Across the full book there are 162 churned merchants — excluding takeovers and now-active accounts, including fraud. This report focuses on the three highest-volume channels (114 of the 162), with a transcript-grounded deep dive on FS Canvas.
Value metric = expected annual TPV (£): the projected annual card volume captured at signup. Section totals sum these projections — several are round-number placeholders — so read them as an upper-bound indication of size, not realised volume.
One chart, grouped by reason, with the new coding on top and the old coding directly beneath it so each reason’s before/after is side by side. The new coding is shown as a share of the 162 churns (takeovers and active accounts excluded); the old coding keeps its own 109-churn base. Each label reads count then (% of churns). The old coding leaned on a broad “fraud / closed by acquirer” bucket and had no silent-churn category — the re-code splits that apart and surfaces silent churn and takeover.
What changed: on the 162-churn base the old “fraud / closed by acquirer” (32%) resolves into Fraud (25%) and Closed/sold (5%), with 19 takeovers reclassified out of churn entirely; Silent churn emerged as a top-3 reason (18%) where it had no category before; “issues with the acquirer / withheld funds” became Held payout; and “bad service” became “delayed delivery / poor service.” Net effect: a clearer view of what’s actually addressable.
Takeovers and merchants since re-activated are excluded from both the counts and the totals. Reasons are multi-select, so a merchant can carry more than one and counts exceed the merchant total.
Each labelled segment reads count · % of that channel’s churns — e.g. “6 (13%)” means 6 merchants — 13% of the channel’s churns.
Fraud is the largest single bucket in every channel. Setting it aside, silent churn leads the addressable reasons in FS Canvas and Cold call; Facebook is a three-way tie of silent, surcharge and competitor. FS Canvas reflects manual corrections: Nakos is a payout churn (not silent) and Osteria is silent (not technical).
The actionable reasons, each with the numbers up front (count, share, and total expected TPV) and the actual Retention / CS history behind the merchants. Merchant IDs throughout.
Diagnostek (10324005), Tobacconists (4568229), RK Gadget Hub (4493219), BIO Britware (10322153), House of Computers (10094324) and Osteria (10290189). None ever called in — they simply stopped transacting with no reason given. The play is early-life activation, not a retention save.
The biggest FS Canvas payout churn. Tuesday payouts don’t fit a cash-and-carry restocking on weekends/Mondays — needs daily. Cancelled free of charge.
Tuesday payouts not suitable for cash and carry visits on weekends and Mondays. Requires daily.Cancellation form
Tagged delayed-service + payout frequency, but her real story is a ~month-long held payout (Shift4, under investigation). Trust lost.
Reclassified from silent — the churn was a payout issue, not silence.
Mobile ice-cream van, busiest at weekends: Sunday takings don’t arrive until Tuesday. Compounded by a stalled bank-details change.
Any payment you take on Sunday, you don’t get [until] Tuesday.Retention call
News Plus was unhappy with payouts and surcharge and says the SM advised him incorrectly. Penn Store cites payout frequency with no further detail captured.
Says the SM advised him incorrectly — one of only four merchants across all channels citing explicit mis-selling.
Surcharge cost-transparency requests, then — decisively — unhappy with the 36-month binding period she says she’d never have signed.
Surcharge unhappiness stacked on a POS gap (no kitchen display) they say they already knew about, plus a Dojo preference.
The rest of the bucket resent the fee without a mis-sell narrative — CNR Trade also disliked the web app and Z-report and “doesn’t want Flatpay at any rate.”
Setting fraud aside, silent, surcharge and competitor account for the bulk of Facebook churn — and surcharge and competitor are tightly linked through a recurring Dojo pull.
MOSHRAHMAN (10318807), MK London (10093555), Munch Now (4673719), Angel Coffee (4532169) and Alba Media (10296676). Only Alba Media ever called — and raised no issue. Early-life silence rather than a hidden problem.
Cancelled in free trial — “feels like his trust has been broken” over the surcharge; the SM had already arranged to collect the machines.
Claims the SM didn’t advise of the surcharge; unhappy with contract length. Declined a 24-month, no-surcharge, 0.49% counter-offer.
Surcharge bundled with a Dojo preference: Arsenal “loved our terminal minus the surcharge” but is locked to Dojo; Ciao Bella pairs surcharge with battery-life and reporting complaints.
Locked to Dojo by a 2-year business loan — loved our terminal but can’t move. Lending is the lock-in we can’t match.
Previously won back, then a Z-report issue sent them looking; declined 0.56% / 24-month and 0.49%-with-a-Dojo-quote offers.
Moreish flatly “will change this company”; Floof Hut cites Dojo’s lower rate plus a connectivity issue on one branch.
The most varied channel: missing features and competitor co-lead, with surcharge, silent and payout close behind. Higher expected-TPV accounts appear here than elsewhere.
Pay by phone / MOTO is the decisive gap. Megna: 75% of payments are over the phone; Pinner Green MOT needs it too. Hot Wok left for a provider that has it.
The rest of the gap list: POS / till (BOOKS wants SumUp’s POS + digital tax; Meadow needs a POS), business loans (Mero’s wants funding, going to Dojo) and monthly invoicing (Mumms).
Wanted Monday cash-and-carry payouts; the Rapyd switch that would enable it disables AMEX — so we couldn’t satisfy both.
The two biggest competitor churns. Moca prefers Dojo’s reporting (+ surcharge unhappiness); Bhanjyang left over price — “too expensive” — and asked for an invoice.
Leaving for Dojo specifically because it offers the business loan they need — the clearest example of lending as the lock-in.
AYVA wants an invoice for the 36-month binding period on the way out; Stucchi is mid-contract with another provider and wants to leave in trial.
“Feels mis-led” on the surcharge — the explicit mis-selling claim, echoed by 4Lebanese Grill on Facebook.
SEB is losing customers over the fee and has a long-standing Dojo relationship; Health Defence wanted the surcharge removed on corporate/foreign cards.
Surcharge rarely travels alone here — it’s bundled with a competitor pull (Moca→Dojo reporting, BOOKS→SumUp POS + rate).
Wants her full revenue paid out, then one invoice at month-end — monthly invoicing, not per-transaction deductions. Mumms Cafe wants the same.
Cash-and-carry needing Monday-morning payouts to restock — and we couldn’t guarantee a morning payout even switching to Rapyd.
Both want daily payouts, which we don’t accommodate; both declined a rate reduction. Perfect Tailoring separately flagged a payout that arrived far short (£65 of a £553 sale).
Ember & Cork, Rendezvous, Life Barbers, Ocean Foods, Auto Arcade, Wendy’s Nails, Seafood For Kings and Viking. Only Viking (10095507) had inbound calls — trading-name edits — and it was Flatpay-cancelled for low TPV. Seafood For Kings (10084197) is the highest-value silent account and worth a direct win-back.
Three moves — two reason-specific fixes plus one structural lever.
Silent is a top-2 addressable reason in every channel and almost none of these merchants ever called — an early-life engagement gap, not a retention save. A first-90-days activation programme, with direct outreach to the higher-value ones (Seafood For Kings).
Surcharge is 11–19% of churns per channel. Proven mis-selling is a minority (4 explicit “SM didn’t tell me”), but it’s the version we control — make surcharge and the 36-month term explicit at signup. Late removal rarely saves the account.
A work-in-progress experiment to cancel merchants’ old providers on their behalf, so a churned merchant can’t simply fall back to a dormant previous account. Especially relevant against Dojo/SumUp, where the old contract often stays open and available. Early-stage — worth tracking as a structural retention lever alongside the reason-specific fixes above.