SB StartupBasket
All ideas
69 /100 VALIDATE Medium complexity

Regulars — finder's-fee challenger for marketplace salons

Proves which 'new' clients were already yours and files the Boost, Fresha and Treatwell fee claims inside the window.

HomeServices SaaS Global Platform-contractFee-recoverySMBSolo-builderBeautyEvidence-fileAI-agent
— views
Evaluation Scores
69/100

VALIDATE

Overall Score

15
Problem
12
Demand
10
Build
11
Distrib.
10
Revenue
7
Time
4
Defense

Regulars

1. One-liner

Proves which “new” clients were already yours and files the Boost, Fresha and Treatwell fee claims inside the window.

2. Trend signal — why now?

Three booking marketplaces run the same billing trick on the same customer. Booksy’s Boost takes 30% of a new client’s first visit ($10 minimum, $100 cap in the US; £5 minimum in the UK). Fresha takes 20% with a $6 / £4 minimum. Treatwell takes 35% plus VAT, and its own partner documentation says a client who hasn’t booked through the marketplace in 365 days becomes “new” again. StyleSeat in the US does 30% up to $50. All four decide unilaterally who counts as “new”. The salon gets the invoice.

The salons say the attribution is wrong, in volume, with dates:

  • Booksy, Trustpilot, 26 Jan 2026: “Clients of mine went to my personal website…clicked my booking button…Despite that, Booksy still charged a 30% Boost commission.”
  • Booksy, Trustpilot, 10 Mar 2026: “My own clients finding me off of social media…Booksy feels like they come from boost…they keep denying my claims.”
  • Booksy, Trustpilot, 17 Aug 2025: “People who literally walk into my shop, scan our QR code, and book…are still counted as Boost.”
  • Booksy, Trustpilot, 8 May 2025: “We have clients who go to our other branch, yet Booksy still counts them as Boost clients.”
  • Booksy, Trustpilot, 3 Jan 2026: “Charged for a client that was not gained through Boost, despite providing proof.”
  • Booksy, Trustpilot, 27 Apr 2025: “You have taken over 20k from me over 9 yrs…you disbelieve me when I explained that the client…was not recommendation.”
  • Fresha, Trustpilot, 30 Aug 2026: “treated that person as a ‘new client’ they had acquired for our business and charged us.”
  • Fresha, Trustpilot (via TimeTailor’s review digest, Aug 2026): “This fee is charged even if customers were introduced to the salon by the salon’s Google, Facebook, Instagram, as well as word-of-mouth who have already been regular customers.”
  • Treatwell, Trustpilot, 13 Mar 2026 (Greek salon): “if an existing client has over a year to visit our salon and he/she uses Treatwell, he/she is considered a new client with a commission 25%”
  • Treatwell, Trustpilot, 3 May 2025: “even you key in client’s detail in the system…you are still charged total 42% Commission”

The platforms admit it in their own help centres. Booksy’s “How do I prevent being charged for Boost?” page opens with “sometimes we can’t keep up with you” and tells owners to add every walk-in “before they leave your shop”. Booksy built a formal in-app claim flow (Marketing → Boost → Boost details → Claim → reason + attachments) with a deadline of “your unique subscription billing period + 7 days” and a refund “within 14 business days” if confirmed. Fresha says the fee is “non-refundable”, but its wallet shows every “New Fresha client fee” with the client’s name and a “Client source history” showing when Fresha claims the client discovered you — which is the evidence you need to contest it. Fresha’s exemption rule is explicit: no fee if “the client already exists in your Clients list when they book their first appointment.”

Money is pouring into the fee-takers, not the fee-payers. Fresha took $80M from KKR at a $1B valuation on 21 May 2026 (130,000 businesses, $15B GMV) and finished killing its free plan by mid-2026, so every salon now gets an itemised monthly bill. Booksy raised $84M in late 2025 at a reported $580M (140,000 businesses). Treatwell claims 150,000 salons and 100M appointments a year, and added a mandatory £39+VAT monthly fee in 2026 on top of commission.

And the only people monetising the complaint are the switching vendors. Tavix, Setora, Slotcut, Heylilo, Blismo, TimeToBook, Calendy and Barber Insights all run “Booksy Boost explained” and “Fresha fee calculator” pages whose punchline is: “The only way to fully avoid commissions is to use a platform that doesn’t charge them.” Nobody sells the audit. Nobody sells the claim.

Provenance:

3. The opportunity

This is the counterparty-picks-the-bill shape again, but on a platform contract rather than a statute. The marketplace decides the variable (“new”), the salon pays the invoice, and the marketplace’s attribution logic is a black box that treats “viewed your Fresha profile once, then booked via your Instagram link” as a marketplace acquisition. Fresha writes that rule down: the fee applies to clients who “view your Fresha marketplace profile first and then choose to book through another online channel including your website or social media pages.”

Three specific things nobody does today:

  1. Nobody proves prior relationship. The salon’s evidence that a client was already theirs lives in five places the marketplace never looks: the old booking system’s export, the card terminal’s receipt history, Instagram DMs, the phone’s contacts, the other branch’s client list. Booksy’s claim form takes attachments. Fresha’s support takes screenshots. The salon owner is not going to assemble a dossier per £18 charge between cuts. Software will.
  2. Nobody runs the clock. Booksy’s claim window is the billing period plus seven days. Treatwell silently re-classifies a regular as “new” after 365 days of not booking via the marketplace. Fresha’s exemption only works if the client is in the list before the first booking. All three are timing problems, and timing problems are what software is for.
  3. Nobody prevents. The cheapest fee is the one that never fires. Keeping the Fresha Clients list and the Booksy contact list populated from every other source the salon already has (POS, old software, phone, DMs) is a nightly sync job, not a habit you can ask a barber to keep.

The incumbents are structurally blocked. Booksy, Fresha and Treatwell are the audited party; they will never sell you a tool that finds their own misattributions. AdminifAI ($300/mo, answers the phone and DMs for salons) integrates with all three and doesn’t touch fees. The switching vendors would rather you leave. The fee calculators (Tavix, RZRV) compute the headline cost and stop.

4. Target market

  • Primary customer: Owner-operators of 1–6 chair salons, barbershops, nail, lash and brow studios in the UK, Ireland, US and Western Europe who list on the Booksy, Fresha or Treatwell marketplace and do £3K–£30K a month. The UK barber on Booksy Boost is the beachhead: Booksy dominates UK barbering, Boost is the most-complained-about feature in its review history, and the claim flow already exists.
  • Why they buy: In their words: “they keep denying my claims”, “despite providing proof”, “charged me for clients they did not send you”, “leads to £100’s”. They lose £18–£100 per misattributed client and have no time to fight it. One reviewer put nine years of it at £20K.
  • Rough TAM reasoning: NHBF counts 61,000+ UK hair and beauty businesses; IBISWorld counts 1.08M US hair salon enterprises (mostly solos). The platforms themselves claim 130K (Fresha), 140K (Booksy) and 150K (Treatwell) businesses. Take 300K marketplace-listed businesses in English-speaking and Western European markets, assume a third run marketplace acquisition in any given month (the only ones who get charged): ~100K addressable. At $25/mo that is a $30M pool. We need 1%.
  • Why now for them: Fresha’s free plan is gone (2025, complete by mid-2026), so every salon now reads a monthly invoice with “New Fresha client fee” lines on it. Treatwell added a mandatory monthly fee in 2026 on top of 35%. Booksy’s own docs now concede the attribution problem and hand owners a claim button with a deadline. The bill got visible and the remedy got real in the same twelve months.

5. Product sketch (MVP)

  • Fee-line ingest: Upload (or forward) the Fresha fee activity download, Booksy Boost details export/screenshots and Treatwell invoices. Every “new client” charge becomes a row: platform, date, client name, service value, fee taken.
  • Prior-relationship match: Upload old client lists (Fresha/Booksy/Vagaro/Square exports), card-terminal receipt exports, phone contacts and DM screenshots once. Each charged client is matched against every source; the product shows what proves they were already yours and when.
  • Verdict per fee: “Yours — claim it”, “Genuinely new — pay it”, “Cancelled/no-show — shouldn’t have been charged”. No maybes without an explanation.
  • Claim drafter with the clock: For Booksy, a ready-to-paste claim reason, narrative and evidence pack, with a countdown to your billing period + 7 days. For Fresha, a support ticket citing the “already in Clients list” rule and the client source history. For Treatwell, a dispute letter per line and a warning list of regulars about to cross the 365-day cliff.
  • Prevention sync: A weekly reminder-and-file to push every client from POS, old software and phone into the marketplace’s Clients list before they book, so the Fresha exemption and Booksy’s “add walk-ins” rule actually fire.
  • Recovery ledger: Every claim filed, its status, the refund date, and the running total recovered — the number that justifies the subscription.
  • Free “Was this client really new?” checker: Paste one fee line and one old client list; get the verdict. This is the funnel.

6. AI angle — what’s load-bearing

The work is turning messy, multi-source evidence into a claim a platform reviewer accepts. Names in the Fresha wallet say “Jess M.”, the old Vagaro export says “Jessica Mullins”, the card terminal says “J MULLINS”, the Instagram DM is a screenshot that says “hey can I book Saturday”. A vision-plus-language model reads the screenshots, normalises the names and phone fragments, ranks the evidence, and writes the two-paragraph claim narrative in the tone that gets approved rather than the tone that gets “they keep denying my claims”. Multimodal inference got cheap enough in 2026 (August price cuts across the major providers) that reading a hundred screenshots per salon per month costs cents. Remove the AI and you have a spreadsheet nobody fills in, which is exactly what exists today.

7. Localization angle (if any)

UK-first, because Booksy Boost and Treatwell both bill in pounds with VAT on top, and UK barbering is where the loudest, most recent complaints cluster. Then US (Booksy, Fresha, StyleSeat’s 30%), then Treatwell’s continental markets (NL, DE, FR, IT, ES) where Treatwell’s 365-day re-classification rule is a distinct wedge. Language is the platform’s language; the claim narrative is generated in the salon’s locale. No payment-rail angle; card subscription is fine for this buyer.

8. Business model — path to $1M–$5M ARR

  • Pricing: £19 / $25 a month per location, or a success fee of 25% of recovered fees for salons who prefer paying only when it works. Treat the success-fee tier as the on-ramp; nudge to flat once recoveries are steady.
  • ACV: ~$300/location/year on the flat plan. Success-fee salons will average less but convert faster.
  • Rough math to $1M ARR: 3,300 locations × $25 × 12 = $990K. That is ~1% of the 300K platform-listed businesses, or ~3% of the ones being charged in any given month.
  • Rough math to $5M ARR: 12,000 locations plus a multi-location tier ($79/mo for 3–10 sites: chains are the ones with the “other branch” misattribution problem), plus expansion into the adjacent fee disputes on the same invoices (no-show fees charged after the salon cancelled, Fresha fees applied to a rescheduled first appointment, Treatwell commission on cancelled bookings).
  • Expansion path: Per-location pricing scales with chains; a “provenance ledger” upsell where every client’s first-touch source is recorded from day one becomes the salon’s permanent defence against any platform’s attribution.

9. Go-to-market wedge — first 100 customers

  • The Trustpilot list. Sixteen named Booksy reviewers and a dozen Fresha/Treatwell reviewers have complained in the last 18 months with their business names visible. Trustpilot pages show the salon; the salon’s Booksy/Fresha profile shows the phone. Message each one: “You wrote on Trustpilot that Booksy charged you Boost on your own clients. Send me your last Boost details export and I’ll tell you which ones you can still claim this billing period, free.” Expect 30–40% to respond; convert 10 of them into founding customers on the success-fee plan.
  • Barbers World and the UK barber Facebook groups. The Barbers World group has a post titled “Why does Booksy steal referrals and previous clients?” — that thread and its cousins are the lead list. Post the free checker, reply to every complaint with a one-line offer to run their export. Target 30 customers from three groups in six weeks.
  • TikTok “how to claim your client on Booksy Boost”. That search phrase already has discover pages. Post 60-second screen recordings: fee line, evidence match, claim pasted, refund landed. Link to the free checker. This channel compounds without spend.
  • The switching vendors’ audience, in reverse. Setora, Tavix and Barber Insights rank for “Booksy Boost commission” and tell owners to leave. Most can’t — “once everyone joins you can’t come off because you’re full”. Sponsor or guest-post the “if you’re staying on Booksy, here’s how to stop paying for your own clients” angle. Ten to twenty customers from one well-placed post.
  • Salon Geek forum threads on Treatwell. Long-running threads with UK salon owners naming the 35% and the re-classification problem. Same play: free run, then convert.

10. Build complexity — justification

Medium. There is no merchant API on any of the three platforms (Fresha’s “API report card” grade is F; the only surface is a paid Snowflake Data Connector exposing bookings, clients and sales tables). V1 lives on CSV exports, invoice PDFs and screenshots, which is fine because the volumes per salon are tens of lines a month, not thousands. The matching layer and the claim drafter are standard LLM work. The Booksy claim itself must be filed by the owner in-app (we produce the reason, narrative and evidence pack, and run the clock); automating the filing is a later, riskier step. A pair ships v1 in 10–12 weeks: ingestion for the three fee formats, evidence store, matcher, three claim templates, the recovery ledger and the free checker.

11. Gating checklist

GatePass?Note
Legal in target market✅Salon uses its own exports and the platform’s own claim process. No scraping of customer accounts, no ToS-breaking automation in v1.
Ethical — no harm / dark patterns✅We help salons pay only for clients the platform actually delivered. Genuinely new clients get a “pay it” verdict, not a template lie.
Market exists (evidence above)✅Dated complaints across three platforms; Booksy built a claim flow because the complaints are real.
1–5 person team can build this✅CSV/screenshot ingest + LLM matching + templates. No APIs to negotiate.
Launchable with <$50K / ₹40L✅Two people, off-the-shelf models, no data licensing.

12. Feasibility score

AxisWeightScoreNotes
Problem intensity2015/20Real, monthly, and quantified in the salon’s own words (£100s a month; £20K over nine years). Not hair-on-fire for everyone: a salon that never turns Boost on never feels it, and a single misattribution is £18. Regular pain with a bad manual workaround.
Demand evidence1512/15Multiple independent signals: dozens of dated reviews across three platforms, a Facebook group thread with the complaint in the title, a TikTok discover page for the claim process, competitor content farms built on the complaint, and the platform itself shipping a claim flow. What’s missing: anyone paying for the audit today.
Build feasibility1510/15No APIs anywhere; everything is exports and screenshots. Fine at v1 volumes, but three fee formats, a matcher and three claim templates is 10–12 weeks for a pair, and the in-app filing stays manual.
Distribution clarity1511/15Named reviewers, named Facebook groups, a pre-existing TikTok search phrase and a free checker that produces a dollar figure in one paste. Conversion from “free verdict” to paid is the guess.
Revenue mechanics1510/15$25/mo is benchmarked against what they already pay Booksy ($30) and Fresha ($15/member). $1M ARR at 1% of listed businesses is credible. Two soft spots: churn when a salon simply turns Boost off, and low ACV meaning we need thousands of logos.
Time to first revenue107/10Success-fee plan means the first invoice lands when the first Booksy refund lands: 14 business days after the first confirmed claim. 4–8 weeks from launch is realistic.
Defensibility104/10Execution moat only in month 3. By month 12 the per-salon provenance ledger (every client’s first-touch source, dated) is real switching cost, but a platform can neuter the whole thing by fixing attribution or restricting claims.
Total10069/100

13. Qualitative modifiers

Founder-fit tags

technical-heavy · sales-heavy

The build is unglamorous data plumbing plus LLM matching. The go-to-market is talking to barbers one at a time in Facebook groups and on TikTok until the free checker carries itself. A founder who finds either of those beneath them will fail here.

Key assumptions to validate (3–5)

  1. Assumption: A meaningful share of Boost/Fresha “new client” fees are contestable with evidence the salon already holds. How to test: Run the matcher on 20 salons’ last three months of fee lines against their old client exports; we need ≥15% of charged clients to match a prior record.
  2. Assumption: Booksy confirms claims when the evidence is organised, despite reviews saying claims are denied. How to test: File 50 evidence-backed claims across 10 salons in one billing cycle; track the confirmation rate and refund timing. Below 40% confirmed, the Booksy half of the business is a prevention product, not a recovery product.
  3. Assumption: Fresha support reverses “non-refundable” fees when shown the client was already in the Clients list or was a returning client. How to test: 20 support tickets with the source history screenshot and the prior record attached; count reversals.
  4. Assumption: Salons will pay $25/mo rather than only a success fee. How to test: Offer both to the first 50; if fewer than a quarter take flat after two recovered months, price on recovery only and accept the lower ACV.
  5. Assumption: The free checker converts. How to test: 500 checker runs from the Facebook/TikTok channels; need ≥5% paid conversion within 30 days.

Risk flags

  1. Platform dependency: All three platforms can change export formats, restrict the claim flow, or quietly fix attribution. Booksy building a claim flow cuts both ways: it validates the pain and signals they may reduce it.
  2. ToS risk on automation: Filing claims on the owner’s behalf inside Booksy Biz could be treated as unauthorised access. V1 keeps the human on the submit button; do not automate that step without reading the partner terms.
  3. Low ACV, high logo count: At $300/year you need 3,300 salons for $1M. Channel throughput, not product, is the bottleneck.
  4. Churn on behaviour change: The best outcome for the salon (turn Boost off, keep clients list complete) makes the subscription look optional. The prevention sync and the recovery ledger have to keep showing a number.

14. Structured verdict

Score:                  69/100
Verdict:                VALIDATE
Confidence:             Medium
Best-fit builder:       Technical founder who will personally work UK barber Facebook groups and TikTok for six months; ex-salon-software or ex-marketplace-ops experience is a bonus, not a requirement
Time to revenue:        4–8 weeks (first Booksy refund confirmed → first success-fee invoice)
Capital to launch:      £15–25K ($20–30K): two people part-time for three months, model costs, a few hundred pounds of paid group posts
Top 3 assumptions to validate first:
  1. ≥15% of charged "new" clients match a prior record — run 20 salons' exports through the matcher
  2. Booksy confirms ≥40% of evidence-backed claims — file 50 in one billing cycle and count
  3. Free checker converts ≥5% to paid within 30 days — 500 runs from two Facebook groups and TikTok
Kill criteria:
  - Abandon if fewer than 10% of charged clients across 20 salons match any prior record (the misattribution is loud but small)
  - Abandon if Booksy confirms under 25% of evidence-backed claims AND Fresha reverses under 10% of ticketed fees (no recovery lever, prevention alone won't carry $25/mo)
  - Abandon if any platform restricts export or the claim flow to the point where the owner must re-key evidence by hand

15. Next step — 1-week validation sprint

  • Day 1–2: Message the 16 Booksy Trustpilot reviewers and the 10 most recent Fresha/Treatwell ones. Ask for last three months of fee lines plus any old client export. Goal: 12 data sets in hand.
  • Day 3–4: Run the matcher by hand (spreadsheet plus a model in the loop). For each salon, produce: fees charged, fees with a prior-record match, £ value contestable, and a drafted claim for each. Send it back the same day. Goal: ≥15% match rate; at least 5 salons say “file it”.
  • Day 5: For the salons that said yes, file the Booksy claims inside their billing window and open the Fresha tickets. Decide go / no-go on two numbers: the match rate across the 12 data sets (need ≥15%), and the count of salons willing to pay 25% of recovery if the claims land (need ≥5 of 12). Refund confirmations arrive over the following three weeks and become the first case studies or the kill signal.

Interested in a detailed proposal?

Get a deep-dive with market research, competitive analysis, and implementation roadmap.

Contact us

info@startupbasket.ai