What a SoundExchange Report of Use Requires (and Where Your Automation Log Falls Short)

If you run an internet radio station, SoundExchange wants a periodic list of everything you played and how many times each recording was heard. That list — the Report of Use— is how featured artists and sound-recording owners get paid for digital airplay. The catch is that the file your automation software exports is usually almost complete: it has the artist, the title, the album, the timestamps and the play counts, and it is missing exactly the two fields that make each line identifiable. This is what the report needs, and where the gap is.

What a Report of Use is

A Report of Use (ROU) is the data file a webcaster submits to SoundExchange listing every sound recording it performed, plus the audience measurement for each, so the royalties can be distributed to the featured artists, non-featured performers, and the sound-recording copyright owners. It exists because non-interactive digital transmissions — internet radio, simulcasts, and similar — are covered by a statutory licence (17 U.S.C. §§112 & 114), and SoundExchange is the collective that administers it. The recordkeeping and rate rules live in 37 CFR Parts 370 and 380. Both commercial and noncommercial webcasters file, as separate rate categories.

The required fields, per track

For every recording you report, SoundExchange asks for the following. The exact column layout is defined by their current template — treat this as the information each line must carry, not a literal header spec:

The reporting period is the filing period itself — a month for census filers, a quarter for sample filers — rather than a per-line column. The two fields that trip stations up are the ones in the recording identifier row: the ISRC, and the label that backs the album-plus-label fallback.

ISRC, or album and label

SoundExchange is explicit that the ISRC is preferred, and that the album-plus-label combination is the accepted alternative when it is missing. In their own words:

So a line is compliant with either the ISRC orboth album and label. The ISRC is the far cleaner path: it is a single unambiguous code, whereas album-plus-label only works when both are present and spelled consistently — and label is precisely the field a play log almost never carries.

Census vs sample reporting

There are two reporting regimes, and which one you fall under decides how much you report and how often:

• Census— the default for essentially all internet webcasters. You report every play, monthly, within 45 days of month-end, using actual total performances.
• Sample— available only to eligible minimum-fee broadcasters simulcasting an FCC-licensed AM/FM station. Two seven-consecutive-day windows per calendar quarter, filed quarterly.

How it's submitted

The report goes in as a delimited plain-text file, not the spreadsheet you built it in. SoundExchange publishes templates to help you assemble it, and submissions go through their Licensee Direct portal — email for smaller files, (s)FTP for larger ones. The precise delimiter and column order come from their current template, so pull that from Licensee Direct rather than hard-coding headers from a blog post. Your job on the data side is simpler and doesn't change: every line needs a clean identifier.

Where your automation log falls short

Here is the practical problem. The big automation and streaming platforms — RadioBoss, SAM Broadcaster, AzuraCast, StationPlaylist, RadioDJ — all log the artist, title, album, timestamp and play count reliably. What they routinely don't carry is the ISRC and the record label: the two fields that turn a play log into a compliant report line.

# A typical automation export — artist, title, album, plays, timestamp
artist,title,album,plays,played_at
Daft Punk,One More Time,Discovery,3,2026-07-01T20:14:00Z
Daft Punk,Digital Love,Discovery,2,2026-07-01T20:19:00Z
# No ISRC. No label. Not yet a compliant Report of Use line.

This isn't a misconfiguration on your end — it is how the tools behave. AzuraCast went as far as building a feature that fetches missing ISRCs from an external database (MusicBrainz) when it generates its SoundExchange report, precisely because the ISRC isn't in the station's own log; operators have reported the ISRC column coming out blank on export (AzuraCast issue #3963). SAM Broadcaster hasISRC and label fields, but only populates them if that metadata was already tagged into your files — an untagged library exports blanks. (These are community-reported behaviours, not official SoundExchange statements, but they match what stations run into.)

Closing the gap

The fix is to enrich the log beforeyou file: take the artist and title you already have, and attach the ISRC, album, and label. That is exactly what SonoVault returns for a track — the recording identifier and the label that backs the fallback, from nothing more than a name:

{
  "results": [
    {
      "id": "315587019",
      "title": "One More Time",
      "isrc": "FRZ030100147",        // the recording identifier SoundExchange prefers
      "releases": [
        {
          "title": "Discovery",           // album — half of the fallback
          "label": { "name": "Virgin" }, // marketing label — the other half
          "release_date": "2001-03-12"
        }
      ]
    }
  ]
}

Fold those two columns back into each row and the same export becomes a report line that identifies its recording:

# The same rows, enriched with the two missing fields
artist,title,album,isrc,label,plays,played_at
Daft Punk,One More Time,Discovery,FRZ030100147,Virgin,3,2026-07-01T20:14:00Z
Daft Punk,Digital Love,Discovery,FRZ030100152,Virgin,2,2026-07-01T20:19:00Z
# ISRC present → identified. (Album + label alone would also satisfy the rule.)

There are two ways to do the enrichment. If you're not a developer, paste your track list into the free Bulk Lookup tool, tick the ISRC and Label columns, and download the enriched CSV. If you want it automated, the same resolution runs over the API. Both paths, plus the stream-monitoring route that skips the export entirely, are covered in the companion guides:

• ISRC vs. album + label: SoundExchange's two ways to identify a recording
• Add ISRCs and labels to a RadioBoss export for SoundExchange
• Monitor your stream URL to auto-build a SoundExchange track list

SoundExchange itself offers ISRC Search, a Repertoire Match service, and ROU Augmentation to backfill identifiers after you submit. Those are useful — SonoVault's angle is the opposite end: getting the ISRC and label onto the line before it goes in, so fewer rows come back flagged as incomplete.