How We Counted: The Method Behind Our Press-Archive Report
This note documents how every number in How Music Blogs Cover Live Sessions: Data from 328 Independent Articles was produced. The count date is 27 October 2026. One script reads the files described below and writes every table the report uses, so the numbers can be rerun on the same files and checked. We publish the method so that anyone citing the report can see what was counted, what was left out, and where the soft edges are.
The sources
The research log. Every article, page and mention we found about our sessions, the records we have worked on and the studio: 732 entries, each with its outlet, date, coverage type, what the article names, and a verification status.
The reviewed link list behind our press page. 424 links that went through two review rounds. 328 of them are marked as counted; that set is what the report analyses, and it is the same set behind the headline on the press page.
An automatic link check from October 2026: whether each link was live, whether an archive copy exists, and whether the page text names Tapetown.
The list of sessions and records with their release years. We store years, not dates.
The 32 fully verified pieces shown as featured on the press page, with publication dates and a note on whether the article links to our site.
The research log and the link list are joined on the article URL, normalised so that www, mobile and tracking variants of the same address match. Every one of the 424 links matches a research entry.
What counts
The 328 counted articles follow the rules we use for the press page:
Independent editorial coverage only, checked by hand or by the automatic page check.
Syndicated copies count once.
A submission network counts once.
Paid placements and press-release copies never count.
Coverage of our founder's own band never counts.
Organisations, festivals and artist sites are listed on the press page but not counted.
An outlet's country is where the outlet is based, not the language it writes in. Eighteen outlets, with 46 articles between them, have no known country. They count as outlets, never as countries.
An article is one counted link. An outlet is one publication. A session or record is one entry in our list; one entry can hold two songs or two releases by the same artist. Three entries are about the studio itself, with 11 articles between them, and appear as "studio" in the tables.
Why not all 732 entries
The research log is a record of everything we looked at, not a list of coverage. Each entry falls into exactly one of these buckets, the first rule that applies:
Bucket | Entries |
|---|---|
Counted: independent editorial coverage of a session, record or the studio | 328 |
Paid or submission-network placement | 62 |
Coverage of our founder's own band | 22 |
Press-release or publicity copy | 40 |
Possibly automated content | 6 |
Syndicated copy or duplicate page | 3 |
Not editorial press: artist, label, platform, database, retailer, venue, festival or institution page | 110 |
Dead link with no archive copy | 13 |
Reviewed and left out: not about the session or record | 55 |
Editorial or unclassified, but outside the press page's sessions and records | 93 |
The last bucket is left out on purpose. Those entries concern artists and releases without a page on our press section, or are still unclassified. They have not been through the press-page review, so they are neither counted nor analysed.
How coverage types were assigned
The research log holds more than 300 free-text descriptions of what each article is. Each counted article was given one of six categories, matched in this order on the description, the kind of evidence recorded, and the article title:
Roundup or list: flagged as a roundup in the link list, or described as a roundup, list, playlist, selection, ranking, year-end list, picks, weekly, mixtape, column or song of the day.
Premiere: described as a premiere or an exclusive.
Review: described as a review, in English, Danish, Italian or German.
Feature or interview: an interview, Q&A, profile, studio visit, feature, spotlight, editorial or credit piece.
News: news, an announcement, an advance, a preview, a brief, a report, a partnership or a mention.
Post built on the session video: the remaining session, embed, video and stream posts.
The line between a feature and a post built on the video follows the labels used during research and is soft. Read the two together where it matters: 163 articles, or 49 per cent.
Who the article names
This classification is built from what was recorded when each page was opened during research; the pages were not re-read for the report.
An engineer or producer is named when the research notes say the article credits, names, quotes or interviews one of the people who recorded, mixed or produced for us, or when the evidence recorded for the article is a named engineer or explicit personnel. Who those people are is listed in Tapetown Session Credits. Notes saying that a credit came from a release page, a database, the artist or our own credits post are not counted, and neither are filming credits or tags.
Tapetown is named, but no engineer when the automatic page check found "Tapetown" or "Bredvig" in the page text, or the title or notes say the page names Tapetown or the Tapetown Sessions.
Only the artist or release when neither applies.
The "engineer named" share is a lower bound. A page that names an engineer without the notes saying so is counted lower, and the automatic check can also find "Tapetown" in a tag or a sidebar.
Outlet country
The country is the outlet's own, as recorded in the link list. The report's "outside Denmark" share uses only the 282 articles with a known country: 223 of them, 79 per cent, came from outlets outside Denmark.
Dates and timing
A publication date to the day comes from the research log first, then from the featured pieces, then from a year-month-day pattern in the article's own address. 257 of the 328 articles have one. Year-only and month-only dates are not used for the week counts.
Because we store release years and not release dates, no article can be timed against a release date. The report uses two measures instead:
Year lag: the article's year against the session's or record's release year, for entries with a single release year. 242 articles qualify; 94 per cent were published in the release year.
Weeks after the first article: for each entry with a single release year and at least two dated articles, the clock starts at the first dated article published in the release year. Articles dated before the release year, such as pre-release singles or older band news, are left out. Every later dated article is measured in days and then whole weeks. That gives 38 sessions and records and 149 follow-up articles. The first article is usually the premiere or the release-day post, so this approximates time from release, but it is not the same thing.
Entries whose releases span two years are left out of both timing measures, because the clock cannot know which release an article follows.
How long the articles last
For the 328 counted articles, a link is archive copy only when the link on our press page is a web-archive copy, or the review recorded it as archived. It is live when the automatic check returned the page, and for the hand-reviewed links when the reviewer saw it. Three articles were blocked or down at the automatic check and were confirmed by hand.
Whether an article links to our site is recorded only for the 32 featured pieces, 25 of which are in the counted set. Four of those 25 link to tapetownstudio.com. The link was not checked for the other articles, so the report gives only that figure.
Rounding
Percentages are rounded down to whole numbers, so shares can add up to less than 100. The headline counts are rounded down in the report as they are on the press page, "more than 120 outlets"; the tables keep exact numbers. Medians are exact.
Known limits
One article is probably misfiled. A 2018 article counted under a record released in 2025 concerns, according to the research notes, an earlier recording made elsewhere, and names Tapetown only through a showcase. It adds one article to that record's count and changes nothing else.
Paid submissions. One counted article carries a note that it may have come through a paid submission category, and 11 of the counted outlets also accept paid submissions. Until our full submission records are matched against the archive, the headline outlet count has a worst case of "more than 110".
Unknown countries. Eighteen outlets with 46 articles, 14 per cent, have no known country.
The counted articles run from August 2015 to 2026. The research log holds entries back to 2009, from before the studio existed and about our founder's own band, and two entries from 2013 that are news items about band competitions. None of them is counted, so the report and the press page both count from 2015.
A few year values are missing in the research log; none of them belongs to a counted article.
Where to go from here
The report itself is How Music Blogs Cover Live Sessions: Data from 328 Independent Articles (2015–2026). Every counted article is listed, with a link, on Tapetown in the press. The script and the tables behind the report are kept with the archive, and we share them on request.
