Most readers use Goodreads. Most Goodreads users are mildly dissatisfied with it and cannot quite say why.
The reason is that it was built to solve a different problem. Goodreads is a discovery and social product: it wants to tell you what to read next based on what other people rated highly. It is reasonably good at that. What it is not built for is being your own record of what you have read, why you read it, what was in it, and how it connects to everything else you have read — and that is what most people actually want from it.
This is about building that record. It is a different activity from curating a physical collection, which is about objects and shelves; this is about data, retrieval, and the annoying fact that any system requiring effort per book will eventually be abandoned.
The star rating is the core design flaw
You finish a demanding novel that reorganised something in your head and gave you a difficult week. Five stars. Two weeks later you finish a comfort thriller that you enjoyed enormously and will never think about again. Five stars.
The number has now destroyed the only information you had. A single ordinal scale collapses several independent dimensions — how much you enjoyed it, how well made it is, how much it changed your thinking, how useful it will be later — into one figure that answers none of those questions when you come back in three years.
What replaces it is not a better scale. It is a sentence. One line on what the book actually did for you, written while it is fresh, is worth more than any rating system, and unlike a number it will still mean something to you a decade later.
What to record
The temptation is to design an elaborate schema. Resist it. Every field you add is a small tax paid per book, and the system dies from accumulated tax.
The fields that have consistently earned their place:
Title and author. Obviously.
When you read it. Not when it was published — when you read it. Your reading has phases, and knowing that you read four books on a subject in one particular autumn tells you something about that period of your life that the books alone do not.
Status. Read, reading, abandoned, want to read. Abandoned is an important category and most systems do not have it; knowing what you gave up on and roughly where is genuinely useful, both for deciding whether to try again and for noticing patterns in what defeats you.
Why you picked it up. One phrase. Recommended by someone, cited in another book, needed it for a project, saw it in a shop. This turns out to be the field I most often want and the one most systems lack, because it tells you which of your discovery channels actually produce books worth reading.
One line on what it did. The replacement for the rating.
A couple of tags. Subject, loosely. Not a taxonomy — three or four words that would let you find it later.
Optional but valuable if you will actually maintain them: quotes worth keeping, and links to other books it connects with. The connection field is the one that turns a list into something more interesting, because your reading has a structure and you cannot see it from a chronological list.
Tool tiers, honestly assessed
A spreadsheet. Set up in twenty minutes, works forever, exports trivially, no vendor can take it away from you or change the terms. Sorting and filtering cover most needs. This is the correct answer for the overwhelming majority of readers and the one people reject because it is not exciting. It becomes unpleasant somewhere past a few hundred entries, mostly because long free-text fields in a spreadsheet cell are miserable to write in.
A note-based system. Obsidian, or Notion, or anything where each book is a document with structured fields at the top. This is where I ended up. The advantage is that the note about the book and the record of the book are the same object, so there is no separate place where your thinking lives, and links between books are native. Obsidian in particular stores everything as plain text files on your own disk, which matters more than it sounds like it does — the ten-year question for any system is whether you can get your data out of it.
A dedicated reference manager. Zotero, free and open source, was built for academic citation management and is genuinely excellent at what it does: capturing metadata automatically, organising by collection, attaching PDFs, generating citations. If a meaningful portion of your reading is papers and technical material, this is the right tool and nothing else comes close. For a shelf of novels it is heavy machinery.
Calibre is a different thing again and worth mentioning: it manages ebook files rather than reading records, and if you own a large collection of ebooks it is essentially the only serious option.
An actual database. A SQLite file and a few queries. I have technical background and I have done this, and I will be honest about the outcome: the queries were satisfying to write and I used them roughly four times. Unless you have thousands of books and a specific analytical question, the effort goes into building the system rather than using it. If you do go this way, the real argument for it is not query power but that a plain database file is the most portable format that exists.
Notes are a separate problem
Cataloguing tells you that you read a book. Notes tell you what was in it, and they are the harder and more valuable half.
Two things I have learned about book notes over a long period of doing them badly.
First, highlights are not notes. A book of underlined passages is a record of what struck you in the moment and it is nearly useless later, because it has no structure and you will not reread it. The valuable artefact is a short summary written in your own words after finishing — a few paragraphs on what the book argued, what you took from it, and what you did not believe. That act of compression is where the retention actually comes from, and it takes twenty minutes.
Second, write it within a few days. The window in which you can still reconstruct a book's argument closes faster than anyone expects. A note written a month later is a note about the impression the book left, which is not the same thing.
Maintenance is the whole game
Setting up a system is enjoyable and creates a strong illusion of progress. Maintaining it is the actual work, and the failure mode is universal: an elaborate structure built in one enthusiastic weekend, maintained for five books, abandoned.
So choose the level you will sustain on a bad week, not on a good one. For most people that is: add the book when you finish it, write one line, add two tags, close it. Two minutes. If you reliably do that for two years you will have something genuinely valuable. If you design something requiring twenty minutes per book, you will have six beautifully documented books and a year-long gap.
Anything more ambitious should be an occasional indulgence rather than a requirement — the longer note when a book deserves one, the periodic review of what you read this year and what the pattern says about you.
What it is for
The point of the record is not the record. It is that in three years someone will ask you a question and you will half-remember that a book addressed it, and you will be able to find not just which book but what you thought about it at the time.
That is a small thing that accumulates into something surprisingly large. A decade of two-minute entries becomes an external map of your own thinking, searchable, in your own words — which is not something Goodreads was ever trying to build for you.
Tags
Taresh Sharan