Flatpay · Retention Analytics

Churn Analysis Overview

Two pages: the United Kingdom page on the full true-infancy master, and a like-for-like Comparison page (official churns only) with UK and France side by side.

Sample — United Kingdom: 185 true infancy churners (official deactivations + shadow / silent churn), the full master. Rate cuts use the 170 that join the signed book; reasons the 109 with a Churn Category.
3,441
Signed since Oct 25
185
True infancy churners
4.9%
Churn rate (170 placed)
20 d
Median lifespan
Development in monthly churn rate
Total portfolio churn — merchants deactivated (> £2,125 TPV) each calendar month ÷ merchants active at month start · Nov 2025 – Jun 2026
0% 1% 2% 3% 1.0% Nov 25 1.7% Dec 25 1.1% Jan 26 1.8% Feb 26 1.2% Mar 26 2.0% Apr 26 2.4% May 26 1.4% Jun 26 Monthly churn rate (%) Month of churn (deactivation) · Nov 2025 – Jun 2026
UK monthly churn rate · rising through 2026 (peak 2.4% May); June partial
Sample & coverage by cut
Master = 185 true infancy churners (MIDs); each cut uses the subset that joins its source file
CutChurnersBaseNote
Time to churn (lifespan)184master lifespanactivation → last transaction
Merchant size, industry, trial, booking, city, signing170÷ 3,441 signed170 of 185 join the signed book
TPV realised170of 170needs Expected TPV from signed book
Reasons (multi-label)109Churn Category by MID76 not in the reason-coded set
Fraud deep dive33cancel_reasonfraud-flagged, full attribute coverage
Part ACohort cuts · churn rates

Rate cuts: true infancy churners ÷ signed within each segment (170 placed in the signed book) · dashed line = 4.9% portfolio average.

Cohort 1 · Lifespan

Two-thirds churn within a month of activation

On the master's own lifespan (activation → last transaction), churn is deep-infancy: 29% are gone within 7 days and 66% within 30; median lifespan ~20 days. A long tail (17%) limps past three months before stopping.

Churn by lifespan
True infancy churners grouped by days from activation to last transaction · n = 184
0 14 28 42 56 70 0% 20% 40% 60% 80% 100% 54 ≤7d 29% of churn 67 8–30d 36% of churn 29 31–60d 16% of churn 19 61–90d 10% of churn 13 91–180d 7% of churn 2 180d+ 1% of churn 29% 66% 82% 92% 99% 100% Churned merchants Cumulative % of churn Lifespan: activation to last transaction (master)
Churned merchants (count, left axis) Cumulative % of churn (right axis)
Cohort 2 · Merchant size

Bigger merchants churn far more

The steepest gradient yet: churn rate climbs 3.1% under £75k → 9.8% at £300k+, over 3×. The £150k+ bands alone are 53% of churn. Large expected volume is a risk marker, not protection.

Churn rate across TPV size cohorts
Churn rate by expected-TPV band · churns/signed under each bar
0% 2% 4% 6% 8% 10% 12% portfolio avg 4.9% 3.1% <£75k 41/1344 24% of churn 4.1% £75–150k 39/945 23% of churn 6.8% £150–300k 51/755 30% of churn 9.8% £300k+ 39/397 23% of churn Churn rate (%) Expected annual TPV at signing · bar label = rate, below = churns/signed
Churn rate by band (deeper = higher) · dashed line = portfolio average
Cohort 3 · Industry

Retail and Hospitality are the churn pool

By rate, Health & Wellness and Wholesale top out at 6.9% (small bases), then Retail 5.8% and Hospitality 4.7%. By volume it's overwhelmingly Retail (53) and Hospitality (69) — together ~72% of all churn.

Churn rate by industry (MCC)
Ordered by churn rate
0% 1% 2% 3% 4% 5% 6% 7% 8% portfolio avg 4.9% 6.9% Health & Wellness 5/72 3% of churn 6.9% Wholesale 7/102 4% of churn 5.8% Retail 53/912 31% of churn 4.7% Hospitality 69/1465 41% of churn 4.4% Services 33/749 19% of churn 3.2% Leisure 1/31 1% of churn 2.9% Tech/Infra 1/34 1% of churn 1.6% B2B Services 1/62 1% of churn Churn rate (%) Industry (MCC) · bar label = rate, below = churns/signed
Churn rate by band (deeper = higher) · dashed line = portfolio average
Cohort 4 · Trial period

Trial merchants churn more

A trial is associated with higher churn — 5.3% vs 4.5%, consistent with the earlier read: the trial enriches for early leavers rather than de-risking.

Churn rate by trial offered
Trial offered at signing vs not
0% 1% 2% 3% 4% 5% 6% 7% 8% portfolio avg 4.9% 5.3% Trial offered 95/1783 56% of churn 4.5% No trial 75/1658 44% of churn Churn rate (%) Trial offered at signing · bar label = rate, below = churns/signed
Churn rate by band (deeper = higher) · dashed line = portfolio average
Cohort 5 · Booking source

Field-sales and paid-social churn hardest

FS-Canvas and Facebook lead at 6.1%, cold-call 5.3%, MGM 5.0% — while TS-Marketing churns least at 2.2%, under half the field-sales channels. Lead source remains a strong predictor.

Churn rate by booking source
Sources with ≥100 signed · ordered by churn rate
0% 1% 2% 3% 4% 5% 6% 7% 8% portfolio avg 4.9% 6.1% FS - Canvas 52/855 31% of churn 6.1% MCC - Facebook 34/561 20% of churn 5.3% MCC - Cold call 47/880 28% of churn 5.0% FS - MGM 16/323 9% of churn 4.3% MCC - Call me 6/141 4% of churn 2.2% TS - Marketing 12/534 7% of churn Churn rate (%) Booking source · bar label = rate, below = churns/signed
Churn rate by band (deeper = higher) · dashed line = portfolio average
Cohort 6 · Signing month

Churn rate by signing month

Mature cohorts show the higher rates; the most recent months are right-censored — recent signings haven't had time to reach infancy-churn, so read them as not-yet-matured.

Churn rate by signing month
Recent months right-censored
0% 4% 8% 12% 16% 20% portfolio avg 4.9% 10.2% Oct'25 10/98 6% of churn 10.2% Nov'25 15/147 9% of churn 14.4% Dec'25 18/125 11% of churn 9.4% Jan'26 30/319 18% of churn 6.5% Feb'26 25/382 15% of churn 7.6% Mar'26 33/433 19% of churn 5.2% Apr'26 32/613 19% of churn 1.0% May'26 7/724 4% of churn 0.0% Jun'26 0/600 0% of churn Churn rate (%) Month signed · bar label = rate, below = churns/signed
Churn rate by band (deeper = higher) · dashed line = portfolio average
Cohort 7 · TPV realised — early vs late churn

Merchants churn before they ramp

Almost all churn at very low realised volume — 44% processed under 5% of expected annual TPV; median just 6.3%. They leave before getting going: an activation failure, not late-stage attrition.

Churn by processed TPV (% of expected annual)
Churned merchants grouped by share of expected annual TPV processed · n = 170 · median 6.3%
0 13 26 39 52 65 0% 20% 40% 60% 80% 100% 14 <1% 8% of churn 61 1–5% 36% of churn 44 5–10% 26% of churn 40 10–25% 24% of churn 9 25–50% 5% of churn 2 50%+ 1% of churn 8% 44% 70% 94% 99% 100% Churned merchants Cumulative % of churn Processed TPV as % of expected annual, at churn
Churned merchants (count, left axis) Cumulative % of churn (right axis)
Part BReasons · multi-label

Master churn MIDs × hand-classified Churn Category from "Merged Data (cancel + churns)", Churn-Analysis-FINAL · joined by MID · multi-select.

Sample. Of the 185 true infancy churners, 109 carry a hand-coded Churn Category (matched by MID); 76 fall outside the reason-coded set. Distribution is over the 109 coded churns, multi-label (avg 1.6 reasons each, so shares exceed 100%).
Reasons · classification

A more balanced mix than the deactivation-only view

On the true-infancy population the drivers even out. Fraud & acquirer closures are still the largest at 32%, but the addressable tier is right behind — overpromised / missing features 26%, surcharge 23%, competitor 22%. Capturing genuine early churn (not just fraud-driven deactivations) surfaces far more product- and price-addressable loss.

Number of reasons cited per churn
Multi-select · n = 109 coded churns
70
22
13
4
70 cite a single reason; 39 cite two or more.
Churn reasons by frequency of mention
Count and % of the 109 coded churns · multi-label
Fraud / closed by acquirer · involuntary
35· 32%
Overpromised product / Missing features
28· 26%
Surcharge
25· 23%
Competitor
24· 22%
Technical issue
12· 11%
Payout frequency
11· 10%
Contract
7· 6%
Closed store / bankruptcy · involuntary
6· 6%
Cancelled by Flatpay · involuntary
5· 5%
Issues with the acquirer/withheld funds · involuntary
5· 5%
Uncategorised
4· 4%
Bad service
4· 4%
Price
3· 3%
Voluntary / addressable Involuntary (fraud / acquirer / Flatpay)
Part CFraud deep dive · subset (50 of 185)
Subset view — not the full sample. This section drills into only the 50 churns flagged as fraud (via the master cancel_reason), a subset of the 185-churn master. Figures below are shares of that fraud subset (“% of fraud”) or of all signed merchants (“fraud rate”), as labelled — not of the full 185.
Fraud · who they are

Fraud is a large-merchant, paid-channel problem

Fraud is the costliest slice: these 50 merchants processed a median £12.0k — ~2.5× a normal churn — and account for 50% of all churned volume. They cluster in large expected-TPV signings from field-sales and paid-social channels.

50
fraud churners — the subset · 35% of the 185 churns, 1.5% of all signed
£12.0k
median TPV processed — ~2.5× the £6.2k of other churns
50%
of all churned TPV (£1.34M) from 35% of churns
66%
of fraud sits in £150k+ merchants — the large-merchant risk
Fraud · by merchant size

Fraud concentrates in large merchants

Fraud rate climbs from 0.4% under £75k to ~1.9% at £150k+, and 66% of fraud sits in the £150k+ bands — a large share of the "bigger merchants churn more" pattern in Cohort 2.

Fraud rate across TPV size cohorts
Fraud churns ÷ signed per expected-TPV band
0.0% 0.9% 1.8% 2.6% 3.5% portfolio fraud rate 1.5% 0.7% <£75k 9/1344 18% of fraud 0.8% £75–150k 8/945 16% of fraud 2.9% £150–300k 22/755 44% of fraud 2.8% £300k+ 11/397 22% of fraud Fraud rate (%) Expected annual TPV at signing · bar label = fraud rate, below = fraud/signed
Fraud rate by band (deeper = higher) · dashed line = portfolio fraud rate (1.5%)
Fraud · by channel

Field-sales and paid-social carry the fraud

Fraud over-indexes on the field-sales and paid-social channels; the marketing channel stays cleanest — the same inversion of the overall-churn pattern seen before.

Fraud rate by booking source
Sources with ≥100 signed · ordered by fraud rate
0.0% 0.9% 1.8% 2.6% 3.5% portfolio fraud rate 1.5% 2.9% MCC - Facebook 16/561 32% of fraud 2.8% FS - MGM 9/323 18% of fraud 1.6% FS - Canvas 14/855 28% of fraud 0.7% TS - Marketing 4/534 8% of fraud 0.5% MCC - Cold call 4/880 8% of fraud 0.0% MCC - Call me 0/141 0% of fraud Fraud rate (%) Booking source (≥100 signed) · bar label = fraud rate, below = fraud/signed
Fraud rate by band (deeper = higher) · dashed line = portfolio fraud rate (1.5%)
Fraud · by industry

Fraud rate by industry

Fraud rate is highest in the higher-ticket verticals; hospitality stays comparatively clean.

Fraud rate by industry
Industries with ≥30 signed · ordered by fraud rate
0.0% 1.2% 2.5% 3.8% 5.0% portfolio fraud rate 1.5% 3.9% Wholesale 4/102 8% of fraud 3.2% B2B Services 2/62 4% of fraud 2.9% Tech/Infra 1/34 2% of fraud 2.1% Services 16/749 32% of fraud 1.5% Retail 14/912 28% of fraud 1.4% Health & Wellness 1/72 2% of fraud 0.8% Hospitality 12/1465 24% of fraud Fraud rate (%) Industry (≥30 signed) · bar label = fraud rate, below = fraud/signed
Fraud rate by band (deeper = higher) · dashed line = portfolio fraud rate (1.5%)
Fraud signature. Large expected-TPV merchant (£150k+) · field-sales or paid-social lead · goes live, processes ~2.5× normal volume, disappears in infancy. Clearest lever: tighter underwriting on high-expected-TPV signings from those channels.
Fraud defined by the master cancel_reason (50 churns) for full attribute coverage — closely aligned with the Churn-Category "Fraud" tag in Part B.
Method · UK. Master: 20260629 UK True Infancy Churners — 185 MIDs (official + shadow). Rate cuts = churners ÷ signed (dashed = 4.9% avg); reasons = multi-label Churn Category by MID (109); fraud = master cancel_reason (33).
Sample — Comparison (like-for-like): official churns only. UK = 106 official infancy churners (shadow/silent excluded, 98 in the signed book) vs France 490 (> €2,500 deactivations) — both on an official-deactivation basis. Same chart style as the UK page (UK rust, France blue), difference noted on the right.
United Kingdom · official — 106 churns (98 in signed book) · 2.8% rate · 24d median
France — 490 churns · 3.7% rate · 53d median
Overall monthly churn rate — UK vs France
United Kingdom France
0% 1% 2% 3% Nov Dec Jan Feb Mar Apr May Jun 0.5%0.3%0.2%0.8%1.2%1.4%1.1%1.0% 1.0%1.7%1.1%1.8%1.2%2.0%2.4%1.4% Month of churn (deactivation) · common window Nov 2025 – Jun 2026 Monthly churn rate (%)
Merchants that churned (deactivated; >£2,125 UK / >€2,500 FR) each calendar month ÷ merchants active at month start, over the common Nov 2025–Jun 2026 window. The UK runs consistently higher every month (peaking ~2.4% in May vs France ~1.4% in April); both climb through 2026, and June dips as a partial month.
Time to churn (installation → last transaction)
United Kingdom · official
0 13 26 39 52 65 0% 20% 40% 60% 80% 100% 63 0–1m 64% of churn 17 1–2m 17% of churn 9 2–3m 9% of churn 5 3–4m 5% of churn 4 4–6m 4% of churn 0 6m+ 0% of churn 64% 82% 91% 96% 100% 100% Churned merchants Cumulative % of churn Months from installation to last transaction
France
0 37 74 111 148 185 0% 20% 40% 60% 80% 100% 183 0–1m 37% of churn 74 1–2m 15% of churn 58 2–3m 12% of churn 40 3–4m 8% of churn 53 4–6m 11% of churn 82 6m+ 17% of churn 37% 52% 64% 72% 83% 100% Churned merchants Cumulative % of churn Months from installation to last transaction
UK churns faster. Median 24 days to churn (installation → last transaction) vs France's 53; 64% of UK churn lands within the first month vs France's 37%, and France carries a much longer tail (17% past six months).
bars = churned merchants · dashed line = cumulative % of churn
Churn rate by merchant size (expected TPV)
United Kingdom · official
0% 1% 2% 3% 4% 5% 6% 7% 8% portfolio avg 2.8% 1.5% <£75k 20/1344 20% of churn 2.1% £75–150k 20/945 20% of churn 4.6% £150–300k 35/755 36% of churn 5.8% £300k+ 23/397 23% of churn Churn rate (%) Expected annual TPV at signing · bar label = rate, below = churns/signed
France
0% 1% 2% 3% 4% 5% 6% 7% 8% portfolio avg 3.7% 2.2% <€75k 110/4895 22% of churn 4.1% €75–150k 130/3157 27% of churn 3.7% €150–300k 102/2783 21% of churn 6.2% €300k+ 148/2382 30% of churn Churn rate (%) Expected annual TPV at signing · bar label = rate, below = churns/signed
Both: bigger merchants churn more. The UK gradient is cleaner and steeper (1.5% → 5.8%). France peaks at €300k+ (6.2%) but dips in the €150–300k band, so the signal is noisier.
bar = churn rate · dashed = portfolio avg · below = churns/signed · % of churn
Churn rate by industry
United Kingdom · official
0% 1% 2% 3% 4% 5% 6% portfolio avg 2.8% 4.2% Health & Wellness 3/72 3% of churn 3.9% Wholesale 4/102 4% of churn 3.8% Retail 35/912 36% of churn 3.2% Leisure 1/31 1% of churn 2.9% Tech/Infra 1/34 1% of churn 2.7% Hospitality 40/1465 41% of churn 1.9% Services 14/749 14% of churn 0.0% B2B Services 0/62 0% of churn Churn rate (%) Industry (MCC) · bar label = rate, below = churns/signed
France
0% 2% 4% 6% 8% 10% portfolio avg 3.7% 8.1% B2B Services 8/99 2% of churn 4.7% Wholesale 16/338 3% of churn 4.0% Tech/Infra 8/200 2% of churn 4.0% Hospitality 260/6505 53% of churn 3.7% Leisure 4/108 1% of churn 3.5% Services 77/2182 16% of churn 3.1% Retail 110/3591 22% of churn 1.9% Health & Wellness 3/154 1% of churn Churn rate (%) Industry (MCC) · bar label = rate, below = churns/signed
Different mix. France's rate is led by B2B Services (8.1%); the UK's verticals cluster lower and tighter. Hospitality sits mid-pack in both, not a rate outlier.
bar = churn rate · dashed = portfolio avg · below = churns/signed · % of churn
Churn rate by trial offered
United Kingdom · official
0% 1% 2% 3% 4% 5% 6% portfolio avg 2.8% 3.3% Trial offered 58/1783 59% of churn 2.4% No trial 40/1658 41% of churn Churn rate (%) Trial offered at signing · bar label = rate, below = churns/signed
France
0% 1% 2% 3% 4% 5% 6% portfolio avg 3.7% 3.8% Trial offered 132/3511 27% of churn 3.7% No trial 358/9709 73% of churn Churn rate (%) Trial offered at signing · bar label = rate, below = churns/signed
Trial matters in the UK, not France. UK trial churn runs 3.3% vs 2.4% without — a real gap. In France a trial makes essentially no difference (3.8% vs 3.7%).
bar = churn rate · dashed = portfolio avg · below = churns/signed · % of churn
Churn rate by booking source
United Kingdom · official
0% 1% 2% 3% 4% 5% 6% portfolio avg 2.8% 3.9% MCC - Cold call 34/880 35% of churn 3.5% FS - Canvas 30/855 31% of churn 3.1% FS - MGM 10/323 10% of churn 3.0% MCC - Facebook 17/561 17% of churn 1.4% MCC - Call me 2/141 2% of churn 0.7% TS - Marketing 4/534 4% of churn Churn rate (%) Booking source · bar label = rate, below = churns/signed
France
0% 1% 2% 3% 4% 5% 6% portfolio avg 3.7% 4.6% FS OG - Own Network 19/414 4% of churn 4.1% IH Canvas 163/3983 33% of churn 4.1% FS OG - MGM 94/2318 19% of churn 3.7% FS OG - Canvas 90/2409 18% of churn 3.4% IH Meta 94/2779 19% of churn 1.6% Upsale - Extra Agreement 3/190 1% of churn 1.0% TS - Marketing 6/607 1% of churn Churn rate (%) Booking source (≥150 signed, top 8) · bar label = rate, below = churns/signed
Marketing cleanest in both (UK TS-Marketing 0.7%, FR 1.0%). The UK's worst channel is cold-call; France's is own-network/referral. The UK spread across channels is wider.
bar = churn rate · dashed = portfolio avg · below = churns/signed · % of churn
Comparison method. UK = 106 official infancy churns (98 join the signed book); rates ÷ 3,441 signed. France = 490 > €2,500 all-time deactivations; rates ÷ 13,220 signed. Caveat: UK is infancy churn vs France all-time. Reasons are not compared — the taxonomies aren’t like-for-like (UK multi-label Churn Category vs France single-label cancel_reason). City and signing cuts also omitted (maturity-confounded).