Africa’s Banked Export Share on Spotify’s Charts Has Taken Only Two Values in 20 Readings. One French Chart Row Decides Which

Across 20 panel days between 11 October and 11 November 2026, the share of African-origin banked Spotify chart streams sitting outside Africa took one of two values and nothing in between. On 15 of those days it read between 1.580 and 1.715 per cent. On the other 5 it read between 0.373 and 0.539 per cent. The interval between the two bands is 1.041 points wide and contains 0 of the 20 readings. The lowest high-regime reading is 2.9314 times the highest low-regime reading. Nothing about the measurement changed between those days: the same 75 published daily Top 200 country charts were collected each time, and the count of African-origin artists charting anywhere on the panel moved only between 106 and 121. What flips the indicator is a single chart row in France.

Banked means cumulative: the streams a track has accumulated in that market over the whole period the chart has been tracked, not one day’s listening. This desk began publishing it last week because it is the only figure available that reaches back past a single reading. The series it produces is not a series. It is a switch.

Fifteen of the twenty readings sit between 1.580 and 1.715 per cent and five sit between 0.373 and 0.539. No reading sits between those two bands.

Sponsored

Show the numbers

Panel day	Banked export share as reported, %
2026-10-19	1.72
2026-10-25	1.68
2026-10-12	1.66
2026-10-27	1.66
2026-10-22	1.64
2026-11-03	1.63
2026-10-26	1.62
2026-10-23	1.61
2026-11-02	1.61
2026-10-24	1.61
2026-10-29	1.60
2026-10-30	1.60
2026-11-11	1.60
2026-11-08	1.59
2026-11-09	1.58
2026-10-18	0.539
2026-10-11	0.517
2026-10-17	0.449
2026-11-05	0.433
2026-10-16	0.373

Source: Computed from this desk's registered dataset african-banked-export-share-series, retrieved 13 Nov 2026 · download the data

The two regimes are one chart row#

On all 15 high-regime days the largest single holder of banked African export is Soolking, an Algerian-registered rapper resident in France, and his share runs from 69.6 to 73.6 per cent. France’s share of banked African export on those days averages 71.37 per cent. On all 5 low-regime days France’s share of banked export is 0 per cent, the largest holder is the Egyptian rapper TUL8TE, and his share runs from 33 to 37.4 per cent.

That the French figure is one row is not inferred from the shares matching. The stored snapshots were read row by row for all 20 days. The French chart carries exactly 1 Soolking row on each of the 15 high days and 0 on each of the 5 low days. It is the same record every time, one Spotify track ID across the window, “Tour du monde”, moving between positions 145 and 198 on between 52,716 and 72,815 streams a day. He holds 0 rows on any other non-African chart on any day.

The record does nothing on the days the indicator halves. Its banked total rises monotonically across the 15 days it is visible, from 67,579,069 to 69,036,029, a gain of 1,456,960 streams. Nothing leaves. The track slips below the 200th row of one chart, the panel stops seeing the 69,036,029 streams behind it, and the published measure reads 0.433 instead of 1.595 per cent.

The denominator is not what moves. Mean banked panel total is 5,890,266,204 streams in the high regime and 5,731,011,863 in the low one. Nor does the population move. Artists holding any lead row outside Africa run from 9 to 11 in the high regime and from 8 to 10 in the low one, and African artists charting anywhere on the panel run from 106 to 121 against 108 to 113. The thing that changes by a factor of 2.9314 is the numerator, and inside the numerator it is one row.

Subtract France’s recorded banked export from each day’s numerator and divide by the same unchanged denominator. All 20 readings then land between 0.373 and 0.539 per cent. The 15 high-regime days, recomputed, run from 0.425 to 0.514 per cent, which is inside the low regime’s observed band. The largest step between any two adjacent readings in the recomputed series is 0.052 points, against 1.041 points in the reported one. The reported range of 1.342 points is 8.0843 times the recomputed range of 0.166 points. France’s share is a measured column, not an assumption, but it is stored to one decimal place. At that rounding error the recomputed share moves by far less than a thousandth of a point; an error of a fifth in the column would lift the recomputed high days above the low regime’s ceiling. The counterfactual is as good as that column and no better.

Why one row can do this, and the daily measure cannot#

The reason is how much history each holding carries. On 11 November, Soolking’s row at position 193 on the French chart carried 69,036,029 banked streams against 55,833 streams that day: 1,236.5 banked streams for every stream it earned. The next deepest holding is Ramy Sabry’s at 238, then Mohamed Hamaki’s at 224.4 and TUL8TE’s at 197.2; the shallowest is Wegz’s at 6.9. A banked measure weights a row by how long it has been tracked, so one long-tracked French catalogue position outweighs nine current Egyptian and Angolan records combined.

Read on the day’s streams, the same artist is a large holding rather than the whole column. Soolking took 73.2 per cent of banked export on 11 November and 27.83 per cent of the 200,628 streams the column actually earned that day, a gap of 45.37 points. TUL8TE takes 23.47 per cent of the day against 9.8 of the bank, Ramy Sabry 13.81 against 7, Anselmo Ralph 9.23 against 3.2. Every holding except Soolking’s weighs more daily than banked.

That is why the one-day measure, computed on exactly the same days and the same charts, behaves like an indicator. Its 20 readings run from 0.866 to 1.503 per cent, a range of 0.637 points, and the largest gap between any two adjacent readings is 0.105 points, which is 16.5 per cent of its range. For the banked measure the equivalent gap is 77.6 per cent of its range. The regime gap in the banked measure is 9.9143 times the largest step in the one-day measure. Split the days by the banked regime and the one-day readings overlap across the split, 1.146 per cent on a high day against 1.170 per cent on a low one. The one-day measure does not know there are two regimes.

The counter-reading#

The obvious objection is that five readings is too few to call a regime. The day-drop test says otherwise. Dropping the single highest reading moves the mean of the banked measure by 1.5 per cent, the highest three by 4.62 per cent and the highest five by 8.35 per cent; the one-day measure moves by 0.93, 2.66 and 4.55 per cent. Neither mean is driven by outlier days, which is the point: the banked measure is not noisy, it is bistable, and a mean is the wrong summary of it.

The second objection is that Soolking should not be in the panel at all, and that removing him is not a robustness test but a correction. He is registered to Algeria by MusicBrainz, which is why this desk counts him, and he is long resident in France. This desk’s own classifier flags him, as it does Angola’s Anselmo Ralph, because neither holds any row on an African chart. Both flags were checked against the markets the artists actually chart in and both were left standing: an African-registered artist charting only abroad is the definition of export, not a classification error. The leave-one-out table makes the dependency explicit either way: on 11 November, removing Soolking cuts banked African export from 94,284,654 to 25,248,625, a fall of 73.22 per cent. Removing the top three leaves 9,370,209, down 90.06 per cent.

The third objection is the one that would have killed this outright: that the 5 low days are days the French chart failed to collect, making the whole pattern a collector defect rather than a property of the measure. It is not. The French chart returned a full 200 rows on every one of the 20 panel days. On the latest panel day, 6 of the 75 charts came back incomplete and all 6 carried 0 African rows, so no missing collection is hiding export anywhere else on the panel either. The wider limits sit the other way and are not fixable by method. Chart streams are tracks inside a market’s published top 200, not catalogue consumption, so an African record selling steadily just below the top 200 in Germany is invisible here. Banked totals are capped by how long the source has tracked each market, so they are not streams since release. Spotify publishes a daily chart for 4 African countries out of 54, and on 11 November 67 non-African charts carried no African lead credit at all. This piece measures the behaviour of an indicator. It does not measure African listening, and it should not be read as doing so.

What an independent label should do with this#

If you are building a case for foreign demand, do not quote a banked or all-time chart figure, and do not accept one from a distributor or a playlist pitch. A banked figure is a position’s tracked age multiplied by its market, and on this panel the difference between a reading of 1.595 per cent and a reading of 0.433 per cent is whether one catalogue record happened to sit inside a top 200 on the day the measurement was taken. Nothing about the other nine artists changed. If your own export case rests on a cumulative number, it will read as 2.9314 times stronger or weaker purely on the sampling date, and anyone checking it will find that out.

The one-day column is the one to quote, and on this panel it is small and it is honest: 200,628 streams across 10 artists on 11 November. It is the measure this desk has reported week after week. It does not jump on a sampling accident: it ranged from 0.866 to 1.503 per cent over 20 readings while the banked measure switched between two states. This desk will report the banked figure alongside the one-day figure from here on, and label which regime each banked reading is in, because a reading of 1.595 per cent and a reading of 0.433 per cent are not a change in African export. They are a change in what was charting in France.