Marketplace-seller records: the safe way in, and whether to buy or build
A focused follow-up to the stress test. It answers the two questions that decide the plan: exactly when UK anti-money-laundering rules bite, and whether Dweise should buy an existing tool or build one. Every claim is evidence-labelled; nothing about "for sale" status is assumed.
Self-service software keeps Dweise clearly outside accountancy-service AML supervision. No neglected UK multi-platform seller-records tool could be verified for sale. The one close acquisition target (Vinta) covers only Vinted, so a purchase shortens time-to-market only partially.
"Verified" = confirmed against a primary/named source. "Estimated" = an inference or a figure from a secondary source to re-check. "Speculative" = a real tool that is not listed for sale, so any approach is off-market. The three legal labels follow the AML research: black-letter law, HMRC/CCAB interpretation, or a genuinely grey area needing a solicitor or AML specialist. This report is decision support, not legal or tax advice.
Executive decision
The stress test asked two open questions: does AML actually bite, and could Dweise buy instead of build? Both now have evidence-based answers, and together they point to a clear path.
The safe model is clear
Self-service software, where the seller decides and owns every classification, sits outside accountancy-service AML supervision. Only a managed clean-up drags Dweise inside. Guidance
No clean acquisition exists
No confirmed UK listing was verifiable. The closest target, Vinta, is Vinted-only and actively run (not neglected). Buying shortens time-to-market only partly. Speculative
The gap is genuinely unowned
Upload-first, ledger-free records across several platforms · crucially including Vinted and Depop · is served by nobody. Verified
The AML & regulatory boundary
The single factor that decides whether Dweise needs AML supervision is: who decides the categorisations and owns the final output? The seller, and you are a software vendor. Dweise's staff, and you are an accountancy-service provider.
Decision flow: which side of the line are you on?
AML boundary map
| Model | AML supervision? | CDD on clients? | Other key duties | Risk | Verdict |
|---|---|---|---|---|---|
| (a) Self-service software seller confirms & owns output |
No Guidance pure software vendor; HMRC software carve-out |
No | ICO fee (Tier 1 £52) + UK GDPR; Consumer Rights Act; subscription rules. No MTAR/agent if no HMRC contact. | Low | Recommended. Keep the user deciding; avoid "we do your books" claims. |
| (b) Managed clean-up Dweise decides & returns books |
Yes Legal "recording, reviewing, analysing... for other people" |
Yes · full CDD/EDD on every seller | HMRC/professional-body registration; MLRO; risk assessment; policies; training; 5-yr records; SARs; £300 + £400/premises. | High | Avoid initially, or run only under a supervised structure. |
| (c) Through a supervised accountant sell to / operate under a practice |
No for Dweise Guidance the practice is the ASP |
Practice does it, not Dweise | UK GDPR data-processing agreement; solicitor-drafted contract allocating AML responsibility in writing. | Low–Medium | Strong B2B alternative. Only if oversight is real, not nominal. |
Data protection Legal
Pay the ICO data-protection fee (Tier 1 £52) and meet UK GDPR duties · lawful basis, security, breach reporting, DPIA if high-risk. Applies in every model.
Tax-adviser registration Legal, phasing
Mandatory Tax Adviser Registration starts 18 May 2026 but only bites if you interact with HMRC on a seller's behalf and are paid. An accountant-ready pack the seller uses does not trigger it.
Consumer & subscription Legal
Consumer Rights Act 2015 + distance-selling info/cancellation rights. Build subscription flows for the DMCCA 2024 regime (clear info, renewal reminders, easy exit), expected ~spring 2027.
Fee figures (HMRC £300/£400; ICO £52) and the verbatim wording of HMRC's software carve-out should be re-checked on the live gov.uk pages before quoting externally. Re-check
The safest initial operating model
Ranked from safest to heaviest for an early-stage business. The gap between rank 1 and rank 4 is the whole AML compliance build.
Pure self-service SaaS
Seller uploads, confirms each classification, owns the output. Outside AML, outside MTAR. Only ICO + GDPR + consumer law. This is the recommended launch model.
B2B: sell to supervised accountants
The practice is the regulated party; Dweise is a vendor/processor. Adds a data-processing agreement, no AML registration. Also the best distribution channel.
Operate under a practice's supervision
Relies on the "work reviewed by a registered accountant" exemption. Safe only if the oversight is genuine and papered by a solicitor. Fragile if nominal.
Managed bookkeeping / clean-up
Dweise decides categorisations. Full AML build: registration, MLRO, CDD on every client, SARs. Viable later, wrong for launch.
Build versus buy
Five routes to market, compared on speed, cost, AML exposure and time-to-market advantage. Select a route to see the detail.
Reuses Dweise's records engine. Builds the differentiator (ledger-free, multi-platform incl. Vinted/Depop) that cannot be bought. Main risk: it is a build, not a shortcut.
Quick to cash but pulls Dweise inside AML supervision and does not scale beyond bespoke consulting. Only as a supervised, later move.
One practice exposes many sellers and carries the AML load. Slower to close each deal, but the best distribution. Pairs naturally with the build route.
No confirmed listing verifiable. Best off-market target (Vinta) is Vinted-only, so it does not deliver the multi-platform wedge on its own.
Buy an existing user base / Vinted-native piece (e.g. Vinta) or a reconciliation engine (e.g. Taxomate), then bolt on the ledger-free multi-platform core. Accelerant, not foundation.
Acquisition landscape
Strictly separated: confirmed live listings versus speculative (real but not-for-sale) targets. Nothing here is assumed to be for sale without evidence.
Speculative targets · real tools, verified to exist, not listed for sale Off-market
| # | Product | Niche | Owner / maturity | Neglect & scale signal | Price | Fit |
|---|---|---|---|---|---|---|
| 1 | Vinta vinta.app | Vinted resellers; UK tax-ready records, labels, backdated import | Solo founder (a UK Vinted reseller); public GitHub repo | Single-operator = key-person risk. Actively blogging, so not neglected | £20/mo or £49 lifetime | Closest fit · "buy the product you'd build", but Vinted-only |
| 2 | Taxomate taxomate.com | Amazon/eBay settlement reconciliation into Xero/QBO; UK/EU VAT | Small indie; owner unverified Est | Cheaper A2X/LMB challenger = small player. Neglect not confirmed | Low tiered | Reconciliation + VAT engine; needs a ledger though |
| 3 | Seller Ledger sellerledger.com | Reseller accounting eBay/Etsy/Amazon | Ex-Outright / GoDaddy Bookkeeping team | Established, US-centric | Low | Mature codebase; UK VAT/MTD would need building |
| 4 | Easy Auctions Tracker easyauctionstracker.com | eBay-only bookkeeping (Excel add-in) | Long-standing indie | Strong neglect: reported winding down (2024); testimonials dated 2020 | $29.99 one-time | Buy for customer list/domain, not tech |
| 5 | Reseller spreadsheet products Etsy / Gumroad / Payhip | UK sole-trader resellers, SA-ready | Solo micro-sellers | Cheap side projects, many stale | £5–£20 | Not software · buy an audience/content, seed a SaaS |
vintasoftware on LinkedIn/Crunchbase is a separate Brazilian dev agency, not the Vinted tool. The target is vinta.app only. Do not conflate them in diligence.Acquisition scorecards
Directional fit scores (out of 5) for the two targets worth a real conversation. Estimated
"Buy the product you'd build" for the Vinted slice, with a user base and HMRC-aligned content. But it covers one platform only and leans on a fragile Vinted scrape; verify it is genuinely his to sell and untangled from the public GitHub repo licence.
Brings a working settlement-reconciliation + VAT engine (months of build). But it is ledger-oriented and Amazon/eBay-focused, so it does not solve the ledger-free, Vinted/Depop wedge. Ownership and financials unverified.
Valuation benchmarks & max-risk thresholds
If a target does become available, price it against 2026 micro-SaaS benchmarks. Smaller deals trade at lower multiples.
2026 micro-SaaS benchmarks Estimated market
| Benchmark | Multiple |
|---|---|
| Acquire.com · confirmed closed SaaS (Jan 2026) | ~3.9× profit (median) |
| Acquire.com · pure-SaaS average asking | 2.6× revenue / 10.7× profit |
| Flippa · micro-SaaS under $1M ARR | 2.85× profit (top quartile 6.1×) |
| Flippa · $10k–$100k deals | ~1.68× profit |
| Flippa · sweet-spot band | $100k–$500k (avg $323k) |
Recommended acquisition criteria
- Price ceiling: ≈ 3× SDE or 2.5× revenue; push toward 1.5–2× profit for sub-$100k deals.
- Must have: genuine UK-seller relevance, transparent Stripe/bank-verified revenue, and code/IP that is genuinely the seller's to transfer.
- Structure: asset purchase via a new Ltd; TOGC if trading continues; escrow + a 12–24 month holdback/earn-out.
- Walk away if: revenue unverifiable, single fragile scrape as the only data route, licence entanglement, or price above ceiling.
Marketplace fees to budget: Acquire.com 6–8% (8% under $250k); Flippa from 3%; Microns 6–10%. Est
Due-diligence checklist
Run this on any target before an offer. Failing a "must-pass" item is a walk-away, not a negotiation.
Financial
- Stripe + bank reconciled to the P&L (verified SDE)
- MRR, churn and trend (12 months)
- Refund/chargeback rate
- Customer concentration
Technical & data
- Code ownership & licences (watch public repos)
- Data routes: official API vs fragile scrape
- Platform-API dependency & enforcement risk
- Security posture; where seller data sits
Legal & regulatory
- Is it (or has it acted as) an AML-supervised service?
- ICO registration / UK GDPR compliance
- Trademark & domain ownership
- Any consumer-terms or advertising claims risk
Commercial
- Real, contactable customers (references)
- Acquisition channels & their durability
- Reviews / reputation (app stores, Trustpilot)
- Brand fit with Dweise
Key-person
- Founder dependency for support/dev
- Handover / transition terms
- Undocumented "tribal" knowledge
- Non-compete on exit
Integration
- What is retained vs replaced
- Migration onto Dweise's engine
- Cost/time to UK-localise (VAT/MTD)
- Does it truly shorten time-to-market?
Risk matrix
The main risks plotted by likelihood and impact. Anything in the red zone needs a mitigation before committing money.
Strongest overall recommendation
The self-service model keeps Dweise safely outside AML supervision. The wedge · upload-first records across several platforms including Vinted and Depop, for sellers who do not run a ledger · is unowned and cannot be bought whole (Vinta is Vinted-only; Taxomate needs a ledger). So the differentiator must be built regardless. Acquisition is worth a parallel look because it could hand you a user base and Vinted-native code, but it is an accelerant, not a substitute for the build.
Primary
Thin self-service upload → normalise → resolve → accountant pack. eBay + Vinted + one more, user-confirmed classifications.
Channel
Marketplace-specialist accountants as the B2B route (they carry AML and expose many sellers).
Accelerant
One exploratory, no-pressure conversation with Vinta; a logged-in sweep of the acquisition marketplaces.
90-day execution plan
Four phases. Each is earned by evidence from the last. Regulatory and validation work run in parallel with the acquisition scan.
Pass / fail gates (the pilot must clear these)
| Gate | Pass condition | Why it matters |
|---|---|---|
| Payment | 3 sellers pay full price, or 2 practices commit real client files | Separates anxiety from genuine demand |
| Automation | ≥80% of transactions normalised/matched automatically | Manual work kills the margin |
| Time | Median handling < 90 minutes per customer | Tests whether it scales beyond consulting |
| Evidence burden | <10% of records need extended chasing | Missing purchase evidence destroys unit economics |
| Margin | Gross margin > 70% after support | Required for a low-ticket product |
| Access | Works on user exports / supported APIs only | Avoids fragile scraping & enforcement risk |
| Regulation | AML/GDPR model reviewed and signed off | Stops a good pilot becoming a liability |
Sources & evidence
Official sources for tax, AML and regulatory claims; vendor and marketplace sources for products and benchmarks. Vendor pages prove current offers, not independent quality.
Legal The statutory definition that pulls a business into AML supervision.
Guidance Scope of "accountancy services", the software carve-out, and exemptions.
Legal MLRO, risk assessment, policies, CDD, records, SARs.
Re-check £300 application + £400/premises figures move; confirm before quoting.
Guidance Flags that software may increasingly be seen as a tax service.
Legal Tier 1 £52; UK GDPR duties.
Legal, phasing From 18 May 2026; only if you interact with HMRC and are paid.
Verified (FOI) 3,988,892 seller reports for 2025, up 272% on 2024.
Verified £50k Apr 2026 → £30k 2027 → £20k 2028.
Verified Vinted-only; £20/mo or £49 lifetime; Chrome-extension data route.
Verified Requires Xero/QBO/NetSuite; no Vinted/Depop.
Verified UK-made; requires Xero/QBO; no Vinted/Depop.
Speculative targets Real tools, not listed for sale; details partially unverified.
Estimated benchmarks ~3.9× profit closed (Acquire); 2.85× profit micro-SaaS (Flippa).
Guidance UK deal-structuring for an asset purchase.
Verified eBay/Etsy/Shopify give usable financial data; Vinted Pro-only, Depop partner-gated → upload-first.
Carry-forward caveats: exact HMRC AML/ICO fee figures and the verbatim software-carve-out wording need a live gov.uk re-check; the "£40m HMRC campaign" figure circulating in accountancy press is not corroborated against a primary HMRC publication; Vinta's "HMRC-compliant" is marketing, not formal recognition; no confirmed live acquisition listing was verifiable from open research and must be checked logged-in. This report is decision support, not legal or tax advice.