01 // THE PROBLEM

MUSIC DISTRIBUTION DOESN'T END AT DISTROKID.

Getting a song onto Spotify is easy. Getting someone to actually hear it is a different problem.

For years, I approached music promotion the conventional way: find stations, locate the right contact, understand what they program, write the pitch, send it, track the response, follow up, and repeat. Multiply that by thousands of stations and the constraint becomes obvious.

The problem wasn't access to contacts. The problem was operational scale. So I stopped treating radio outreach as a collection of emails and started treating it as a data and workflow problem.

02 // THE INPUT

4,095 CONTACTS ARE USELESS IF YOU CAN'T DECIDE WHO SHOULD RECEIVE WHAT.

RADIO COMMAND converts a public radio directory into a working targeting database. Each record can carry station, contact, genre, location, eligibility, match score, priority tier, contact status, and campaign history.

The database can be searched, filtered, scored, and segmented before a campaign ever sends.

• Station + contact
• Genre + location
• Eligibility + contact status
• Match score + priority
• Campaign history

03 // THE SCORING ENGINE

DON'T BLAST 4,000 PEOPLE. DECIDE WHO ACTUALLY FITS.

Every station is not equally relevant to every release. RADIO COMMAND assigns contacts a 0–100 match score and organizes targets into PRIME, A, B, C, or SKIP.

The system decides who belongs in the campaign before the campaign reaches the queue.

04 // CAMPAIGN CREATION

A RELEASE BECOMES A CAMPAIGN.

Tracks are reusable entities. Title, release, genre, mood, Spotify, Apple Music, YouTube, EPK, and description travel with the track, so the pitch infrastructure does not have to be rebuilt for every release.

• KILL CONTRACT — TERMS OF SUFFERING
• REWARD FUNCTION — SINGLE
• USE THIS SOUND — SINGLE

05 // PERSONALIZATION

ONE CAMPAIGN. THOUSANDS OF INDIVIDUAL MESSAGES.

Eight templates route outreach across Industrial / Metal, Electronic, Alternative, Experimental / Freeform, College Radio, General Broadcast, Personal DJ / Host, and Generic Inbox contexts.

Structured station data, track metadata, and the right template produce outreach personalized with station, contact, location, release, and listening links.

06 // THE TRANSMISSION LAYER

THE QUEUE HANDLES THE REPETITION.

Campaign outreach enters a controlled transmission workflow instead of firing as one mass send. Queue processing, daily limits, sender warm-up, status tracking, bounce handling, suppression, unsubscribe handling, and follow-up logic become reusable infrastructure for every future release.

07 // CAMPAIGN CONTROL

EVERY RELEASE GETS ITS OWN DISTRIBUTION OBJECT.

Campaigns maintain their own target population, progress, transmission controls, and performance state: targets, sent, replied, interested, bounced, and unsubscribed.

The first production campaign promotes KILL CONTRACT from TERMS OF SUFFERING.

08 // THE RESPONSE LOOP

SENDING THE EMAIL ISN'T THE END OF THE WORKFLOW.

Inbound responses return to a unified inbox and become structured outcomes instead of disappearing inside email. Positive opportunities move into a submissions workflow.

• INTERESTED
• SEND EPK
• SUBMISSION PORTAL
• ADDED TO ROTATION
• PASS
• WRONG CONTACT
• UNSUBSCRIBE
• SAVE SUBMISSION

09 // INTELLIGENCE

EVERY CAMPAIGN MAKES THE NEXT CAMPAIGN SMARTER.

The analytics layer asks which genres respond, which score buckets perform, which markets engage, which campaigns create responses, and which assumptions were wrong.

• Conversion funnel
• Replies over time
• Performance by genre
• Performance by score bucket
• Performance by region
• Campaign performance

11 // ARCHITECTURE

UNDER THE INTERFACE.

React and Tailwind power the mission-control interface. Base44 provides application infrastructure. TypeScript serverless functions handle scoring, campaign generation, queue processing, and workflow logic. Gmail OAuth supports controlled sending, Gemini assists response classification, and scheduled workers operate the queue.

• React + Tailwind
• Base44
• TypeScript serverless functions
• Entity-based application store
• Google Gmail API / OAuth
• Gemini
• Scheduled workers + queue processing

13 // MY ROLE

DESIGNED AND BUILT END TO END.

I architected and built RADIO COMMAND as part of the larger NULL PERSONA experiment: product architecture, data structure, contact ingestion, scoring, segmentation, tracks, campaigns, templates, personalization, queues, Gmail integration, replies, AI classification, submissions, suppression, analytics, and interface design.

14 // WHY I BUILT IT

I DIDN'T WANT TO BUILD AN AI SONG.

I wanted to see how far one person could take an idea when the creative work, software, automation, data, and distribution infrastructure were treated as parts of the same system.

NULL PERSONA became the identity, music, visual system, website, content engine, release infrastructure, radio database, outreach platform, and measurement layer. Music is one output. The larger project is the machine around it.

15 // THE BUSINESS LESSON

AUTOMATION ISN'T THE INTERESTING PART.

Sending thousands of emails automatically is not difficult. Building a system that decides who should receive something, what they should receive, when they should receive it, what happened afterward, and what the system should learn is a much more interesting problem.