Why RepoRanker retired its credit economy

One card payment has ever bought a position on this board. It was $3, on 2026-08-27, on the founder's own repo, and it is still the only one. Every other paid row up there was granted by a credit conversion rather than sold, and no card was involved in a single one of them.

Those rows were funded from balances the 2026-08-27 credit conversion granted, and they were written onto the board in bulk rather than bid for one at a time. That is not revenue and it is not demand. It is the settlement of a debt. As of 2026-09-03 the board carried 196 paid rows: 195 of them from the conversion, and one of them that $3. For completeness, because "one payment" should not be read as more than it says, the records also hold two older Stripe checkouts dated 2026-04-25, both on the retired boost product and both the founder's own. No account other than the founder's has ever paid RepoRanker anything.

Leading with that is not modesty. The whole pitch of this board is that every price on it is public and checkable, so the one thing we do not get to do is quote our own gross number as if it meant customers. The full accounting, down to a per-account CSV, is on the pivot page. This post is the other half of it: what the old product failed at, and what a public price list replaced it with.

What the old product was

Until 2026-08-27, RepoRanker was a peer-review platform with an internal economy. You submitted a repo. Other developers wrote reviews of at least 800 characters. A review sat in a 48-hour dispute window before it counted. When it released, the reviewer earned credits, and enough of them moved the reviewer up through reputation tiers. Credits bought five kinds of boost: leaderboard, featured, newsletter, review request, and backlink. Reviews fed the ranking formula, so writing them moved repos around the board.

All of that is retired. The 800-character minimum, the dispute window, the release step, the credit award, reviewer reputation and tiers, and all five boost types. Nothing earns credits and nothing spends them.

What it actually failed at

Not the reviews. Reviews still exist, anyone signed in with GitHub can write one, and they read exactly as they did before. What failed was the economy wrapped around them.

The mechanism had two prices and neither one was real. A review cost 800 characters of somebody's evening and paid out a token the platform minted at will. A boost cost that token. At no point in the loop did anyone have to decide whether a position on this board was worth actual money, which means nothing the board displayed was evidence of anything. The repo on top was the repo holding the most of a number we could have set to whatever we liked.

The volume said the same thing from the other direction. 319 reviews, ever, written by 16 accounts, against 307 repos submitted and 254 registered users on the day of the change. A ranking formula fed by reviews was ordering a board that almost nobody had reviewed. Asking a few hundred people to keep writing 800-character reviews to keep that formula fed was never going to compound.

What replaced it

One mechanic, and it fits in four sentences. Listing a repo is free. A place near the top of the board starts at $3. Rank among paid rows is cumulative dollars spent and nothing else. A placement never expires: when somebody spends past you, you move down the board rather than off it.

Everything below the paid rows is ordered on GitHub momentum, meaning stars gained since the repo was listed plus a push-recency bonus. Reviews are not a ranking signal and will not become one. The exact ordering is written down on the ranking rules page, and the price of every position, including what it costs to pass the row above you, is on the placement ladder.

What a public price changes

It makes the board falsifiable. Under credits, "top of the board" meant a score computed from inputs a reader could not audit and could not have bought even if they wanted to. Now every paid row states its total in dollars, and the amount required to pass it is printed on the same page. Where a total is not money somebody handed over, that is disclosed rather than averaged in: the conversion rows are named as conversion rows, and on 2026-09-03 every fractional total on the board was raised to the next whole dollar at no charge to its holder, which is a gift from the house and is written into the ledger as one. If you think the position above you is overpriced, you can look at the price and decline. That is not a thing you can do with a ranking formula.

It also changes what failure looks like, which is the part that matters more. A formula fails quietly: the numbers keep moving and nobody can tell whether they mean anything. A price list fails out loud. If nobody is paying, the board says so, in dollars, on the front page. That is what it is doing right now. It is the mechanic working, not a bug to design around.

Where that leaves the number

195 of the 196 paid rows the board carried on 2026-09-03 came from the conversion rather than a sale, so the board looks busier than the business is. The honest read of demand to date is one payment. We are not going to average that out, bury it inside a gross total, or describe the conversion as a launch. Those balances existed because people had earned credits, the credits were being retired, and the value had to go somewhere other than our pocket. Paying it out as placement dollars was a debt being settled, and settling a debt is not a sale.

Which is why this post quotes no total for the board, and will not. A gross figure pinned into a static post is stale the moment anybody pays, and worse, it moves for reasons that have nothing to do with anybody paying: the 2026-09-03 round-up lifted the board's sum purely by rounding, with no card charged and no balance debited. The count of card payments is the number that holds still. Check the rest against the ladder, which shows what every position costs today, and the board, which shows who holds them.

The number that would actually mean something is the second card payment for a position on this board, from an account that is not the founder's. When it lands it will be visible the same minute, at the amount it cost, sitting next to every other price on the board.

List your repo for free, or read the full pivot accounting first. Reading it first is the better order.

Related: The pivot: what changed and what it cost · Leaderboard ranking rules · The placement ladder.