House Track: Business Plan (revision 2, after red team)

Your brand's own music, licensed for every store, every post and every ad, under one agreement.

Prepared November 2026, revised after a three-person red team (finance, sales and lean operations). Every forward-looking number in this plan is a projection. The assumptions sit next to each number and in model.json. House Track has no customers, revenue, partners, pilots or letters of intent today, and nothing here says otherwise. Market and comparable figures carry a source link. Where the data is thin, the plan says so.

What changed in revision 2. The first version built its volume top-down (7,000 locations by year 5) and valued the business at 4.0x forward ARR ($28M). The red team showed that the funding did not reach the plan's own gate, that the multiple was applied to the wrong base, that the low tier was priced 3-6x above a direct AI competitor, that the lead fitness pitch overlaps incumbents, and that the operating cost of devices, compliance and franchise sales was missing. This revision rebuilds volume from deals, restructures pricing around a headquarters license, removes proprietary hardware from year 1, adds the missing cost lines and makes the engine's provenance a day-0 gate. The result is a smaller, slower and more honest base case: about $2.8M of revenue in year 5, monthly EBITDA break-even in month 32, about $0.89M of peak capital, and a base-case exit of roughly $8-9M. Section 17 lists every critique and the response.

1. Executive summary

The pitch. Franchise systems and multi-unit groups have a music problem that sits at headquarters, not in the store. Hundreds of franchisee and location social accounts post video with commercial music the brand has no right to use, and courts have now found brands liable for exactly that (Bang Energy, Marriott's franchised hotels, DSW; sources in §2). In the stores themselves, brands pay a background-music service, often pay performing-rights blankets as well, and still sound the same as their competitors. House Track sells one owned, AI-generated catalog per brand, tuned to the brand, organized by zone and daypart, streamed to every location, and licensed for in-store play, on-hold, social video and ads, with documented provenance and a capped indemnity.

The unit of sale is the brand, not the speaker. The core product is a Headquarters (HQ) Brand License ($1,500-6,000 a month; modeled at $2,200) that covers the brand catalog, the social and ad license for the franchisor and every franchisee account, and the indemnity. In-store streaming is sold on top, per location. Direct multi-unit groups (med spas, dental and veterinary groups, independent fitness groups, dealerships) buy a per-location Brand plan that bundles the same rights for their own accounts.

Who buys first (targets, not customers): 1. The general counsel or CMO of a 50-300-unit franchise brand, buying the HQ license after a lawsuit headline in their category. 2. The COO of a 10-40-location med spa, dental or veterinary group, buying a premium brand feel plus on-hold at $99 a location. 3. Independent multi-unit studio groups and emerging fitness brands under 100 units that do not already have a branded-music provider. Large fitness franchisors are not a year-1 target (§6).

What House Track does not sell. It does not lead with PRO savings. Savings appear only as a hand-off on the calculator's output page ("your counsel may then decide..."), and the calculator does not ship until current rate cards from all four PROs are in hand (§3).

Headline numbers (base case, all projections, unit-level build in model.json):

• Capital: peak cumulative need about $887k (month 37). Committed in three gated tranches: $450k at launch, $300k at the month-9 gate, $150k at the month-21 gate, plus a 25% contingency (about $220k) held back and not drawn in the base case. Total exposure if everything is drawn: about $1.1M. A lean-launch variant cuts peak need to about $828k (§11).
• Break-even: monthly EBITDA positive from month 32. Cumulative cash does not turn positive inside five years in the base case (it reaches -$293k at month 60). This is a build-to-sell or build-to-harvest plan, not a quick payback.
• Exit (base): 3.0-3.5x trailing recurring revenue of $2.64M = $7.9-9.3M enterprise value. The EBITDA cross-check on a buyer's normalized "harvest-mode" EBITDA (about $1.15M, 38% margin, growth spend removed and a market-rate CEO paid) at 6-8x gives $6.9-9.2M. Upside case at 4.0x: about $14.6M. Signature Sound is excluded from exit revenue.
• Probability-weighted view (judgmental): 50% stopped at the month-9 gate (loss about $0.42M), 30% downside (net about $3.2M), 15% base (net about $6.5M), 5% upside (net about $10.0M). Expected net value about $2.2M after tax and fees, on $0.45M at risk at the first gate and up to $1.1M if every tranche and the contingency are drawn. §12 shows the full returns table, after tax and after a charge for the founder's time.

2. Problem and customer

Who exactly

1. Franchise systems of 50-300 units (lead HQ segment). Fast-casual, fitness, salon, med-aesthetics, home services and retail franchisors that run a brand fund and set brand standards. The buyer is the general counsel or CMO; the pain is system-wide social-video and ad exposure across independently run accounts.
2. Med spa, dental (DSO) and veterinary groups of 10-40 locations (lead direct segment). AmSpa counted 10,488 med spas (2023 data), with about 13,000 projected by the end of 2026 (AmSpa). About 17% of US dentists were DSO-affiliated in 2025 (Becker's Dental). These groups play ambient music all day, run on-hold audio, are often acquisitive and buy centrally.
3. Independent multi-unit fitness groups and emerging fitness brands under 100 units. Low-BPM concepts (Pilates, yoga, stretch, recovery) where generative instrumentals suit the floor, lobby and on-hold. Class music stays with whatever the instructors use today unless a blind test (§15) shows otherwise.
4. Dealership groups and salons. Central buyers who care about brand consistency.
5. Destination resorts (Signature Sound). One discounted case study in year 1, then a fixed-scope product (§3). No savings claim, because resorts keep PRO blankets for live acts.
6. Cannabis dispensaries (year 2, not year 1). Moved back because payment processors and banks often decline licensees, and because the segment is small (107 licensed dispensaries in Nevada, the first state targeted, CCB).

How they buy

• Franchise HQ: legal and brand-standards review, a pilot, often a franchise advisory council, then an approved-vendor listing. Approved-vendor listings typically take 6-18 months (red-team estimate from franchise sales practice; we found no published benchmark, so treat it as a range, not a datum). A listing does not mean adoption: each franchisee is a separate buyer unless the franchisor mandates the vendor. The model assumes optional adoption (§11).
• Direct groups: COO or operations lead with procurement; 2-6 month cycle; a 30-day pilot on 1-2 locations.
• Referral partners: insurance brokers and franchise-compliance consultants see the exposure first. Channel partners: AV integrators (installation and tier-2 device support) and, potentially, class-music providers (§6).
• Trigger events: a lawsuit headline in the category, a takedown, a rebrand, a new-location build-out, a franchise compliance audit, or a PRO letter.

Evidence the pain is real

• Social-video liability is a litigation pattern. UMG and Sony won liability rulings against Bang Energy over TikTok posts; the court held that TikTok's license did not cover commercial use. Sony sued Marriott over 931 alleged infringements, including posts by franchised hotels (settled). Warner and Sony sued DSW in 2025, and rightsholders have pursued Crumbl and Chili's (Reed Smith; Pryor Cashman; MBW on DSW).
• Unlicensed in-store use is widespread. Soundtrack's November 2026 survey found 79% of US businesses that play music use consumer streaming services licensed only for personal use (MBW). That figure comes from a vendor with an interest in the answer.
• What has not been tested: House Track has run no customer interviews or pilots. Willingness to pay for an HQ license, franchisee adoption, churn and member and instructor acceptance are all unvalidated. The first 90 days are built to test them, cheapest test first (§15).

3. Solution and offer

Product. A per-brand catalog generated by a licensed engine with documented commercial and public-performance rights (§10, a day-0 gate), screened and sample-reviewed by human curators, delivered through a web player, casting, or the customer's existing system (Sonos and similar). No proprietary hardware in year 1. From year 2, one certified off-the-shelf device is available through integrators, who keep the hardware margin and handle installation and first-line fixes. Programmed by zone (lobby, floor, treatment room, restrooms, on-hold) and daypart.

How catalogs stay distinct without per-brand generation every month. Each brand gets a "brand DNA" preset and an exclusive core set (about 150 tracks) refreshed quarterly. Monthly rotation draws from shared, pre-screened genre pools filtered through the brand's parameters. This cuts generation and human-QA load (§9) while keeping the exclusive set that makes a brand sound like itself.

Changes from revision 1: Essentials ($49, shared library) is replaced by Core at $29 with no social rights. Brand per location becomes $99 for direct groups. The social and ad license moves from an 8% usage add-on into the HQ license, which is where the value sits. Hardware revenue is gone. Hospitality zones now carry the same net-down as locations.

Compliance, done as a product, not a promise. - Whole-premises inventory for any customer who asks about PRO licenses: every place music plays, which sources House Track covers, which still need a license. The decision to drop any blanket stays with the customer and their counsel. House Track supplies documentation, not legal advice. - Quarterly re-attestation by automated survey. If a location does not re-attest, its whole-premises status lapses on paper, which limits House Track's exposure when an instructor plugs in a phone or a TV goes up in the lobby. - PRO-response packet: a template, a one-click provenance export and a counsel letter, so a PRO letter becomes a 30-minute task and not an open case. Outside counsel is retained on a fixed fee per incident, and letters per 100 blanket-dropped locations is tracked as a KPI.

The calculator, demoted. The savings calculator becomes a hand-off on the proposal's last page, not the pitch. It shows $0 in PRO savings whenever any live act, DJ, TV, radio or other repertoire remains, whenever the customer already uses a PRO-inclusive service, and for every casino. In fitness it is shown only for the Core tier. It does not ship until current rate cards from ASCAP, BMI, SESAC and GMR are on file. The only hard rate behind revision 1 was BMI's 2019 schedule, and BMI's current page says only that its fitness minimum is "less than a dollar a day" (BMI). Savings claims must be substantiated under FTC Act Section 5.

Indemnity, capped and insured. Capped at 12 months of fees paid by that customer. Offered only after (a) a bound IP-defense insurance quote under the launch threshold (§15) and (b) melody-level screening is in the QA loop (§9).

4. Why now

1. Brands are being sued over social music, and franchise systems multiply the exposure. See §2. This is the trigger the HQ license is built around.
2. Labels are moving from suing AI companies to licensing them. UMG and WMG settled with Udio in late 2025, and WMG partnered with Suno (MBW; Bloomberg Law). That makes licensed-data engines more available and raises the provenance bar, which favors an operator that can document its engine.
3. Integrators are deploying AI-generated in-store music. M-Cube moved luxury in-store radio to AI-generated music in 2025 (sixteen:nine).
4. Incumbents are consolidating. Mood Media bought PlayNetwork (2021) and Vibenomics (2023), Soundtrack bought Ambie and Tunify (Feb 2026), and Stingray bought DMI (Oct 2025) (BusinessWire; mainsights; Stingray). Roll-ups create buyers.
5. A funded competitor has the same thesis. Tonada raised a $3M pre-seed led by Antler (announced June 2026) to generate original, royalty-free tracks from a brand's identity, and reports paying customers in Scandinavia, DACH and Singapore (Pulse 2.0; tech.eu). That validates the category and means House Track's edge has to be US franchise distribution and the HQ license, not the idea.

5. Market size (bottom-up)

Top-down context: estimates of the global background-music market range from about $1.9B (IMARC) to about $3.0B (Polaris) for 2025. Definitions differ; treat as order of magnitude.

Bottom-up, US. Location counts come from the cited sources. Multi-location shares are assumptions.

• Sourced SAM: about 76,000 locations. At a blended ~$75 net per location per month that is roughly $68M ARR of in-store streaming, before HQ licenses. The count of US franchise systems with 50-300 units, which sizes the HQ license, has not been sourced; that is the most important market number still missing.
• Base-case year-5 target: 2,642 locations (about 3.5% of the sourced SAM) and 18 HQ licenses. Revision 1's 7,000 locations (9% of SAM) is retired.

6. Competition and positioning (honest)

What revision 1 got wrong. It said competitors "must build or buy an engine." Rightsify has its own model, Tonada has one, and licensed engines are becoming available. The engine is not the moat.

The real moat (to be earned, not claimed): 1. US franchise distribution: approved-vendor listings, franchise-specific self-serve signup and franchisee billing. These take 6-18 months each to win, which is a barrier for the next entrant too. 2. The HQ social and ad license: one signature that covers every franchisee account, with an indemnity. Incumbents license in-venue playback; creator libraries license content but do not program stores; nobody we found sells both to franchise headquarters. 3. The provenance ledger: every track carries its engine, model terms, generation parameters, screening results and license ID. Counsel and insurers ask for exactly this.

Fitness, re-scoped. Large fitness franchisors already have branded or curated class music, and instructors keep it. House Track does not try to displace it. The fitness offer is the floor, lobby, on-hold and social license ("you keep class; we supply everything else"), sold to independent multi-unit groups and brands under 100 units. FitRadio, Pure Energy Music and Feed.fm are treated as possible channel partners or acquirers, not targets.

Where House Track loses: buyers who want recognizable hits, buyers who want only the cheapest legal background music (Rightsify, CloudCover), and enterprise RFPs that require Mood-scale footprints.

7. Business model and pricing

• Recurring: HQ licenses, franchise-location streaming, direct-group Brand locations, Core, and hospitality zones.
• Project revenue: Signature Sound, excluded from exit revenue. Year 1 has one discounted case study, carried at its $15k production cost with no revenue counted. Then 2, 3, 4 and 4 packages a year at $30-35k, with producer and PM cost at 45%.
• Engine royalty: 8% of recurring revenue, under an assignable license (§14).
• Franchisor rebate: 6% of franchise-location revenue to the franchisor's brand fund (approved-vendor programs commonly charge 5-10%, per the red team; we found no published benchmark).
• Integrators: a $125 spiff per activated location they source, plus 10% trailing margin on integrator-sourced revenue only. They keep the hardware margin. No margin on deals House Track sources itself. Program capped at 3-5 integrators in 2 metros through year 2.
• Contracts: annual terms with location-minimum true-ups for groups; annual prepay for Core; 30-day pilot on the first 1-2 locations of a group; HQ licenses on annual terms.

Unit economics (base, projection; month-60 cost structure)

Revision 1's table had two arithmetic errors: the royalty at 8% of $67.80 is $5.42, not $5.90, and payments at 3% are $2.03, not $2.20. The revised offers are shown below.

Fixed cost of goods sold not shown per unit: similarity and melody screening plus IP insurance, $25k in year 1 rising to $110k in year 5 (a placeholder until a carrier quotes, §15).

Churn and CAC are untested assumptions. LTV is capped at a 60-month horizon so that low churn does not inflate it.

Churn assumptions, by segment

Sensitivities at 3% churn on every segment, and at 1.0-1.25% on annual contracts, are in §11.

8. Go-to-market

First 10 targets, by title (targets, not signed)

1. General counsel, a 50-300-unit franchise brand in a category that has seen a social-music lawsuit (fast-casual, retail, hospitality). Offer: HQ license.
2. CMO, a 50-150-unit emerging franchise brand (salon, med-aesthetics or boutique fitness under 100 units). Offer: HQ license plus franchise-location streaming.
3. COO, a 10-40-location med spa group. Offer: Brand at $99 with on-hold.
4. VP Operations, a regional DSO (20-100 offices). Offer: on-hold plus lobby.
5. COO, a regional veterinary group.
6. Owner, an independent multi-unit Pilates, yoga or stretch group (5-20 studios), not under a franchisor with a music deal. Offer: floor, lobby, on-hold and social; class music stays as is.
7. Marketing director, a dealership group (5-15 rooftops).
8. Partner, an insurance brokerage serving franchise systems. Referral partner.
9. Principal, a franchise-compliance consulting firm. Referral partner.
10. VP Brand or guest experience, a destination resort. The single Signature Sound case study.

Channels

• Direct to HQ (general counsel and CMO). Lead line: "Your franchisees are posting TikToks with major-label songs on them." One signature covers the system, which lowers CAC.
• Referral partners: insurance brokers and franchise-compliance consultants, who see the exposure first.
• AV integrators: 3-5 in 2 metros through year 2; self-serve certification, deal registration and automated payouts through a partner portal.
• Class-music providers as partners: "we supply your floor, lobby and on-hold; you keep class."
• Gym-software marketplaces are upside, not a year-1 channel. Mindbody requires partner-team review and approval and an active business subscription before go-live (Mindbody developer FAQ).
• Trade shows: franchise expos, the AmSpa Medical Spa Show, and dental and veterinary group conferences.
• The House Track Studio prototype (§10): the prospect hears their brand's sound in the first 30 seconds.

Sales capacity

• Year 1: the founder plus the Success and Ops lead. 10 direct groups signed in months 4-12 (one a month, two in month 12) and 1 HQ license signed in month 11.
• Year 2 onward: a commissioned enterprise closer (base $90k plus heavy variable) from month 13, which revision 1 had in year 4. Direct groups: 20, 28, 34 and 38 a year in years 2-5. HQ licenses: 3, 5, 6 and 7 a year.
• Signature Sound: founder-sold case study in year 1; afterwards a fixed SKU delivered by a contracted producer-PM, or white-labeled to a sonic-branding agency for a 30-40% referral share. It continues only if margin after PM time stays above 50%.

9. Operations with minimal headcount

Automated

• Catalog pipeline: brand brief, DNA preset, quarterly exclusive set, monthly rotation from shared pools, loudness normalization to venue targets.
• Screening: fingerprint and similarity screening against major catalogs plus melody-level screening (transcription-based or a vendor). Fingerprinting catches near-copies of recordings but does not reliably catch a lifted melody, which is the real risk for generated music (arXiv 2601.21260). Vendors describe their AI checks as pre-screening that still needs human review (ACRCloud).
• Player telemetry: heartbeat, auto-restart and auto-resync, before the first pilot.
• Franchisee onboarding: franchise-specific signup link, auto-provisioned zones, card on file.
• Support: AI chat for tier 1. Tier 2 for devices goes to integrators.
• Billing, dunning, rebates and partner payouts.

Human, contracted, paid per unit of work

• Curators, paid per track for sample-based QA of each catalog's quarterly exclusive set and of the shared pools. Customer approval is capped at one round. Budgeted at $100 per active catalog per month (about $160k in year 5 across about 130 catalogs).
• Producer-PM for Signature Sound (inside its 45% cost).
• Outside counsel on a fixed fee per PRO incident, plus fractional counsel and a fractional CFO.
• Contract ML-ops and app development in year 1, until the Ops/ML engineer is hired.

Org chart (W-2), hiring triggered by volume, not by calendar

Year-5 team: 6 W-2 plus about 2-3 FTE of contractors (curation, producer-PM, fractional counsel, contract dev). Year-5 recurring ARR per W-2 is about $0.50M (run-rate $3.0M ÷ 6), or about $0.35M per FTE including contractors, which is inside the $250-400k range the red team cited for lean SaaS.

Three runbooks are written before the first pilot: player outage, takedown and replace within 24 hours, and PRO-letter response.

Founder's weekly hours (estimate)

• Months 0-12: about 27 h (engine diligence, first HQ conversations, first groups).
• Year 2: about 17 h. Year 3: 15 h (franchise relationships stay with the founder). Years 4-5: about 10 h.
• The founder takes no salary until monthly EBITDA has been positive for three months in a row (month 35 in the base case), then $150k a year. §12 charges the unpaid time at $150 an hour.

10. Technology, the engine gate and the prototype

The engine is a gate, not an asset (yet)

Revision 1 said "a production-grade pipeline already exists." That is not true for this use today. The parent music-AI company's current production pipeline depends on a third-party consumer generation service with throughput caps. Its consumer product has been in maintenance mode since October because of one of those caps, and its licensed and self-hosted engines are still in development and staging. A B2B public-performance catalog resold with an indemnity cannot be built on consumer-account access to a third-party service.

Day-0 engine gate (before any capex): 1. An engine with documented commercial and public-performance rights, either a licensed-data model under written commercial terms or a self-hosted model trained on licensed or owned data. 2. A one-page provenance memo (model, terms, training-data posture, upstream warranties), written before the first sales call. 3. A measured cost per finished, QA-passed track and a measured capacity of at least 50 catalog-refresh equivalents a month (about 2,500 QA-passed tracks). The $1.50 per location per month generation cost in the model is unverified until this test runs. 4. If the engine is self-hosted, a contract ML-ops engineer is in the year-1 budget ($60k), and the W-2 Ops/ML engineer follows when ARR reaches $1.0M.

Step-in cost if the parent's engine is replaced by a third party (sensitivity, §11): generation cost of $4-6 per location per month plus a 5-8% license fee in place of the royalty. Year-5 EBITDA falls by up to about $0.14M; the plan survives it.

Stack

• Agent orchestration for briefs, screening, rotation and refresh.
• Catalog store on object storage behind a CDN with signed URLs; web player with offline cache; casting and Sonos-type paths.
• From year 2, one certified off-the-shelf device under device management at $4 per device per month (Esper pricing lists $2, $4 and $6 per device per month by plan, with a 25-device minimum).
• A console for brands, locations, zones, dayparts, social-asset downloads with a license certificate per asset, and franchisee billing. SSO and a console API for franchisor tech teams in year 2 ($50k capex).
• The provenance ledger: every track carries its engine and model version, model terms reference, generation parameters, screening reports and license ID.
• Year-1 build: $100k (web player, console, ledger, billing, franchise signup), up from $55k.

The prototype: House Track Studio

A browser-based venue sound designer: brand inputs, a generative Web Audio engine, a 3-note sonic logo derived from the brand name, a daypart planner, a zone map and a one-click brand sound kit. It is a sales demo, not the production engine, and it uses no third-party generation service.

11. Financial plan

5-year projection, base case (PROJECTION)

Volume is built from deals: HQ licenses signed × units per system × franchisee adoption over time, plus direct groups signed × locations per group, plus Core self-serve, all net of segment churn (§7).

Why the margins are lower than revision 1 (79% gross, 52% EBITDA). Revision 1 left out variable support, device management, human curation, melody screening, the franchisor rebate and integrator spiffs, and it sized costs against 7,000 locations that the unit build no longer supports. Year-5 EBITDA of 17% reflects a business that is still spending to grow at about 40% a year. With growth spending removed (marketing, commissions, the closer and half the channel role), the month-60 run-rate supports about $1.15M of EBITDA (38%), which matches the red team's 35-40% steady-state estimate.

Key assumptions (all in model.json): - HQ: about 120 units per system; franchisee adoption (optional vendor) of 30% of units in the first 12 months, 10% more in the next 12, then 0.3% of units a month; franchise-location churn 1.75%/month; HQ churn 1%/month; first HQ signature in month 11. - Direct groups: 14 locations on average, deployed over 3 months (40/30/30); same-group expansion of 0.5% a month (med spa and DSO groups are acquisitive; AmSpa's count implies about 7% a year industry growth); churn 1.25%/month. - Core: launches month 6; 5, 10, 20, 25 and 30 gross adds a month in years 1-5; churn 3%/month; annual prepay. - COGS: hosting, streaming and generation $4.50/location (2x per zone); variable support and fleet $2.50/location; device management $4/device at 30% attach from month 13; curation $100/catalog/month; royalty 8% of recurring; payments 3%; integrator trailing margin; screening plus IP insurance $25k, $45k, $70k, $90k and $110k (paid annually in advance); Signature producer-PM 45%; year-1 case study $15k. - Opex: W-2 as in §9 (loaded cost $85-160k, 3% annual raises); founder $0 until three straight EBITDA-positive months, then $150k; contractors $84k-130k (reduced by $60k when the Ops/ML engineer joins and by $24k when the finance generalist joins); legal $40-60k a year plus $1,500 per PRO incident (3 incidents per 100 blanket-dropped locations a year, with 15% of locations assumed blanket-dropped); marketing $40k-200k; commissions at 10% of first-year contract value on gross adds, plus integrator spiffs; software, including sales-tax compliance, $36k-75k; insurance (E&O, cyber, GL) and admin $15k-40k; a $2k blind test in month 1. - Working capital: monthly-billed revenue collected 30 days in arrears; Core prepaid annually (deferred revenue); IP insurance prepaid each January. No hardware inventory (integrators hold devices). The month-37 cash trough is the year-4 insurance prepayment.

Break-even: monthly EBITDA positive from month 32. Cumulative cash troughs at -$886,730 in month 37 and is -$293k at month 60.

Scenarios (PROJECTIONS)

The downside is shown for completeness. The gates in §15 are designed to stop it before most of its $1.33M is spent. The red team expected a 12-month HQ delay to push the trough to $450-500k against revision 1's model. In this rebuild, direct groups carry the early revenue, so the delay adds about $107k to peak need (from $887k to $994k), not $150-200k.

Sensitivities (base case, one change at a time)

The two largest single risks to capital are IP insurance pricing (+$250k of peak need at 2x) and churn. That is why both carry hard gates.

Capital required and use of funds

Tranche 1, year-1 uses of funds (base case):

Tranche 2 and 3 can come from the founder or from a small SAFE with angels or a strategic partner. A strategic partner could be an AV distributor, a franchise-software platform or a class-music provider. This is more than the $150-300k that revision 1 said was self-fundable. The plan is still within a founder-plus-small-round range, but the founder should decide with the full $1.1M exposure in view.

12. Returns

Comparable multiples (cited)

• SaaS M&A medians are about 4.0-4.2x EV/TTM revenue (SEG, 2Q25-3Q25; 2Q26 summary at 4.0x) (SEG; SEG 3Q25 PDF). These are trailing multiples. Small deals ($0-5M ARR) sit around 3.3x (Aventis Advisors).
• Audio and media strategic: Stingray-TuneIn at 5.8x adjusted EBITDA, $175M EV (GlobeNewswire).
• Category: Mood Media bought Muzak for a reported $345M in 2011 (Wikipedia). Terms were not disclosed for Mood-PlayNetwork, Soundtrack-Ambie/Tunify or Stingray-DMI, so the data is thin.
• Natural buyers: Mood Media, Soundtrack, Stingray, Spectrio (mergr), FitRadio or other class-music providers, franchise-software platforms, and hospitality-tech platforms. A buyer acquires the engine license, not the engine. That is why §14 requires the license to be assignable, with a buyout option and source escrow.

Valuation method

• Base: 3.0-3.5x trailing-twelve-month recurring revenue (HQ licenses, franchise and group locations, Core, and hospitality zones at net prices). No usage add-ons, no Signature Sound, no hardware.
• Cross-check: 6-8x a buyer's normalized EBITDA. At month 60 the year-5 reported EBITDA ($0.37M normalized) supports only $2.2-3.0M, because the business is still spending to grow. Harvest-mode run-rate EBITDA of about $1.15M supports $6.9-9.2M. A buyer paying 3x revenue is paying for that margin plus about 40% growth. If growth stalls, use the EBITDA number.
• 4.0x and above is upside only.

Scenarios and returns (PROJECTIONS)

IRR convention: annual free cash flows, with the year-1 flow at t=0 and the exit at the end of year 5. Positive flows and the exit are reduced by 20% for tax and deal fees. Tax treatment depends on entity form and the founder's situation; get tax advice, including on whether qualified small business stock rules could apply to a C-corp. Founder time cost: about 1,400, 900, 750, 500 and 500 hours in years 1-5 at $150 an hour, less salary actually paid. That is about $430k of unpaid time over years 1-3.

Revision 1's bear-case IRR (57%) is removed. The gates stop that path, so quoting a return for it was misleading.

Probability-weighted view (judgmental, not data)

On $0.45M committed at the first gate, and up to $1.1M of total exposure, an expected net of about $2.2M is attractive but not extraordinary. About two-thirds of that expected value comes from the 20% of outcomes at base or better. The decision rests on the cheap early tests in §15 (engine, blind test, insurance quote, the first HQ conversations), not on the spreadsheet.

Cash-flow alternative. If the founder does not sell, the base case reaches about $0.43M of free cash flow in year 5 while still growing, and about $1.1M a year in harvest mode from about year 6. Cumulative cash turns positive in year 6.

13. Risks and mitigations

14. Legal and regulatory checklist (factual, not legal advice)

• ☐ Entity: a separate C-corp (a SAFE is likely for tranche 2 or 3); cap table; founder IP assignment.
• ☐ Engine license from the parent, written and approved by the parent's board:
• royalty-bearing (8% of recurring revenue) with arm's-length support;
• assignable on a change of control without consent, and not terminable for convenience;
• a perpetual, paid-up buyout option at a set multiple of trailing royalties (for example 3x), exercisable by House Track or an acquirer;
• source and model escrow, with release on insolvency, breach or a change of control at the parent;
• warranties on commercial and public-performance rights and on training-data provenance;
• a right to regenerate the catalog on another engine without penalty.
• ☐ Upstream model terms: confirm commercial and public-performance use for every model in the pipeline; document training-data provenance. No consumer-account generation in the House Track pipeline.
• ☐ Contributor agreements: work-for-hire and IP assignment; contributors agree not to register House Track works with any PRO.
• ☐ No PRO registration of catalog works; maintain the provenance ledger.
• ☐ Opinion letter on whole-premises replacement; compliance kit and PRO-response packet reviewed by counsel.
• ☐ US Copyright Office guidance: register human-authored elements with AI disclosures.
• ☐ 17 U.S.C. §110(5)(B) small-establishment exemption reflected in the calculator.
• ☐ Customer terms: license grant (in-venue, on-hold, social, ads), sublicensing to franchisees under the HQ license, indemnity capped at 12 months of fees, takedown-and-replace SLA, location-minimum true-ups.
• ☐ Franchisor agreements: rebate (modeled at 6%), brand-standards obligations, data and SSO terms.
• ☐ Insurance: IP-infringement (defense) cover bound before any indemnity is offered; tech E&O, cyber, GL. Standard CGL usually excludes IP (Insureon).
• ☐ Advertising claims: any savings statement substantiated with current rate cards (FTC Act Section 5).
• ☐ Sales tax: SaaS taxability varies by state; compliance software budgeted.
• ☐ Integrator agreements: spiff, trailing margin on sourced deals only, territory, support split.
• ☐ Cannabis (year 2): processor and banking rules for serving licensees.
• ☐ Privacy: console user data, player telemetry; privacy policy and DPA.

15. Launch plan, gates and kill criteria

Day-0 gates (before tranche 1 is spent on build)

1. Engine gate (§10): documented rights, provenance memo, measured cost per QA-passed track and capacity of at least 50 catalog-refresh equivalents a month. Fail: do not launch.
2. Insurance gate (by day 30): a bound or firm IP-defense quote under $60k a year for year-1 limits. Fail: launch without any indemnity, or stop.

Days 0-30

• Form the entity; sign the engine license (§14 terms).
• Blind test (first 2 weeks, about $2k): 10 instructors and 50 members across 3 studios, plus staff and managers at 2 med spas, hear generated tracks against their current music, scored for fit and class flow. Nothing is spent on devices or class-room features until this passes.
• Write the three runbooks (outage, takedown, PRO letter).
• Run 30 discovery calls: 10 franchise HQ (general counsel, CMO), 12 direct groups, 4 referral partners, 2 integrators, 2 class-music providers.
• Request current rate cards from ASCAP, BMI, SESAC and GMR.

Days 31-90

• Two direct-group pilots on 30-day terms, on the web player.
• First HQ license in legal review.
• First integrator and first referral partner signed.
• Hire the Success and Ops lead.
• Report weekly: locations live, net ARPU, uptime, tickets per location per month, pilot-to-paid conversion, and venue-manager and member feedback.

Kill or pivot criteria

• Day 14 (blind test): fewer than 60% of members rate the generated tracks as good as or better than current music for floor and lobby use → drop fitness entirely and focus on med spa, dental, vet and HQ.
• Day 90: fewer than 2 paid direct groups, and no HQ license in legal review → stop new spend; keep only the Signature Sound case study.
• Month 9 (tranche-2 gate), all of:
• at least 60 paying locations (base model: 89);
• at least 2 HQ agreements signed or in legal review;
• pilot-to-paid conversion of 50% or more;
• at least 1 signed approved-vendor listing or signed HQ license;
• logo churn under 1.5% a month on annual contracts;
• under 0.15 support tickets per location per month.
• Fail: do not release tranche 2. Wind down, or sell the catalog, contracts and provenance ledger as an asset deal.
• Month 21 (tranche-3 gate): recurring run-rate of at least $450k ARR (base model: about $536k), at least 3 HQ licenses active, and churn still under 1.5% a month. Fail: cut to a cash-flow-positive team (no closer, no new hires) and run for harvest.
• Any time: a credible infringement finding traced to the engine that regeneration cannot cure, or an insurer non-renewal without a replacement → stop offering the indemnity and re-scope.

16. Domain names (proposals; availability not checked)

1. housetrack.com (ideal; likely already registered)
2. gethousetrack.com
3. housetrackhq.com
4. playhousetrack.com
5. housetrackaudio.com
6. housetracks.com

17. Red team and response

Three reviewers examined revision 1: a financial red team, a sales red team and a lean-operations red team. Each finding below is marked Accepted, Partly accepted or Pushed back, with what changed.

Financial red team

Sales red team

Lean-operations red team

Net effect of the red team

What the red team did not change: the problem is real and documented; contribution margins per unit are healthy (69-84%); Signature Sound stays out of exit value; and gates are the right way to run the business. With them, House Track is a real but smaller bet: a likely $8-9M outcome if it works, sized against up to about $1.1M of exposure, with half the probability mass ending at a gate for about $0.42M.

Files: model.json (assumptions, monthly 60-month build for every scenario, annual outputs, sensitivities, probability-weighted view, savings-calculator rules marked not-to-ship, red-team log and sources), house_track_model.py and build_model.py (reproduce every number in this plan), and brand.json.