Editorial Notes · Internal
Notes on the editorial choices in the horsegiirL Artist Insights build — what was surfaced, what was reframed, and why.
The Artist Insights deck uses raw data from Strategic Analytics. This doc walks through the editorial choices that turn data into narrative — which numbers got highlighted, which connections were drawn, and what the reasoning was. Think of it as a making-of for the insights process.
The point isn't that the data changes. It doesn't. Every number in the deck comes directly from the spreadsheet. What changes is which numbers get put in front of people, in what order, and with what framing.
The track_stream_trends tab has weekly ODA for every track. Sorting by recent average (Feb–Mar 2026), the top 3 tracks are: ONLY THE BEST (112K avg), My Little White Pony (81K avg), My Barn My Rules (55K avg).
Tracks #2 and #3 are from 2022. They're four years old and still pulling 81K and 55K weekly. That's not normal for a developing artist — usually the most recent release dominates and catalog fades. Here, catalog is actually sustaining the baseline between release spikes.
artist_stream_trends: US 8-week avg is 318K, WW 8-week avg is 911K. That's 34.9% US share.
35% US share is unusually high for a developing artist in this genre. Most electronic/hyperpop acts skew heavily international (European festivals, UK scene). Flagging this as an anomaly is more useful than just reporting the number, because it tells you where this artist sits relative to expectation.
partner_share: SoundCloud at 0.9% share, index 279.
Most people would skip SoundCloud at 0.9% share — it's a rounding error in the total. But the index is 279, meaning this artist's audience uses SoundCloud nearly 3x more than the benchmark. For the hyperpop/electronic lane, SoundCloud is still a discovery platform. That 0.9% share at 279 index is a cultural signal, not a business metric.
global_streaming_markets: Germany 145, France 148, Poland 167, Netherlands 155, Denmark 263. All over-indexing.
These aren't random European markets. They're the five countries with the strongest electronic music infrastructure — major festivals (Primavera, Sonar, Dekmantel, Unsound), established club scenes, and audiences that actively seek out this genre. Naming it the "European electronic corridor" connects the data to a geographic strategy rather than just listing numbers. It also tells International exactly which conversations to have.
audience_preferred_genres: Techno at 0.1% share, index 6,795.
0.1% share is negligible volume. 6,795 index is the single highest number in the entire dataset. That means horsegiirL's audience listens to Techno at 68 times the rate of the average listener. Combined with Dance/Electronic at 318 index and the Fans Also Like list (SOPHIE, A.G. Cook, Shygirl, Slayyyter, VTSS, Brutalismus 3000), this confirms the genre positioning conclusively.
Pop leads by raw share (29.5%), and a pure share-based read would call this a Pop artist. The index tells you the audience's identity; the share tells you the market's default. The identity is electronic/experimental. Pop is just what these listeners also consume.
social_followers: TikTok +180% YoY, 401K followers (surpassed Instagram at 310K).
engagement: TikTok 2.7% engagement, -6.6% vs benchmark. Instagram 10.0%, +4.8% vs benchmark.
If you only looked at follower growth, TikTok is the clear winner. If you only looked at engagement, Instagram is. The insight is in the disconnect: TikTok is driving discovery (people finding the artist) but Instagram is driving connection (people caring about the artist). They serve different functions.
This changes the content strategy recommendation. TikTok doesn't need less investment — it needs different content. Growth-at-all-costs content (trends, hooks, algorithmic optimization) gets followers but not engagement. Scene-specific, personality-driven content would trade some growth for depth.
The data exists (track_stream_trends, 2,595 rows) but horsegiirL has 21 tracks. A chart with 21 overlapping lines is unreadable. The bar chart (Week 1 debuts) and the "top tracks by recent average" commentary convey the same information more clearly.
Trimmed to top 9. The bottom 12 tracks all debuted below 4K. Including them would compress the bar chart and make the acceleration story harder to see. The editorial judgment: show the trajectory, not the inventory.
The data has 145 markets. The deck shows 11. Anything below 1% share and without a notable index signal was cut. The Denmark exception (1.0% share but 263 index) made the cut because it completes the European electronic corridor pattern.
Listed in partner_share at 0.0% share, index 1. Mentioned once in commentary to make a specific point: "this audience doesn't exist on lean-back radio." Then dropped. The absence is the insight.
Every action callout in the deck is tagged to a specific department. The routing logic:
The point of tagging departments is that someone scanning the deck can find the actions that are relevant to them without reading every slide. A person from Commercial Partnerships can look at the green ACTION bars, see their name, and know exactly what to do. Everyone else can skip it.
For reference, here's what happened between receiving the spreadsheet and delivering the deck:
1. Data intake: 14-tab Excel file with streaming trends, track debuts, partner share, demographics, market data, genre profiles, social followers, and engagement metrics.
2. Analysis pass: Calculated 8-week averages, YoY growth, identified top tracks by current streaming (not just all-time), found the biggest week-over-week jumps, compared share vs. index across all tables.
3. Insight extraction: Identified the 5–7 findings that would change someone's behavior — not just interesting facts. "Catalog is sustaining the floor" changes playlist strategy. "SoundCloud at 279 index" opens a new conversation. "TikTok growth vs. engagement disconnect" changes content strategy.
4. Department routing: Matched each insight to the team best positioned to act on it.
5. Output generation: Three formats from the same content — interactive HTML (for web sharing), PPTX (for meetings), PDF (for attachments).
Total data points in the spreadsheet: ~5,000+ rows across 14 tabs. Total data points that appear in the deck: ~80. The editorial value is in the 98% that gets cut.
Editorial Notes — horsegiirL Artist Insights — March 16, 2026