Why your Spotify Wrapped number is always lower than your real total.

Every December, Spotify tells you how many minutes you listened to this year. If you have ever compared that figure with a full data export, you will have noticed a gap. Sometimes small, sometimes large. It is not a glitch. It is a predictable consequence of how Wrapped is built, and once you understand the mechanism the difference makes complete sense.

Wrapped is a marketing campaign first and a statistics product second. It is designed to produce a shareable number, not an auditable one. The extended streaming history, which Spotify is required to provide under the GDPR, is the auditable record. It contains every stream since your first play, with millisecond-level durations, skip flags, device identifiers, and the country where each stream happened.

The cutoff problem

Wrapped has a cutoff date that Spotify does not publish officially. Based on community reports it falls somewhere in late October or early November, meaning six to eight weeks of listening are excluded before Wrapped is even calculated. Your export, by contrast, runs to the date Spotify generated it, which is typically within a day or two of your request.

Six to eight weeks of November and December listening simply do not exist inside Wrapped. They are not missing. They are in your export.

Counting rules: streams versus plays

Wrapped applies a minimum-stream-duration filter. A track you played for less than thirty seconds typically does not count. Your export records every stream regardless of duration, including the three-second accidentals you skipped immediately and the songs you restarted seventeen times before letting them play through.

The table below shows how those counting rules play out across one sample export. All figures are from the export itself: none are estimates, averages, or projections. Your own numbers will differ.

What the export actually contains

Each line in the extended history JSON file is a single stream event. Here is a representative entry, anonymised and lightly truncated for readability. The field that matters most for the total-minutes calculation is ms_played.

{
  "ts": "2023-08-14T07:22:11Z",
  "ms_played": 214893,
  "master_metadata_track_name": "Midnight Rain",
  "master_metadata_album_artist_name": "Taylor Swift",
  "reason_start": "fwdbtn",
  "reason_end": "trackdone",
  "skipped": false,
  "conn_country": "GB"
}

The reason_start and reason_end fields tell you how a track began and ended. The combination of fwdbtn into trackdone is how a genuine skip is derived, because Spotify’s own skip flag was not populated for years of older history.

How to read your own gap

If you have your export, you can calculate the gap precisely. The steps are:

1. Sum ms_played across every entry for the year, then divide by 60,000 for minutes.
2. Filter to entries where ms_played is at least 30,000, sum again, and you have the Wrapped-style count.
3. Exclude entries whose track name is empty or whose artist field is a podcast. That approximates the music-only figure Wrapped shows.
4. Compare the two. The larger is your real total, the smaller is close to your Wrapped headline, and the difference is the four factors above.

In practice most people find the export total is between 25 and 45 percent higher than the Wrapped figure. The exact ratio depends on how much podcast listening the account holds and how late in the year the Wrapped cutoff fell.