We post our work. Read it, dispute it, copy it.

Most royalty estimators in the indie space hand you a single dollar figure with no footnotes. We don't. The Royalty Reckoning tool returns a range, a slider, and this page. Free to read, free to critique, free to fork into your own tool. We publish it because the industry doesn't.

The promise we won't make

We will not tell you "you are owed $4,200." We will tell you the recoverable range is, for example, $2,700 to $5,700, with a midpoint around $4,200, and we will show you a slider for the recovery window because nobody knows in advance whether your unclaimed pool is three years deep or ten. The single-number version is a marketing artifact. The range is closer to the truth.

That range is a stated ±35% band around one central estimate, and it is deliberately not wider. A range can fail in two directions. Too narrow pretends to a precision the data doesn't support. Too wide is just a shrug with dollar signs on it — and it's the easier mistake to make look rigorous, because every individual assumption in it is defensible. If you ever see a spread from this tool that's ten times wide or more, that's a bug, not caution. We shipped one on 10 August 2026 and it's in the revision log below.

Where the numbers come from

The model takes one input you give us — an artist name or a Spotify URL — and pulls everything else from public sources. From there it has four stages.

1. Catalog scan. Spotify's public catalog API is used to pull the artist's primary-artist tracks, their popularity scores (0–100), their ISRCs, and their earliest release dates. "Appears on" credits are excluded from the primary count; we track them separately.
2. Per-track stream estimate. Each track's popularity score maps to a per-year stream range. The mapping is fit against real publisher royalty statements from independent songwriter catalogs Doom Tide has recovered for — not against industry-published guesses. Older tracks get scaled down by a track-age multiplier, because a popularity-30 track from 2010 doesn't stream like a popularity-30 track from 2024.
3. Dollars per year. Stream counts get converted to dollars across two royalty types: mechanical (US-share-aware, built from the real per-DSP rates The MLC publishes each month and weighted by a typical platform mix, since the DSPs don't pay the same rate per play) and performance (PRO blanket pools, global). The figures behind those conversions come from real rate sheets and statements; they are not the U.S. statutory headline rate, because the headline rate is not what an indie songwriter actually receives.
4. Recovery window. The floor is three years — the active unclaimed pool the MLC and U.S. PROs hold before market-share redistribution kicks in. The ceiling extends up to ten years for never-registered catalogs, because the MLC absorbed pre-existing unclaimed mechanical pools when it came online under the Music Modernization Act of 2018. The slider on the result page lets you pick where in that three-to-ten-year window your catalog actually sits.
5. Registration branch. Then the model asks whether you're already registered, and prices two completely different things depending on your answer. If you've never filed, the whole stack across that window is unclaimed and the number is what you'd recover starting from scratch. If you have filed, your works are already collecting, and what's recoverable is only the part that leaks past a live registration. Those are different questions with answers an order of magnitude apart. Until July 2026 this tool only asked the first one.

The change we made in July 2026, and why

Updated 11 August 2026. The section below describes a recalibration we ran in July. Most of it still stands, but one conclusion we drew from it was wrong, and the correction is worth reading before the rest: we treated "already registered" as meaning "already collecting." Those are not the same thing, and the difference is most of the money. A registered catalog with a dead publisher on the filing collects nothing at all. The gap figures below apply only to songwriters who are actually receiving payments — which, among our own paying clients, is one in eight. Everyone else is priced on the full stack.

Between June and July 2026 we audited nine catalogs work by work — not estimated, audited: every composition checked against its MLC registration, its PRO registration, and its actual matched recordings. About seven hundred works. Then we compared what we found against what this tool would have told those same songwriters.

All nine were already registered. Not one had simply never been filed. That is the assumption this tool was built on, and it held zero times out of nine.

It matters more than it sounds. On one catalog, the estimate-style number came to somewhere between $270,000 and $470,000 recoverable. The work-by-work audit put the provable figure at $18,938 to $35,170 — seven to fourteen percent of the estimate. Nothing was wrong with the arithmetic. The catalog was registered at 100% and already collecting. The problem was that we'd been answering a question the songwriter hadn't asked.

What a registered catalog actually leaks

A registered catalog still leaks. It just leaks differently, and much less. What we found:

• Unclaimed publisher share on a live filing. The registration exists and looks fine, but only part of the ownership is on it. One catalog was missing its publisher entirely from thirty of thirty-six filings.
• Recordings matched to somebody else's work. Your recording, registered correctly, matched at the MLC to a different songwriter's composition. That money pays out to them. We found this on five of the nine catalogs.
• Recordings not matched to anything. The work is registered, the streams are real, and no recording connects the two. On one catalog this was the single largest category of loss.
• Writer identities that don't resolve. Placeholder writers, duplicate IPIs, one person existing in the registries as nine different people. This was the most common finding across all nine catalogs and, in most cases, the least recoverable in dollars — the money is usually still flowing, just not through anything you control.

Across the seven catalogs where we could compute it cleanly, the leak ran 2% to 12% of catalog income, clustered around 5%. That is the range the tool uses when you tell it money is already reaching you — and only then.

One thing to be clear about, because we got it wrong ourselves. Those percentages are what we could prove from public registration data. Every one of those nine audits says in its own text that the figure is a floor and that its largest findings carry no dollar amount at all. Where we have since gotten inside the actual royalty statements, real recoveries run roughly an order of magnitude above the public-data floor. If you are collecting and this number looks small, that is why. It is the part we can show our work on, not the part we think is there.

Why we don't publish the average

The average across all nine is 17.9%. We don't use it, because two of the nine had a systemic filing defect — a publisher share missing across nearly the whole catalog in one case, co-writers with no PRO affiliation at all in the other — and those two came in at 48% and 58%. The rest came in under 12%. The distribution has two clumps and nothing in between, so the average describes none of the nine.

If you have that kind of systemic defect, your number is 40% to 60%, not 5%. This tool cannot detect it, because it reads Spotify and the defect is visible only at the MLC. We're telling you it exists rather than quietly averaging it into everyone's headline.

What the audits found that we still can't price

Every number above is a floor. Three of the most common findings across the nine catalogs carry no dollar figure at all, because public data can't size them: recordings matched to the wrong work, writer-identity fragmentation, and foreign recordings contaminating a registration. On three separate catalogs the most important finding in the report was one we refused to put a number on. We would rather publish a number we can defend and say plainly what it excludes.

One more limit worth stating. The 2% to 12% range was measured on the mechanical side. The tool applies it to mechanical and performance together, which assumes performance leaks at the same rate. The audits didn't measure that.

What we exclude on purpose

Every exclusion below is a number a less careful estimator would inflate the headline with. We don't, because each one is harder to recover than it sounds.

• Cover songs. Royalties flow to whoever wrote the song, not whoever recorded it. If the catalog has covers, those royalties belong to the original writers — not the artist on this report. Spotify metadata doesn't reliably distinguish covers from originals, so we don't auto-subtract them. Subtract them mentally.
• Pre-2021 mechanicals already paid. The MLC's "historical unmatched" pool gets cited a lot, but most pre-2021 mechanicals were already paid through HFA or major-publisher channels. Treating "every stream before 2021" as recoverable is the most common overestimation in indie royalty marketing. We don't.
• Synchronization income. If you already have a sync placement, you already have the contract. Our estimate doesn't try to predict future sync deals you haven't signed.
• Neighboring rights and master-side digital performance income. These pay master-side, not composition-side. The Reckoning is a composition-side tool. Doom Tide does not currently administer this revenue stream.
• Foreign collections routed through a sub-publisher. If you have a foreign sub-publishing chain, those territories' royalties flow through your sub-publisher's deal — not directly recoverable by you. The per-DSP mechanical rates we use are the U.S. (MLC) rates, so the estimate stays inside Doom Tide's U.S. scope; we don't double-count by assuming foreign CMOs are also leaking the same income.

What can throw the estimate off

• Popularity 60+ is extrapolated. Our calibration sample is mostly indie/rock songwriters in the popularity 0–60 range. Above that, the mapping is extrapolated. Big-catalog or high-popularity artists should treat their numbers as the least precise on the page.
• No territory split per artist. The model assumes a U.S. share of streams. Artists with a heavily non-U.S. listener base will be over- or under-estimated depending on which way their split runs. Spotify popularity is global; U.S. share is not artist-specific in our current model.
• Classical, instrumental, and ambient catalogs. The popularity-to-streams fit is calibrated on indie/rock. We haven't yet fit a separate curve for classical or ambient catalogs. A revision is open.
• Brand-new artists. Below three years of catalog age, the recovery window collapses to its floor. The estimate gets conservative on purpose — there's no ten-year tail to recover from yet.
• Missing Spotify release dates. A handful of older tracks have no release date in the Spotify catalog. We treat those as zero years old in the decay model, which is conservative. Real release dates from your distributor would tighten the estimate.

The assumptions that matter most

Two assumptions are doing most of the work. If either one is wrong for your catalog, the estimate moves a lot.

1. 100% composition ownership. The estimate assumes you wrote the songs and own 100% of the publishing. Co-writers and assigned publishing change that. If you wrote half the songs and half of those have a co-writer, multiply by your actual composition share to get a defensible number.
2. Whatever you told us about registration. This used to read "the estimate assumes the works are not currently registered." It no longer assumes — it asks, and prices the two cases differently. But the answer is still yours, and it's the input that moves the number most. If you're not sure, we price you as registered, because that's the better bet, and show the never-registered figure as a separate line rather than stretching one range across both. Searching your name at The MLC's public portal takes about two minutes, costs nothing, and collapses the range to one end or the other. Do that before you rely on any figure here.

Why we publish this

The industry standard for independent-songwriter royalty estimators is a single big number on a marketing page, framed as money you're owed, with no model behind it that anyone can read. The reason for that standard is obvious: the bigger the number, the more conversions. The reason we don't follow it is also obvious: independent songwriters have been getting that pitch for two decades, and most of them still aren't recovering their actual income.

So we publish the method. It's not proprietary because the methods aren't actually scarce — what's scarce is the discipline to refuse the inflated number. If a competitor wants to copy the structure of this page, they should. The work is in maintaining it.

We update this page when the math changes. Sometimes that's monthly. Sometimes it's quarterly. The revisions section below is the record.

Revisions

Every change to the model, the disclosure, or the wire response gets a dated row. The most recent change is at the top.

Suggest a correction

Found a mistake, a missing exclusion, or a fresher data point? Send it. We read every submission. The ones that lead to an actual model change get a dated row in Revisions above — with attribution if you give us a name.