Reading trackers and the streak

Somewhere in December a person is choosing a short book. Not because they want a short book, but because the yearly challenge says fifty and the shelf says forty-six and there are nine days left. Nothing dishonest is happening. The tracker is doing exactly what trackers do: it made reading visible, then made it countable, and the counting has started to read back.

Start with what the trackers get right, because it is real. Reading is a habit competing with machines engineered to win the evening, and a habit that is recorded happens more often than one that is not. Self-monitoring is one of the oldest and least controversial moves in behaviour change; you do not need a laboratory to observe that the run you log beats the run you meant to go on. A reading tracker app applies the same lever to books, and for a great many people the lever genuinely moves.

The field splits into two families. The shelf-keepers — Goodreads is the standing example, StoryGraph the thoughtful challenger — track books: read, reading, want-to-read, ratings, a yearly challenge with a progress bar. The session-keepers, which include most app-based readers and a crowd of dedicated reading habit trackers, count minutes and days: tonight’s twenty minutes, this week’s total, the streak. Both families measure honestly. The interesting question is what the measures mean.

StoryGraph deserves its reputation as the considered one: it tracks moods and pace as well as titles, lets you set the yearly challenge low or not at all, and its statistics lean towards description rather than judgement. Goodreads carries the weight of scale — the shelves your friends are actually on — and the design inertia that comes with it. Neither is doing anything underhand. They are both, in the end, census instruments: superb at what, silent on how well.

What the number quietly does

A measure is a proxy, and a proxy under pressure stops being a good one — the observation is old enough to have names attached to it in economics. “Books finished” rewards short books, punishes long ones, and turns abandoning a bad book — a skill every serious reader needs — into a small defeat. “Minutes logged” rewards time in the chair regardless of what the time contained. Neither number knows whether anything was understood, and understanding was the point.

You can watch the redefinition happen around any yearly challenge: December sprints of novellas, the guilty re-shelving of an abandoned tome, audiobooks run fast to keep the total alive. None of this is a scandal. It is what proxies do to conscientious people. The reader who games a book count is not lazy — they are doing exactly what the instrument asked, and the instrument asked for the wrong thing.

The second-order effects are subtler. Once finishing is the unit, the definition of finishing starts to flex — do skimmed chapters count, does a book abandoned at ninety per cent, does the audiobook you half-heard while cooking? Communities argue these cases with real heat, which tells you the number has become a stake rather than a note. A measure you have to litigate is a measure that has started to govern.

In defence of the streak

The streak deserves a fairer hearing than it usually gets from people who write essays about reading. A streak measures showing up. That is all it measures, and showing up is not nothing — practice that does not occur has no effect size. For a reader rebuilding a habit after years of fragmentary scrolling, a chain of ordinary days is a legitimate thing to protect, and the mild dread of breaking it is doing honest work.

The dishonesty arrives when showing up is dressed as achievement — when the interface implies that day forty of a streak means more than day three, or that the number measures you rather than your calendar. It also arrives at the reset: a missed Tuesday erases six weeks of true fact — you did read on those days — because the instrument values continuity over total. Elsewhere in habit software, streak repairs are sold outright, which at least prices the fiction honestly. The generous reading is that a streak is scaffolding. Scaffolding is for taking down.

It helps to notice which unit of time each instrument answers to. The yearly challenge is an annual instrument: it makes January generous and December desperate, and it says nothing at all about any given evening. The streak is a daily instrument: it cannot see the year, only the chain. Neither is wrong, but they pull in different directions — a person protecting a streak will read ten minutes of anything, while a person chasing a yearly number will skip the slim Tuesday session that no longer moves the bar. The instrument you display is the tempo you will keep.

The reader who games a book count is doing exactly what the instrument asked.

Proxies, stated plainly

Pages are not comprehension

Here is the gap all of these instruments share: nothing in a shelf, a page count, a minute total or a streak knows whether the material went in. The testing literature is blunt on the adjacent point — being tested on material does more for retention than re-reading it — and yet the field’s measurement effort points almost entirely at quantity. We measure what is easy to log, then treat it as the thing we cared about. Finishing is legible. Understanding is not, unless you ask.

None of this makes quantity tracking worthless — a proxy can be wrong in detail and right in direction, and someone logging twenty minutes a night is almost certainly reading more than they were. It means the proxy needs a companion. Speed has the same problem in our corner of the field: a words-per-minute figure without a comprehension figure next to it is exactly as hollow as a book count, which is a cost we take apart in the comprehension tax.

So the honest use of any tracker starts with saying, out loud, which proxy you have chosen. If the goal is a durable evening habit, minutes and streaks are the right proxies and book counts are noise. If the goal is finishing the reading a course demands, the count matters and the streak is decoration. And if the goal is understanding — for work, for study, for the sake of the book — then none of the standard numbers touch it, and you need an instrument that asks, which is the argument of testing your own comprehension.

What we log instead

Our own answer is narrow and worth describing plainly. After a passage, the app runs a recall check — words that genuinely appeared mixed with plausible distractors — and shows the score next to the speed, because either number alone flatters you. Sessions are kept on your device and exportable as a file you own; there is no server copy, no feed, no follower who can see your shelf. The pairing is the point: speed with recall attached is a measurement. Speed alone is a boast.

There is also a social cost we should name rather than wave at. Goodreads and StoryGraph are not really measurement tools; they are company — friends’ shelves, reviews, the pleasure of a public year of reading. Our log is private by construction, which protects you from the audience and also deprives you of it, and for many readers the audience is precisely what makes the habit hold. That trade is stated, not solved, in offline by design.

The test for any reading habit tracker, ours included, is whether you can say what its number is a proxy for without the sentence going soft. “Days I showed up.” “Books I finished.” “Words I could pick out of a line-up ten seconds later.” All three are honest when stated and slippery when implied. Keep the tracker that lets you finish the sentence — and every December, notice what length of book you are reaching for.