This desk’s Spotify chart roster carried 2,052 artists with no resolved country of origin on 25 September 2026, 35.8% of its 5,728 charting artists. That is barely moved from the 35.4% this desk measured on 15 September, when it first reported that the roster’s MusicBrainz resolver had failed on more than a third of the panel. In the ten days between those two counts, 342 new artists joined the roster, and 148 of them, 43.3%, arrived unresolved too, a worse rate than the roster as a whole.
The jump between 5 and 15 September was the resolver failure this desk reported on 15 September. The line has been flat, not falling, for the ten days since.
Show the numbers
| Point | Unresolved share, % |
|---|---|
| 25 September (this collection) | 35.82 |
| 15 September, before that day's corrections | 35.50 |
| 15 September, after that day's corrections | 35.40 |
| 5 September (baseline) | 1.20 |
Source: This desk's own artist roster (afw_data_roster), counted directly from the stored option using the same test ($r['country'] === '') the roster resolver's --unresolved flag uses, retrieved 25 Sep 2026 · download the data
| # | Point | Charting artists on roster | Unresolved-origin artists | Unresolved share, % | Resolved to an African country | African share, % |
|---|---|---|---|---|---|---|
| 1 | 5 September (baseline) | 4,908 | 61 | 1.20 | 119 | 2.40 |
| 2 | 15 September, before that day's corrections | 5,386 | 1,912 | 35.50 | 129 | 2.40 |
| 3 | 15 September, after that day's corrections | 5,386 | 1,904 | 35.40 | 136 | 2.50 |
| 4 | 25 September (this collection) | 5,728 | 2,052 | 35.82 | 145 | 2.53 |
Source: This desk's own artist roster (afw_data_roster), counted directly from the stored option using the same test ($r['country'] === '') the roster resolver's --unresolved flag uses, retrieved 25 Sep 2026 · download the data
New artists are entering unresolved at nearly the same rate as the backlog#
The 15 September piece described a one-time resolver failure: a batch of new artist IDs that MusicBrainz could not place on first attempt, written into the roster as permanently unresolved rather than left pending for a retry. Ten days of further panel refreshes were a test of whether that failure was a single bad batch or a standing condition of the resolver. It is the latter. Of the 342 artist IDs added to the roster since 15 September, 185, 54.1%, resolved to a non-African country, 9, 2.6%, resolved to an African country, and 148, 43.3%, could not be placed at all.
Newly added artists are failing to resolve at almost the same rate, 43.3%, as the ten-day-old backlog, 35.8%. The gap is not closing on its own.
Show the numbers
| Outcome for the new artist ID | Artists |
|---|---|
| Resolved to a non-African country | 185 |
| Left unresolved: no MusicBrainz match | 148 |
| Resolved to an African country | 9 |
Source: Computed from this desk's registered roster-resolution-stagnation-05-to-25-september-2026 dataset: each row is the 25 September total minus the 15 September (after-corrections) total, retrieved 25 Sep 2026 · download the data
This desk’s roster only ever writes a country for an artist ID new to it; the resolver does not revisit or silently clear an ID it has already assigned a country to, so a rising unresolved count cannot come from artists losing a classification they once had. It can only come from new artist IDs failing resolution on their first attempt, which is what the arithmetic above shows directly: the three outcome counts for the 342 new IDs sum to exactly 342.
Half of it sits in one market’s chart#
| # | Name | Total chart streams | Unresolved-origin streams | Unresolved-origin lead artists | Unresolved share, % |
|---|---|---|---|---|---|
| 1 | Nigeria | 15.04m | 1.88m | 30 | 12.50 |
| 2 | Egypt | 6.14m | 810,891 | 20 | 13.20 |
| 3 | Morocco | 2.08m | 954,067 | 41 | 45.87 |
| 4 | South Africa | 5.37m | 3.64m | 94 | 67.83 |
Source: This desk's afw_data_roster matched against the 25 September 2026 Spotify chart snapshot for each of the four African markets Spotify publishes a daily Top 200 for, retrieved 25 Sep 2026 · download the data
South Africa’s own Spotify chart is the reason the desk-wide figure looks as bad as it does. 67.8% of South Africa’s chart streams on 25 September belonged to artists this roster cannot place at all, up from 65.2% on 22 September, three days earlier. Morocco is the next worst at 45.9%. Nigeria and Egypt, this desk’s two largest African markets by chart volume, sit at 12.5% and 13.2%. South Africa alone held 94 of the four markets’ 185 combined unresolved lead artists, and
The counter-reading#
A four-market total driven by one market’s chart is exactly the kind of finding that needs a robustness check before it is trusted as a four-market story rather than a South Africa story wearing a four-market label.
| # | Scenario | Rows | Total | Largest row share, % | Change in total, % |
|---|---|---|---|---|---|
| 1 | All rows | 4 | 7.29m | 49.99 | 0 |
| 2 | Excluding the top 1 (ZA) | 3 | 3.65m | 51.59 | -49.99 |
| 3 | Excluding the top 3 (ZA, NG, MA) | 1 | 810,891 | 100 | -88.88 |
Source: Recomputed from panel-origin-resolution-by-market-25-september-2026, retrieved 25 Sep 2026 · download the data
South Africa is 50.0% of the combined unresolved total by itself. Drop it and the remaining three markets, Nigeria, Egypt and Morocco, still carry 3,645,690 unresolved streams between them, a fall of exactly half. Drop the three largest markets by unresolved volume, South Africa, Nigeria and Morocco, and Egypt alone accounts for the remaining 810,891, an 88.9% fall from the full panel. So the finding is really two findings stacked: the roster has a general, ten-day-old resolution problem that is not improving, and South Africa’s chart concentrates that problem more than any other market this desk tracks, for reasons this dataset cannot explain, since an unresolved artist ID carries no information about who the artist actually is.
What this means for reading this desk’s other South Africa numbers#
Every African-share figure this desk has published for South Africa’s own chart, including the 15.8% African share in this desk’s current market-share panel, is computed by treating unresolved artists as not African by default, because an unclassified artist cannot be counted as a confirmed African-origin stream. That default is a limitation of the roster, not a finding about those artists, and on South Africa’s chart specifically the gap between what is confirmed and what is simply unclassified is not a rounding error: 67.8% of that chart’s streams, nearly seven in ten, carry no confirmed origin at all. A South African label reading this desk’s African-share numbers for its home chart should treat the published 15.8% as a floor this desk can prove, not a ceiling on what is actually there. For Nigeria and Egypt, where under 14% of chart streams are unresolved, the gap between the published African share and the true one is small enough that the headline numbers are close to reliable as stated. For Morocco and especially South Africa, they are not, and this desk will not have a fix to report until the roster’s unresolved share actually starts falling, which it has not done in the ten days since it was first measured.
The limits#
This compares four snapshots taken on different collection dates rather than a continuous daily series, so a single volatile day inside that window could exaggerate or understate the true ten-day trend; the next several collections will show whether 35.8% is a stable plateau or a number still climbing. The 43.3% new-entry-unresolved figure rests on the roster’s own no-regression behaviour, that an artist ID already assigned a country is never revisited or cleared, which is how the resolver is built but which this piece has not independently audited row by row against every one of the 342 IDs. Unresolved is a statement about what this desk’s own MusicBrainz-based resolver can confirm, not a statement about an artist’s actual nationality: the seven rows flagged by this desk’s own anomaly check on today’s collection were resolved correctly by name search or manual override despite carrying no identity link, and the reverse, a wrong classification that looks confident, is a separate failure mode this piece does not measure.
