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

Every December, Spotify tells you how many minutes you listened that year. Add up the same year from your extended streaming history and the export comes out higher. It is not a glitch, and most of it is not a mystery either. Spotify has said what Wrapped counts, and the export lets you measure what it leaves out.

What Spotify says Wrapped counts

Spotify's newsroom set this out when Wrapped 2025 launched:

• The window runs from 1 January to mid November, to leave time to check the data before launch. Spotify does not publish the exact cutoff date.
• A play counts once you have listened for more than 30 seconds.
• Total minutes include music, podcasts and audiobooks, Private Mode sessions and anything you excluded from your Taste Profile.
• Offline listening is included.

One thing it does not say is whether the time from plays shorter than 30 seconds goes into your minutes. Keep that in mind, because it turns out to matter far less than you might expect.

The cutoff, measured

These are four calendar years from one export, with every stream included, private sessions too, since Wrapped counts them. The second column stops at the end of 15 November, which is our reading of "mid November". A stream is placed by when it started, because the export's timestamp records when a stream ended.

Between an eighth and a fifth of each year's listening happened after Wrapped had stopped counting.

The share moves because it depends on how you listen at the end of the year, not on anything Spotify changes. In 2024 this listener's last six weeks ran at nearly twice their usual daily pace, so more than a fifth of the year fell outside the window. In 2025 the pace barely changed all year, and the same six weeks came to about an eighth, which is roughly their share of the calendar.

The 30-second rule changes plays, and barely touches minutes

Plays under 30 seconds are common. They are also, by definition, short, so they are many and they weigh almost nothing:

So if you are comparing play counts, the 30-second rule can explain a quarter of the difference. If you are comparing minutes, it explains about one per cent at most, and the calendar explains nearly all of the rest.

Two things that do not explain the gap

Podcasts. A common explanation is that Wrapped keeps podcasts out of the total. Spotify says the opposite: they are in it. In this export they were also too small to matter, eight minutes across four years.

Private listening. Wrapped counts Private Mode time. So does the export, where it is marked incognito_mode. Quilore leaves private sessions out of its totals, which pushes our figure slightly below the export's rather than above it: 856 minutes in 2025 for this listener.

Checking your own gap

1. Open the files named Streaming_History_Audio for the year you want and add up ms_played for every entry. Divide by 60,000 for minutes.
2. Add it up again for streams that started before 16 November. The timestamp is when a stream ended, so subtract ms_played from it first.
3. Compare both with your Wrapped figure. The second should sit much closer to it.
4. Whatever difference remains is inside what Spotify has not published: the exact cutoff date, the timezone it uses for it, and how it treats time from plays under 30 seconds.