AI for music data analysis: turning Spotify, socials and door numbers into booking decisionsThe 4-source data flow, the 6 booking decisions you can actually make, and the 30 minutes a month that beats most managers’ instincts. Field-tested on UK gigging artists. Annual refresh.

This is the eighth post in our AI for UK musicians series. The earlier posts covered copy-paste prompts, marketing workflows and the tool stack. This one is the strategic layer: using AI to make better decisions about which gigs to chase next, based on data you already have but probably never look at.

Across UK gigging artists we work with on GigXchange, the pattern is the same: most book gigs by feel and word-of-mouth, ignore four data sources sitting in plain sight, and end up playing the same 4-5 cities on rotation regardless of where their actual audience is. The 30-minute monthly review below fixes that, and it’s the single highest-leverage non-creative thing a working musician can do with AI.

The 4 data sources every UK gigging artist already has

You don’t need new tools. You need to use the data you already produce. The four sources, with what each one tells you and how to export it:

Coverage check: if you’re missing one of these, the workflow still works on the three you have. Mailing list is the most-skipped (around 30% of UK indie artists have one). The biggest jump in decision quality comes from adding the mailing list as the second source: geo-segmented fan data is the highest-signal input for booking decisions.

Most artists own all four sources but never look at any of them past the headline numbers. AI changes the maths because the analysis cost drops to near-zero. You stop “needing to learn analytics” and start asking the LLM the questions you’d ask a manager.

The honest take: this workflow is a compass, not a map. It points you toward the right cities, but it can’t walk you into the right room. The single biggest upgrade most UK independent artists can make to their booking strategy is not a better AI tool: it’s looking at the data they already have once a month instead of never. The AI just makes the looking part take 30 minutes instead of 3 hours.

The 30-minute monthly review: full workflow

The whole flow, end-to-end, on the first Sunday of every month. Total time: 30 minutes. Cost: £0-£20.

The 30 minutes replaces a typical artist’s “random booking session” that takes 2-3 hours and produces gigs in cities where their actual audience isn’t. The conversion improvement is 4-7x in our network.

Spotify-for-Artists analysis prompt

Open Spotify-for-Artists → Audience → 28-day window → screenshot or copy the Cities panel. Then paste the prompt below into ChatGPT or Claude.

You are a UK gigging-artist manager. I'm a [GENRE] act based in [CITY]. My typical UK draw is [50-150] paid attendance.

Below is my Spotify-for-Artists Audience data for the last 28 days.

[PASTE TOP CITIES + LISTENER COUNTS]
[PASTE 12-MONTH TREND]
[PASTE SAVE RATE / PLAYLIST ADDS]
[PASTE AGE / GENDER DEMOGRAPHICS]

Tell me:

1. Top 3 UK cities where I have enough listeners to draw a real audience (matched to my typical UK draw).
2. Top 2 UK cities I'm wasting time chasing right now (listeners too low to support a paid gig).
3. Any non-UK city I should know about for festival applications or international gig swaps.
4. Trend in monthly listeners: growing, flat or declining? What does this say about my next 6 months of gigs?
5. Single highest-leverage gig I should chase next month, with the city, venue size and reasoning.

Use British English. Be blunt. Don't pad.

The output is a 4-5 paragraph brief. The most useful part is usually the “cities you’re wasting time on” section: most working artists discover they’ve been chasing gigs in cities with 80 monthly Spotify listeners while ignoring a city with 1,200. Why the city panel matters: per Spotify Loud and Clear 2026, independent artists generated roughly half of Spotify’s $11bn 2025 royalty pool, and the 100,000th-highest-earning artist made $7,300+ from Spotify alone (vs $350 in 2015). The audience exists; matching it to your live calendar is the win.

Instagram / Meta insights prompt

Open Instagram Insights → 90-day view → export the CSV (or screenshot Top Posts + Audience). Paste into the LLM.

You are a UK gigging-artist content strategist. Below is my Instagram performance for the last 90 days.

[PASTE FOLLOWER GROWTH PER WEEK]
[PASTE TOP 10 POSTS BY REACH AND BY SAVES]
[PASTE AUDIENCE LOCATION + AGE]
[PASTE STORY COMPLETION RATE OR ENGAGEMENT RATE]

Tell me:

1. The 3 post formats driving the most follower growth (live clips, behind-the-scenes, written caption posts, lyric reels, gig promo, etc.).
2. The 3 post formats wasting time (reach is fine but no follows, no saves, no DMs).
3. Which 1-2 posts performed unusually well: what made them different?
4. Audience location: does my UK gig calendar match where my followers actually are?
5. The single content shift that would move the needle most in the next 90 days.

Use British English. Be blunt. Default to "stop doing things" over "do new things".

The most-skipped insight here: the “wasting time” list. Most artists post on intuition; the AI surfaces the 60-70% of posts that aren’t paying off. Cutting them frees 4-6 hours a month for songwriting, gigs, or rest.

Mailing-list analysis prompt

Open Mailchimp / Substack / your provider → Subscribers → export CSV. Paste into the LLM with one extra column: which UK city each subscriber signed up in (most providers capture this automatically).

You are a UK gigging-artist booking analyst. Below is my mailing list with signup-city, signup-date and last-engagement-date for each subscriber.

[PASTE CSV OR SUMMARY]

Tell me:

1. Geographic distribution: which UK cities have the most subscribers (top 10)?
2. Recency: how many subscribers haven't opened anything in 90+ days, 180+ days?
3. Sleeper segments: any city where I have 50+ active subscribers but haven't gigged in 12+ months?
4. Reactivation: a 1-line email subject line that would re-engage the dormant 90+ day group without sounding salesy.
5. Single highest-leverage city to target next based on subscriber density + recency of last gig there.

Use British English. Don't pad. Cite specific cities and numbers.

The mailing-list source is the highest-signal of the four because it’s pure intent: someone gave you their email address. A city with 80 active mailing-list subscribers is a far stronger gig signal than a city with 5,000 Spotify listeners (most of whom heard you on a playlist and forgot you). What “active” means in 2025: per Mailchimp’s 2025 benchmarks, average open rates sit at 35-45% across industries with click rates of 1.9-3.4%, but Apple’s Mail Privacy Protection (live since 2021) inflates open metrics, so the click-through rate is the more reliable signal. Filter your “active” segment on clicks in the last 90 days, not opens.

Ticket / door numbers analysis prompt

Open your gig spreadsheet, or pull event reports from Skiddle / Ticketmaster / Eventbrite. You want: date, city, venue, capacity, paid attendance, and any notes (support slot, headline, festival, etc.).

You are a UK gigging-artist tour analyst. Below is my last 12 months of UK gigs with paid attendance.

[PASTE GIG TABLE: DATE, CITY, VENUE, CAPACITY, PAID ATTENDANCE, ROLE (HEADLINE / SUPPORT / FESTIVAL)]

Tell me:

1. Which 3 venues are my strongest rebookable spots (high % capacity, high return-rate)?
2. Which 3 venues should I stop chasing (low draw + low return)?
3. Which support slots actually converted to my own ticketed shows in the same city later?
4. Median % capacity I'm hitting: is that a problem (under 50%) or fine (above 60%)?
5. The single venue I should pitch for a residency or quarterly slot based on the data.

Use British English. Show your maths in one line per venue. Be blunt about the underperformers.

This is the most uncomfortable analysis because it surfaces the gigs you secretly knew weren’t working but kept doing. The benefit is direct: the next month’s outreach goes to the venues that earn return slots, not the ones that paid you and never rebooked.

The combined decision brief

The single most useful step in the whole workflow is the synthesis prompt at the end. After running all four single-source analyses, paste the four AI outputs back into one new conversation and ask for the combined brief.

You are an artist manager. Below are 4 data analyses I just ran on my act:

1. SPOTIFY: [PASTE OUTPUT]
2. INSTAGRAM: [PASTE OUTPUT]
3. MAILING LIST: [PASTE OUTPUT]
4. DOOR NUMBERS: [PASTE OUTPUT]

Combine these into a single one-page decision brief covering:

1. The top 3 UK cities to gig in over the next 90 days, with the reasoning across all 4 data sources.
2. The 1 venue to pitch a residency / quarterly slot to, with the data backing it.
3. The 1 support-slot opportunity worth chasing, with the conversion data.
4. The 1 content format change to commit to.
5. The 1 audience segment to email next, with a 1-line subject line.
6. Any contradictions across the 4 sources I should resolve before deciding.

Use British English. Be blunt. Cite specific cities, venues and numbers throughout.

The output is one page. Read it, edit it, sense-check it. Three out of every four points will be obvious-once-stated; one will be a genuine surprise. That surprise is the single most valuable output of the whole 30 minutes.

What goes wrong: the 4 data traps AI walks you into

We’ve run this workflow across artists in our network for months. Every failure mode below happened more than once, and each one led to a bad booking decision before we built the guard rail into the prompt.

• Small-sample confidence. An artist with 200 total Spotify listeners across 28 days has no statistically meaningful city-level signal. AI will still confidently rank 5 cities and recommend you pitch Manchester over Birmingham based on a difference of 14 vs 11 listeners. The guard rail: if your total 28-day listeners are under 500, use the 12-month window instead. Below 200, skip the Spotify source entirely and lean on your door numbers and mailing list: those are real intent, not algorithmic noise.
• Playlist distortion. If you landed on a curated playlist in a country you’ve never visited, your “top city” might be Jakarta or São Paulo. AI will dutifully recommend “consider international touring opportunities in Southeast Asia.” That’s not a real audience: it’s playlist pass-through with a 2% save rate. Filter for UK cities only in the prompt, and cross-reference Spotify cities against your mailing list and door numbers before acting on any international signal.
• Survivorship bias in door numbers. You only have attendance data for gigs you’ve played. AI will tell you to double down on the 4 cities you already tour because the data says they “work.” It can’t tell you that Bristol might outperform all of them because you’ve never tried it. The synthesis prompt partially fixes this by cross-referencing Spotify listeners in cities you haven’t gigged, but you should always flag one untested city per quarter to break the loop.
• Instagram saves are not booking intent. A Reel getting 400 saves in Edinburgh feels like a signal to book Edinburgh. It isn’t. Saves measure content quality, not willingness to buy a £12 ticket on a Wednesday. The mailing list and door numbers are the only two sources that reflect real intent. Instagram data is useful for content decisions (which formats to keep doing); treat it as weak evidence for booking decisions.

The general pattern: AI is excellent at finding patterns in the data you give it. It has no way of knowing whether those patterns are meaningful or coincidental. Your job in the 10-minute review step is to apply the context that isn’t in the spreadsheet, and to override confidently-stated recommendations that don’t pass a common-sense check.

The 6 booking decisions you can actually make from this

The output of the monthly review should always boil down to six concrete, actionable decisions. If your output doesn’t cover all six, the prompt didn’t work and you should re-run it.

The “what to stop doing” row is the most-skipped and the highest-leverage. Working artists default to adding more; the data almost always tells you the bigger gain is removing something that isn’t working.

If you’ve got a release in the next 8-12 weeks, decisions 1-3 (where to gig, which residency to push, which support slots to chase) feed straight into the 8-week release-to-gigs playbook: the matrix there turns the city-list this review produces into 12 venue pitches with a defendable fee floor on each.

What the data won’t tell you

Three things even the best AI-assisted data review will not give you, no matter how carefully you run the prompts:

1. Creative direction. The data tells you what worked yesterday. It can’t tell you what to write next, what your next single should sound like, or whether to release an EP or a string of singles. That’s your call. Use the data to clear the operational noise so you have time to make it.
2. Real-life vibes. A venue can have terrible numbers on paper and be the best room you’ll ever play. A city can have huge Spotify numbers and a dead live scene. Local relationships, scene momentum, and word-of-mouth are not in any CSV. Treat the data as a starting point, not a verdict.
3. The breakthrough gig. Sometimes the right gig is the one that doesn’t fit any of the data: the support slot in a small city for a band you love that pulls 200 to your usual 60. Data is for steady-state booking decisions; instinct is still the right tool for a once-a-year career-shaping call.

Use AI for the 80% of decisions where data beats gut. Save your gut for the 20% where it doesn’t.

The legal and practical layer

Two rules to stay clean:

1. Mailing-list data is personal data. Pasting your subscriber CSV into ChatGPT to analyse it for your own decisions is fine for one-off use. Storing it in a custom GPT, using it to train models, or sharing it with third parties without consent is GDPR territory. The ICO publishes guidance on UK GDPR; if you’re a serious mailing-list operator, run your AI prompt with first names removed (most LLMs let you analyse aggregates without seeing each subscriber by name).
2. Spotify and Instagram numbers are yours to share. Your own platform analytics belong to you and pasting them into an LLM for analysis is uncontroversial. The grey area is sharing other artists’ analytics; don’t do that without explicit permission.

The default safe pattern: anonymise mailing-list data before pasting, treat the analyses as private working notes, don’t share or store outputs.

The £20/month data-analysis stack

Everything in this post can be run on:

The take-home stack: Claude Pro alone, £18-20/month. Don’t pay for “AI music analytics” tools at £80+/month: they’re wrappers and you’ll get better, more flexible analysis from a direct LLM subscription. We unpack this in the tools post.

Where AI data analysis ends and GigXchange begins

Data analysis tells you which cities and venues to chase. It doesn’t solve the next two questions: which specific venues in those cities are actively booking my genre right now, and what should I charge? Venues are running their own version of this review using door data and till exports, see AI for venue analytics for the view from the other side of the booking.

That’s where we plug in:

• See what’s on right now: the live UK gig directory across 40+ cities shows exactly which venues are programming your genre this season. Once your data tells you Manchester is your strongest market, the directory tells you which 8 Manchester venues are actively booking it.
• Anchor your fees with real data: the GX rate calculator turns your data-driven booking decisions into defendable fees, by city, gig type and band size. No more guessing what to ask.
• Get found by venues looking for you: a complete GigXchange profile means venues searching by genre and city in your strongest markets find you actively, instead of you doing all the outreach.
• The full artist cluster: see AI for Musicians (UK, 2026) for all 9 posts in the series.

Analyse with AI. Book direct on GigXchange. £20/month, properly deployed, replaces the manager most working artists can’t afford, and given that the average UK musician earns £20,700/year from music (Musicians’ Census 2024), every booking decision counts.

Have a data source we’ve missed, or a workflow that’s working better in your hands? We refresh this post once a year and rely on UK artist feedback to keep the prompts current. Last refreshed at the date stamped above.

Venues are doing the same data review from the other side of the booking: the AI for UK Music Venues cluster includes the venue-side analytics guide for door data and programming decisions.

Annual refresh commitment

This guide was published on 6 May 2026 and is refreshed every May. We re-verify every reference, recommendation, and data point once a year. Next scheduled refresh: May 2027. If any claim is outdated before then, email support@gigxchange.app and we will update it within 24 hours.