch with us to discuss your next project. Follow our work on Dribbble.