Avalon Build Package
Business Case
Market thesis, customer profiles, revenue model, wedge strategy, competitive landscape, and risk register. Written for Charles and Cliff. Not build input; the feature matrix and tech stack are the build input. October 2026 update adds GloVerse — a celebrity-backed on-demand beauty marketplace that entered beta this month and is now the most relevant supply-chain risk in the landscape.
$8–9B
Global market estimate 2025–2026
$18M+
Raised by direct competitors (excl. GloVerse)
5
Direct / near-direct competitors as of Oct 2026
9x
More spend on hair care vs general pop
$130K–$400K
Estimated range to launched product (Rev F scope)
18–20M
Adult Black women: consumer planning range (Pew 2024)
The Thesis
A Black woman deciding on her next style is making a decision with unusually high switching costs. A set of knotless braids is four to eight hours in a chair and commonly $150 to $400. Locs are a multi-year commitment. A silk press is a heat-history decision. A big chop is irreversible. And the inspiration photo she is working from is almost always on someone else's face, someone else's texture, and someone else's density.
The gap is not that she lacks inspiration. Instagram and Pinterest have solved inspiration. The gap is that inspiration does not transfer, and the cost of finding out is measured in hours and hundreds of dollars.
Avalon closes the gap between the reference photo and her own face. The visualizer is the wedge. The business is what happens in the ninety seconds after she sees herself in the style and thinks "I want that": who she books, what she buys, and who gets paid for the referral.
The gap is not that she lacks inspiration. Instagram and Pinterest have solved inspiration. The gap is that inspiration does not transfer, and the cost of finding out is measured in hours and hundreds of dollars.
Avalon closes the gap between the reference photo and her own face. The visualizer is the wedge. The business is what happens in the ninety seconds after she sees herself in the style and thinks "I want that": who she books, what she buys, and who gets paid for the referral.
Ecosystem Structure
Sourced baseline compiled by Cliff Smith's team, September 24 2026. These are counts from different methodologies and sources — do not sum across rows. Each figure measures a different unit in a different year from a different data collection method.
| Segment | Count | Source | Date | What it measures |
|---|---|---|---|---|
| U.S. online hair-product sellers (Shopify) | 2,618 | EachSpy directory | Sep 23 2026 | Storefronts in one niche on one platform — not all DTC hair commerce |
| U.S. wig shops and boutiques | 5,842 | SmartScrapers | Apr 1 2026 | Location records; overlaps with online sellers and service providers |
| Extension-focused service providers | 12,684 | SmartScrapers | Apr 1 2026 | Business/location records including salons and studios; not a count of individual installers |
| Employer beauty salons (NAICS 812112) | 84,176 | U.S. Census 2023 | 2023 | Establishments by location — broad universe, not unique companies |
| Nonemployer beauty businesses | 838,264 | U.S. Census 2023 | 2023 | Provisional secondary-source figure; primary-table extraction outstanding |
| Hairdressers, hairstylists, cosmetologists | 595,000 jobs | BLS 2025 | 2025 | Broad occupation count; ~286,000 self-employed (48%); not a count of unique people |
| Adult Black women (consumer planning range) | 18–20M | Pew 2024 + Census | 2024 | Derived: 49.2M Black residents, 73% adult, 50–55% assumed female — demographic range, not an adoption estimate |
Five counting rules
- Business accounts are not professional users
- Storefronts are not companies
- Specialties are not additive
- Consumer relationships are not unique consumers
- A stylist is not the same as a salon license
Data hypothesis
Cliff's working hypothesis: consumer data may be the most valuable long-term outcome for ER and, at sufficient adoption, for Avalon -- more valuable than the visualization product itself. This is a hypothesis to design for, not an established business or validated metric. Illustrative arithmetic: 1,000 business accounts x 200 annual active consumers = 200,000 consumer relationships; 100 accounts x 10,000 = 1,000,000. Neither number is a forecast. Four open research priorities remain: reconcile business base to ownership-level register, estimate service specialties separately, replace the female-share assumption with direct demographic data, and quantify the specialization rate within the licensed universe.
Avalon and Ebene Regne are distinct businesses
ER is the charter customer and validation environment for Avalon -- not the same legal entity. ER's objectives for the platform: social-network growth, simpler purchasing, additional market channel, and consumer insights. White-label and grey-label deployment to other businesses is a commercialization option under consideration for Avalon, separate from the ER deployment. Cliff is meeting with legal to establish a second entity specifically for Avalon (flagged September 24 2026).
Customer Profiles
Three distinct ICPs. Only the supply side pays at launch. Charge for the qualified lead, not the transaction.
Primary consumer
She's choosing her next style
- Black woman, 18–44, US metro or suburban
- Changes protective styles every 4–8 weeks
- $100–$400 per install, plus product
- Discovers styles on TikTok, Instagram, Pinterest
- Owns a phone with a decent front camera
"I don't want to sit for six hours and then hate it."
Supply side (who pays)
The braider or stylist
- Independent braider or natural hair stylist
- Booth renter or home-based
- Books through Instagram DMs or link in bio
- Pain: no-shows, unrealistic reference photos, deposits
- Marketing budget is small, but she will pay for bookings
"She sent me a picture of 24-inch waist-length knotless on a girl with a completely different density and got upset when it didn't look like that."
Brand ICP (later)
The hair care brand
- Indie and mid-market textured hair care brands with DTC motion
- Wig and extension retailers
- Multinationals with textured lines
- Want proof of authentic reach
- Phase 4: don't chase this first
"Which styles are rising by region, texture type, and season?"
Revenue Model
Four lines, in the order they should be switched on. Don't try to turn on three at once.
01: Switch on first
Stylist subscription
$29–$79/month
Claimed profile, appearing in the directory, receiving consultation handoffs with the render attached, deposit collection. Value is legible: leads. Fastest path to revenue.
02
Consumer subscription
$6–$10/month
Free tier with a daily render cap. Paid tier for unlimited renders, lookbook history, and comparison tools. Low ARPU, useful for cost recovery and demand signaling, not the business.
03
Product attach and affiliate
Affiliate margin or placement fee
Every style has a maintenance product set. Attach at the moment of intent. Constrained by not compromising recommendation integrity (a real tension worth naming in advance).
04
Demand data (aggregate only)
High margin, near-zero delivery cost
Which styles are rising by region, texture type, and season. Brands and distributors have no good source for this today because the demand lives in DMs. Strictly aggregate per C9.
Not recommended: Taking a percentage of the appointment. Requires payment rails, invites disintermediation (she books once through Avalon and thereafter through the stylist's Instagram), and puts Avalon in direct collision with StyleSeat, Booksy, Vagaro, and Fresha. Charge for the lead, not the transaction.
Wedge Sequencing
Four options. Pick one. The competitive research below directly informs which of these is the right lane. Read together.
Option A
Consumer-first
Build the audience, monetize later. Highest ceiling, longest runway. Requires distribution Avalon does not have yet. HairHunt has a 14-month head start and a content funnel.
Option B
Stylist-first
Fastest revenue from the supply side. In-salon surface (A-13) is cheaper to build than consumer PWA and monetizes immediately. Competitive research strengthens this option; the stylist-facing lane is uncrowded.
Option C
Brand-first
Fastest cash. Cleanest match to Swift's existing retainer motion. Risk: turns Avalon into an agency deliverable and the consumer product never gets built.
Option D
Wig commerce first
Render units, attach a buy button, take retail margin. No stylist dependency. Risk: becomes a store rather than a platform. Parfait is already here and executing poorly, a customer-service and supply-chain play, not a technology one.
Recommendation: Consumer-free for demand signal, then pick one paid motion. If cash is the constraint: wig path monetizes soonest. If the long-term asset is the constraint: stylist-paid builds the supply relationships that are the actual moat. Don't chase the brand deal first, even though it is the easiest money, because it is the one path that quietly kills the actual company.
The market is funded and moving. Five companies have now raised or launched against variations of this thesis. October 2026 update: GloVerse — a celebrity-backed on-demand beauty marketplace from Grammy-nominated GloRilla — entered beta this month. It is not a visualization competitor, but it is the most funded distribution machine targeting the same stylist supply relationships Avalon needs at Phase 3. The premise in early drafts was wrong: textured hair tech is a contested category, not an open field. That does not kill Avalon. It sharpens the strategy and changes what counts as a real position versus what is marketing.
The Open Lane
Nobody found in this research combines all four. Individually, each exists. Together, they don't.
1
Style visualization tuned for textured hair
HairHunt has the closest version but user reviews confirm face alteration at scale. Table stakes, not moat.
2
Feasibility verdict on whether she can actually get the style
No funded competitor has this. The feature that tells her if her hair can reach the style, not just what it would look like. Strongest differentiator in the package.
3
Connection to a stylist who does that style
Swivel Beauty has this independently. Partnership is more realistic than rebuilding ST-01 and ST-02 from scratch.
4
Honest, measured fidelity reporting
Everyone claims face preservation. HairHunt proves claiming is cheap. Measuring and reporting by hair type (CIELAB gate) may still differentiate.
Direct Competitors
Research conducted August 2026, updated October 2026. GloVerse added after Forbes launch coverage shared by Cliff Smith, October 3 2026.
Myavana
AI hair analysis for Black women. Founded 2012 by Candace Mitchell Harris (Georgia Tech).
Website
Investors
Ulta Prisma Ventures (lead), Amazon, H/L Ventures, Reform Ventures, New Age Capital
Technology
HairAI strand analysis, personalized care plans, integrated into Ulta e-commerce. ~80% Black women customers, $10M+ revenue anticipated 2024.
Gap
Analysis only. No style visualization.
Plausible partner or acquirer as much as competitor. 14 years of textured hair data is the moat Avalon proposes to start building. Also proof the market is real and fundable.
Parfait
Selfie-driven wig customization. Pivoted to B2B, licensing AI to hair brands.
Website
Founders
Isoken & Ifueko Igbinedion, Marlyse Reeves, Simone Kendle. Black women founders, explicit equity-in-AI mission.
Investors
Upfront Ventures, Serena Ventures, Ulu, Unshackled, Contrary, Visible Hands, TRUE Capital Culture Fund, Tristan Walker.
Technology
Four selfies drive cap sizing, lace tinting, head shape analysis. Indique brand partnership. Now licensing the AI (P-10, R13 overlap).
Mixed to poor consumer reviews: sizing issues, color inaccuracy, unresponsive service, at least one unfulfilled paid order. Technology story is ahead of execution.
Wig path is occupied by a funded, technically credible competitor executing badly. Opportunity exists but it's a customer-service and supply-chain play, not a technology one.
HairHunt
The most direct product competitor. 320k+ users in 5 months. Moving fastest.
Website
Background
Owned by The Right Hairstyles, a large hair content site. Content-to-app funnel Avalon doesn't have. Launched November 2025. April 2026 update added 360-degree video preview and AI Style Match quiz.
Core loop
Selfie in, AI hairstyle preview out, roughly 10 seconds. Publishes detailed textured hair content including cornrows, locs, twists, afros.
"The face shape was changed enough that the result did not resemble her." User reviews confirm the failure mode Face Lock was designed against is real, at scale, and users are articulating it in exactly those terms.
This is the attack surface. The failure is confirmed by real users. Building the thing that doesn't do that is a real position, if it can be delivered.
Swivel Beauty
Digital marketplace connecting women with vetted natural-texture stylists.
Website
Background
Founded by Jihan Thompson, former magazine editor. Searchable by desired service (braids, cut, rod set), hair type, time, and price. Also surfaces product recommendations.
Overlap
This is the stylist directory and handoff wedge (ST-01, ST-02) already built and operating, with the supply relationships Avalon identified as moat item two.
Partnering with Swivel is more realistic than rebuilding it. Swivel has the stylists. Avalon has the render. That's a distribution relationship, not a competition.
GloVerse
New entrant — October 2026
On-demand beauty marketplace launched by Grammy-nominated rapper GloRilla. Targeting Black women in underserved markets. Beta live October 2026.
Website
What they are
Service marketplace ("Uber for beauty"), not a visualization tool. Connects consumers with stylists for on-demand, at-home professional beauty services. GloRilla as founder and face provides instant cultural distribution Avalon cannot replicate. Targeting cities where established beauty networks are thin — the same geographic underserved gap Avalon's stylist discovery addresses.
What they have that Avalon doesn't (yet)
On-demand mobile service mode (professional travels to consumer), men's grooming and barbering, multi-category beauty (nails, lashes, makeup), celebrity-driven social proof and press coverage, native app at launch, brand partnerships built in.
What Avalon has that GloVerse doesn't
AI render engine, Face Lock and CIELAB fidelity gate, feasibility verdict (can she actually get the style), photo-based hair type classification, style-to-stylist match logic. GloVerse answers "who should do my hair." Avalon answers "what should I do with my hair."
Supply-chain threat: GloVerse is onboarding stylists right now with celebrity distribution. Phase 4 was when Avalon planned to build stylist supply relationships (ST-03/05/07). Rev F accelerates those to Phase 3 in direct response.
Not a visualization competitor. The strategic play is integration, not opposition: Avalon renders → GloVerse books. The risk is that GloVerse locks up stylist supply before Avalon reaches Phase 3. The Rev F roadmap moves stylist features forward explicitly to prevent that.
Adjacent Landscape
| Company | What they do | Relevance to Avalon |
|---|---|---|
| OurX | DTC data-tech platform for personalized care for tightly textured hair. ~$2.5M seed, 2021. | Competes for the personalization and data narrative |
| Ruka Hair | Extensions company, UK. ~$1.6M raised 2021. Founder targeting ~$12M next round. | Product side of the wig and extension market |
| HaloBraid | Braiding materials as an engineering problem. | Materials, not software. Competing for the same investor attention. |
| Rebundle | Plant-fiber braiding hair. Raised grant and venture capital. | Materials, not software. Part of the $32M+ category raise. |
| HairstyleAI, Fotor, ArtSmart, Chroma Studio, VEED | Generic AI hairstyle filters, several explicitly marketing braids and locs. | The commodity layer. Positioning problem: Avalon must be visibly different within 5 seconds. |
| Ma Coiffeuse Afro | Booking app for people of African heritage to find appropriate hairdressers, France. ~EUR 350k seed. | Proves the model works internationally and the taxonomy question is live. |
| Oddity Tech / SpoiledChild (Nasdaq: ODD) | Public DTC beauty tech company. IL MAKIAGE (cosmetics) and SpoiledChild (personalized haircare, skincare, supplements). ODDITY LABS internal biotech using AI-driven molecule discovery. Quiz-first personalization (PowerMatch AI). $679M revenue, 658 employees. Profitable since 2020; stock ~54% off IPO price. Current weakness: customer acquisition cost crisis from ad-platform algorithm changes. Flagged by Cliff Smith, September 2026. | SpoiledChild's personalized haircare quiz overlaps with Avalon's discovery and recommendation layer. ODDITY LABS gives them a proprietary ingredient pipeline that could reach textured-hair formulations. Scale and data moat are significant. Not a direct visualization competitor today, but worth tracking as the space matures. |
Note on the commodity layer
HairstyleAI's app store listing markets "Realistic Face Preservation: your facial features stay completely natural, only the hair changes" as a headline feature. Face Lock is therefore not a differentiator. It is table stakes marketing. What may still differentiate is that Avalon proposes to measure it (the CIELAB fidelity gate) and report it by hair type, rather than merely claim it.
Strategic Implications
What the competitive research changes about the wedge decision above.
In-salon surface (A-13) strengthens
Consumer visualization is crowded. HairHunt has 320,000 users and a content funnel. The stylist-facing consultation tool is not crowded. Competitive research makes the stylist-first wedge a higher-priority decision than it appeared before this analysis.
Wig path weakens as a tech play
Parfait owns the technology narrative and is reportedly fumbling delivery. Competing there means competing on operations, not on technical superiority. Still an opportunity, but a different business than the one in this brief.
Partnership before building
Swivel has the stylists. Myavana has the brand relationships and a decade of textured hair data. For a small team, partnering on one or two layers is a materially better risk profile than building all four. Evaluate before Phase 1.
The honest strategic position
"The pieces exist and nothing is joined up" is the actual position. That's workable; most successful products start there. It requires better execution rather than just being first, and it changes the investor conversation.
GloVerse changes the stylist supply timeline
GloVerse is onboarding stylists now, backed by GloRilla's reach. Avalon's original plan to build stylist relationships in Phase 4 was written before a celebrity-backed competitor entered the category. Rev F moves ST-03 (stylist verification), ST-05 (portfolio profiles), ST-07 (geo-match algorithm), and new feature ST-10 (on-demand mobile service mode) from Phase 4 to Phase 3. The integration pitch — Avalon renders, GloVerse books — is the preferred outcome. But that partnership only happens from a position of supply relationships, not before them.
Risk Register
| Severity | Risk | Mitigation |
|---|---|---|
| Critical | Biometric privacy litigation (BIPA-style) | Consent architecture in Phase 0. This is the risk most likely to end the company and the cheapest to prevent. |
| Critical | A model lightens skin and it goes public | Face Lock (C1), fidelity gate (C3). Structural, not policy. Must be enforced before delivery. |
| High | Render quality insufficient on 4B/4C and braids | Curated launch catalog, honest confidence bands, Phase 3 fine-tune. Don't launch publicly until beta says it looks like them. |
| High | Corpus cannot be licensed at acceptable cost or speed | Stylist contribution program with real compensation (C8). This is the critical path, longer than the code. |
| High | Perceived as extraction from the community | Cliff as principal, compensated contribution, advisory group (R14), aggregate-only data (C9). |
| High | Corpus cold start never resolves | Commission a seed shoot (D-09). The stylist-only path has an unsolved cold-start loop. This costs money, which is why D-10 blocks D-01. |
| High | Retention collapses after novelty | A-12 lifecycle mechanics and the feasibility growth verdict. Instrument day 7 and day 30 from the first beta cohort. |
| High | Cliff's requirements differ from this document | D-03. This package is written from a secondhand verbal brief. Treat as a proposal to Cliff, not a plan agreed with him. |
| Medium | Incumbent builds textured-hair support | Moat is corpus and supply relationships, not the filter. HairHunt proves the filter is a commodity. Move fast on Phase 3. |
| Medium | Booking platform collision (StyleSeat, Booksy) | Feed the funnel, don't replace it. Charge for the lead. |
| Medium | Capture gate excludes part of the audience | P-12: measure rejection rate and break out by skin tone cohort. Treat a rising rejection rate as a defect, not user error. |
| Medium | Generation endpoint abuse | P-14: input and output classification with a named human who owns reports and incidents. |