Media
5 items
Description
Scores every Medium post against the posts around it, so a good week doesn't just mean you had more followers than last time. A thousand views at 5,000 followers is a hit. The same thousand at 10,000 is a quiet disappointment. Medium shows you the numerator and hides the denominator, so your own archive becomes unreadable within about a year — recent posts always look better, because you're always bigger. Breakout fixes that with two corrections. AGE-MATCHING. Posts are compared at the same age: views in their first 7 days, not views-to-date. Otherwise old posts win by having existed longer. A MOVING BASELINE. Each post is scored against the five posts you published before it and the five after — not against your all-time average. Your audience size, your posting rhythm and Medium's distribution mood at that moment are all baked into those neighbours, so they cancel out instead of accumulating. What you get is one number per post. Reach × — 1.0× is a normal post for you, right then. 4× is not. Breakout and underperform thresholds are learned from your own spread, so a writer whose numbers swing wildly needs a bigger spike to earn the label than one whose numbers are steady. THREE SCREENS, FOUR NUMBERS • The graph — every post you've published, publish date across, reach up a log axis, your own normal range shaded. Breakouts sit above it. The posts you wrote when you were small stop being invisible. • All posts — your archive re-sorted by reach instead of raw views. • One post — its pace against the 25–75% band of the posts before it, plus depth, off-Medium share, and follower conversion. The four numbers: reach × (D7 views over your local baseline), depth (read ratio against your own median), off-Medium share (search and external links versus the Medium feed), and follow rate (new followers per 1,000 views). A 4× spike with a collapsed read ratio is distribution, not resonance — and it is worth knowing which one you got. Deliberately absent: claps, earnings, per-day tables, CSV export, settings that change the maths. Each is a column that would need explaining, and none of them changes a decision. HOW MANY POSTS YOU NEED Five, before anything can be scored — there have to be neighbours to compare a post against. About ten before Breakout will call anything a breakout, because a threshold learned from your own spread needs enough posts to learn from. Below that it shows you the real reach numbers and tells you how many posts are left to go, rather than labelling noise. PRIVATE BY CONSTRUCTION There is no server, no account and no analytics. The extension reads your own author statistics from medium.com using the session you already have, caches them on your machine, and sends nothing anywhere. It touches no site other than medium.com. Open source, MIT licensed: https://github.com/sbalnojan/medium-breakout Breakout is an independent project. It is not affiliated with, endorsed by, or connected to Medium.