Weekly analysis · 2026-W41Wikipedia pageviews · MusicBrainzJul 2015 – Oct 2026

How much attention does a new album buy — and for how long?

On release day, views of an artist's Wikipedia article jump to a median 4.7× their usual level, against 1.3× on an ordinary Friday. Half of the jump is gone in 5 days. Over eight weeks a typical album adds about 28 days' worth of normal attention. 144 albums by 38 artists; press play to step through them.

Typical release-day peak

4.7×

vs 1.3× on an ordinary Friday · middle half 2.7–7.5×

Half of the rise gone in

5 days

median curve · 45% of the 8-week extra comes in week 1

Extra attention, 8 weeks

28 days

of usual views · ordinary Friday 3 · middle half 12–58

Daily data collected

4,114 days

51 artists · Jul 2015 – Oct 2026 · 1,538 ordinary Fridays as control

—

This album

—

Each dot is one album: extra attention in its first eight weeks. Click a dot, or use ← → and Space.

01 · The typical album

A spike on release day, half gone within a week

Take every album's curve and the median of them day by day: it peaks on day at the usual level, half of that rise is gone after , and it is back to by day 28 and by day 55. About of an album's eight-week extra attention arrives in its first seven days.

Why here and in the headline? The headline is the median of each album's own highest day; albums peak on different days, so the day-by-day median curve is flatter. Usual level = median daily views on days −35 to −8. The band on the stage is the middle half of all albums.

02 · Compared with an ordinary week

An album week looks nothing like a random Friday

I took the same artists on ordinary Fridays with no album within 90 days and measured them the same way. Their median peak is ; only of them reach the median album's . Albums: middle half .

Highest day in the first two weeks, × usual level

Blue: albums. Grey: ordinary Fridays. Log scale. Fridays are not empty: news, tours and awards also move attention; their 90th percentile is , and of albums are above it.

03 · Size and jump

Bigger artists get bigger jumps, not smaller ones

I expected well-known artists to have less room to grow. The data say the opposite: the rank correlation between an artist's usual daily views and the album's peak is . The total extra attention over eight weeks, in multiples of the usual level, is unrelated to size ().

Both axes log. Click a dot to see that album on the stage.

04 · Can it be predicted?

An artist's past albums do not predict the next one

Each album's eight-week extra attention was predicted using only albums released before it ( forecasts). The plain median of all earlier albums was off by a factor of . Using the artist's own previous album, or their median so far, was worse: and .

Error = mean absolute difference in log10, shown as a factor: 2× means the guess is typically half or double the actual value.

05 · Eleven years at a glance

Most big months are albums; some are not

Monthly views of each artist's article, compared with that artist's own median month. Ticks mark the albums analysed. Deaths, tours and award shows make the other bright cells.

Colour: views in the month ÷ the artist's median month; white = at or below the median, darkest = 20× or more. Grey = the article did not exist yet. Ticks: albums analysed; click one to open it.

What I found

An album lifts the artist's article to a median 4.7× its usual daily views within two weeks (middle half 2.7–7.5×). Ordinary Fridays: 1.3×.
The jump is short: half gone after 5 days on the median curve, still 1.2× at day 28.
Bigger artists get bigger relative jumps, not smaller ones (rank correlation 0.31 between usual views and peak).
An artist's past albums do not help predict the next one: the median of all earlier albums is typically off by 2.7×, the artist's own history by 3.4× (103 out-of-sample forecasts).

Read with care

Wikipedia views measure attention, not listening, sales or chart position. A news story moves them as much as a record.
English Wikipedia only, so attention from non-English readers (e.g. Bad Bunny, Diljit Dosanjh) is undercounted.
51 well-known artists, mostly US/UK; no Turkish artists, no newcomers. Results do not generalise to small artists.
Album dates are MusicBrainz's earliest release date; a few are a day or two off the official date (e.g. a surprise drop or an early regional release).
This is a before/after comparison, not proof that the album caused every view: promotion, singles and tours cluster around releases.
When singles or a tour lift the four weeks before release, the 'usual level' is too high and the album looks smaller: 10 albums end up with negative extra attention (e.g. 1989 (Taylor's Version), released during the Eras Tour).