You sign up, choose a category from a short list, and the app hands you a form. Records, cards, comics, games. The form turns up with fields already in it, chosen by whoever built the app, and those fields decide what you can record, what you can sort by, and what you can still prove about an item two years from now.
That list is the product. The rest of the app is a view onto the same data.
The form you did not write
A record makes the trade visible. To add a release to Discogs you supply a main artist, a title, a label, a catalogue number, a format, a genre, and the full tracklist. Style, country and release date are optional, and the app requires the rest.
Those requirements are correct for a public database of pressings. They are also a decision about what a record is. By that list, a record is a release, and the thing that separates one copy from another is the label and the catalogue number printed on it.
Where does the shop you bought it from go, or the note that this is the copy your dad played to death? Into a free-text box, where no one can sort it or count it. The same problem applies to the pressing plant, the sleeve condition, and the price you paid.
Nothing has gone wrong. You joined a database of releases and got a database of releases.
Why the field list stays the same
Card apps fix the taxonomy in a different place, and the mechanism is worth understanding because it explains the rest of them.
Collectr covers more than twenty trading card games against a catalogue of 200,000 products. You scan a card and it matches the set, the number and the variant, then prices the card from real sales. TCGplayer works the same idea from the selling side. Listing a single card means searching their catalogue for the product and attaching your price to that entry, or matching it to their market price.
The catalogue match is what makes the pricing work. It is also the boundary. Collectr can only hold what its catalogue holds, and TCGplayer can only list what has an item page. Anything without a manufacturer, a set code and a number has no path in. The card you drew yourself, the misprint and the toy in the shape of a card sit outside the model, and the app cannot hold them.
A fixed field list is what makes a scanner, a price feed and a shared public catalogue possible. Those features only work because items mean the same thing everywhere in the database. The taxonomy is the price of the database, and for a single hobby it is a price worth paying.
The problem turns up at home
The costs arrive when you own more than one kind of thing, which most collectors do. Vinyl, cards, coins, LEGO sets, games, books, keyboards. Own three of those and the fixed-list route wants three apps.
A vinyl app, a card app and a LEGO app would handle three of them and leave the rest nowhere to go. That is three subscriptions, three logins, and three totals with no relationship to each other. The question “what this collection is worth” has no single answer, because there is no single place it lives.
The other half of the problem is quieter. Hobbies with no app end up in a spreadsheet, and a spreadsheet is not a worse version of an app. It is a different thing that happens to accept text in cells. You can write anything in a column and sort none of it by anything that matters.
A schema you build yourself
Collectd starts from the other end. A collection is a name, a description, and the fields you define, and you decide each one when you create it.
You name each field, choose what type it is, and set the order it appears in on the add form. There are ten types to work with: text, long text, number, dropdown, date, currency, rating, year, range and true or false. The types do the work that free text cannot. A dropdown gives you a fixed set of options you wrote, so condition becomes something you can filter rather than read. A currency field holds a number you can total. A date field holds a date you can sort, which free text cannot do.
Two collections in the same app can share no fields. Round pound coins track the design, the year and the mint. Mechanical keyboards track the model, the switch and the layout. Each list comes from whoever built the collection, and neither has to bend to fit the other.
Tags run across collections rather than inside one. A tag for items you plan to sell, or items you inherited, or items that were gifts, can hold items from any hobby you track. Public galleries turn a collection into a page you can send to someone, and you can copy a schema from another collector's collection when theirs matches yours. That last one is how a schema gets shared without anyone sitting down to design it. Someone builds a Needoh collection, and the next person to buy a squished fidget toy copies it.
A set you can finish, and a set that keeps going
Some collections come with a number attached. The round pound coin ran from 1983 until the twelve-sided coin replaced it in 2017, and it carried twenty-four reverse designs in that time. Collectors chased the set rather than the coin, so what matters is which designs are in the album and which are still missing.
An app that only records what you own cannot show you that gap. Turn it around instead. Add the whole set, give the collection a true or false field for the ones you have, and the gap becomes a filter rather than a memory test. Twenty-four items, seventeen ticked, and the seven you are missing sit behind one click. A tick box is a filter like any other here.
Other collections have no end. A LEGO theme ships new sets each year, a game ships expansions, and a run of books keeps going. The schema does not assume a final count, so those collections grow as you add to them, and the same app that tracks twenty-four coins holds a shelf with no final number in it.
You can also copy a public collection and take the items with it, if the owner permits copying. The twenty-four designs someone else is chasing become the twenty-four you chase, and two collectors work from the same list to see what each of them still needs. That is a kind of competition a shelf on its own cannot produce, and a set gets finished more often when someone else is finishing it. A public collection comes with its copy count and its star count, and the owner decides on both.
What a custom schema costs
The trade is real, and it lands in the same place the catalogue apps win.
Fields you design do not come with a price feed. No market catalogue will look your item up and fill in a value, because none of them covers Needohs, round pounds and mechanical keyboards in the same schema. You describe the collection your way and you do the valuing yourself.
You also carry the design work. A schema that holds one item and a schema that holds a thousand are the same amount of setup, and the cost of getting it wrong shows up later, when the collection is large enough to be worth sorting and the field you need does not exist. That is the part worth thinking about before item fifty, and it is what the schema postcovers.
The apps that fix the field list are not wrong. They are specific, and specificity is how most of them get good. The question is whether the specificity matches the shape of what you own, and for a collector with a shelf that does not fit one category, it does not.
A collection app should be able to hold anything you collect, without asking you to call it something else first.