Skip to content
ForeGaze research · Trust · Part 1 · Fraud

What Fraud Costs a Small Business

16 min read · Data: FBI, FTC, Federal Reserve, FinCEN, Nacha, card networks and surveys, read 23–25 Oct 2026 · US, with California noted; one UK figure, labeled

ForUS small businesses, from solo owners to 500 people, that take paymentsUS (California noted)Canada (short section)Global and UK figures labeled

Fraud in 2026 for US small businesses, from solo owners to 500 people: what it costs, how it happens, and why payments still mostly work.

Losses reported to the FBI rose 26% to $20.9 billion in 2025 (all complaint types; cyber-enabled fraud was $17.7 billion of it). Payment systems (“rails”) mostly work; the big losses come when staff are tricked into sending money, as in business email compromise ($3.05 billion in 2025). Checks are the most-attacked payment type (58% of US organizations, AFP survey, 2025); online merchants worldwide put fraud at 3.5% of accepted orders in 2026.

$20.9BLosses reported to the FBI in 2025 across all complaint types, up 26% from $16.6B in 2024, from 1,008,597 complaints. Complainants from other countries account for almost $1.6B (FBI IC3).
$3.05BBusiness email compromise losses reported to the FBI in 2025, up 10%. Average $123,005 per complaint.
ChecksThe most-attacked payment type in 2025: 58% of US organizations saw check fraud attempts, against ACH debits 30%, wires 25% and corporate cards 21% (Association for Financial Professionals, AFP, survey of organizations of varying sizes).
76%Of US organizations saw attempted or actual payments fraud in 2025 (AFP 2026 survey, 465 treasury staff, organizations of varying sizes, self-reported). Not a loss rate.
Part 1 · What fraud costs · you are hereThe numbers$20.9B reported to the FBI in 2025 · $3.05B from business email compromise.
Part 1b · When fraud hitsWho pays, what to keep, what tools costThe business pays unless an exception applies · 2 banking days to return a business ACH debit · bank fraud controls cost $0–$75 a month.
The answer

The payment rails mostly hold. The weak point is people being talked into paying.

  • Most card volume is not lost to fraud. The newest published card loss rate is for 2024, and 2025 is due in January 2027 (Figure 1). The Fed has published no current fraud rate for bank transfers (ACH) or checks; its 2026 payments study has no fraud findings yet.
  • Push fraud is rising. That is when someone at a firm is tricked into sending (pushing) the money. Reported losses to the FBI rose 26% to $20.9 billion in 2025. Investment fraud (41% of losses) and business email compromise (15%) were the largest categories we checked.
  • Checks are the softest target. 58% of organizations saw check fraud in 2025. Stolen checks get washed, copied or forged.
  • Card fraud lives online. Online merchants worldwide put fraud at 3.5% of accepted orders in 2026, up from 3.0% (Merchant Risk Council and Visa Acceptance survey, global, merchant-reported). In the Fed’s latest US debit split, 2023, 54.3% of losses were card-not-present.
  • The cheapest control rates best. 94% of firms call back on a known number before paying a changed request, and of those that do, 94% rate it effective or very effective (63% very, 31% effective; AFP).
First step: never change who you pay because of an email. Call back on a number you already had, every time.
1 · Success vs failure · US and worldwide

How much fraud there is: start with the 2025–26 reports

The newest evidence is survey and FBI data from 2025–26. The card loss rates further down are 2024 data published in January 2026, the newest available; the 2025 rates are due in January 2027.

Read this firstWhy the numbers differ

Each figure counts something different

  • Value or cases. A loss rate counts dollars; a case count counts events. UK card fraud rose 3% by value and 13% by cases (UK Finance, 2025).
  • Gross or net. Some figures include money later recovered; the Fed counts losses to all parties before they are shared out.
  • Debit or all cards. The Fed’s 17.6 bp is US debit and prepaid at large issuers (2023); Nilson’s 6.43 is all cards worldwide (2024), and its volume includes ATM cash.
  • Place and denominator. US, world and UK are different pools, and the base can be card dollars, orders, transactions or complaints.
  • Survey rates are not loss rates. The survey figures below count how many firms or orders were hit. They cannot be set beside cents per $100 of card volume.

The merchant survey’s first-party-misuse share of fraudulent disputes was 17% in 2026, down from 20% (MRC and Visa Acceptance 2026 report, p. 8; merchant-reported, global).

SurveyUS organizations, AFP

76% hit by payments fraud in 2025

  • Share of organizations that saw attempted or actual fraud of any payment type. It is not a loss rate and not card-only.
  • AFP 2026 Payments Fraud and Control Survey: 465 treasury staff, fielded Jan 2026, organizations of varying sizes, self-reported (US). AFP release, 14 Apr 2026.
Survey, primaryOnline merchants, worldwide

3.5% fraud rate by order (2026)

  • Merchants’ own average: the share of accepted orders that turned out fraudulent (a trimmed average), online sales only. North America 3.4%, small and mid-size merchants 3.3%. Global, not US, merchant-reported. Up from 3.0% in 2025 (p. 27).
  • 64% of merchants report rising first-party misuse (2026, p. 8). First-party misuse is 17% of fraudulent disputes, down from 20% in 2024 and 2025; merchants put the average cost to resolve each such dispute above $80 (p. 8).
  • 61% report a rise in refund and policy abuse (2026, p. 9).
  • Method: 1,278 merchant professionals (69 are MRC members) in 37 countries, fielded Nov–Dec 2025 by an independent research firm. MRC and Visa Acceptance, 2026 Global eCommerce Payments and Fraud Report, copy supplied 25 Oct 2026 (page numbers are the report’s own). MRC release, 18 Mar 2026; report PDF.
SurveyConsumers, Canada

13% of Canadians had payment fraud

  • Consumer incidence only, over the six months before the Feb–Mar 2025 survey of 1,500 Canadians (±2.5 points). Canada, not US.
  • It gives no business figure and no loss rate. Payments Canada, Aug 2025.
Survey, primaryUS banks and credit unions

75% saw debit fraud attempts (Q4 2025 survey)

  • Debit was about 40% of payment fraud losses (p. 4). Check fraud attempts: 63%, and 31% of those had losses (p. 4).
  • Share reporting rising fraud, ACH: account-holder scams 41%, business email compromise 19%, account takeover 12% (p. 8). Wire: 38%, 20%, 14% (p. 9). The report’s charts order some rows differently from its text; we use the text.
  • Instant and real-time payments were “a relative bright spot,” with few attempts or losses (p. 5).
  • Method: 403 institutions (banks 75%, credit unions 25%), surveyed Q4 2025, no exact publication date. Institutions, not businesses; no small-business breakdown. Federal Reserve Financial Services, 2026 Risk Officer Report (Key Findings).
Vendor claim, method not statedUS businesses, Experian

60% say fraud losses increased (2026)

  • Experian 2026 U.S. Identity and Fraud Report blog post, 5 Aug 2026. Vendor, US; the page states no sample or method and does not split businesses by size.
Vendor studyUS retail and ecommerce firms, LexisNexis

$5.13 total cost per $1 of fraud loss (2026)

  • Includes costs beyond the loss itself. The US figure is $5.13 and Canada’s $5.23 (LexisNexis release, 24 Jun 2026).
  • Based on 513 risk and fraud leaders at consumer-facing firms in the US and Canada. The release does not state firm size, fielding dates or method, and we have not read the full study.
  • Vendor claim. Do not set it beside the loss rates above; it counts a different thing.
Vendor claim, method not statedUS financial institutions, NICE Actimize

ACH fraud volume +38%, value +52% (2025)

  • Vendor data from financial institutions, US payments, 2025; sample size and method not stated. ACH only, not all payments, and not a small-business or loss figure. Total ACH volume rose 23% and value 11% over the same year (p. 9). ACH was 0.8% of 2025 fraud value and 0.6% of volume in the same data (p. 4).
  • The fraud rate rose 2 basis points to 0.13%, up 12%. Attempted fraud per payment was $4.65, up 23% (p. 9). The report does not define the rate base, so do not set it beside the loss rates on this page.
  • The average ACH payment was about $3,240 (down 10%); the average fraud attempt was about $3,713, up 11% (p. 9).
  • NICE Actimize, 2026 Fraud Insights Report (U.S. payments), copy supplied 25 Oct 2026; page numbers are the report’s own.
Vendor network data, ecommerceOnline merchants, Signifyd

Fraud pressure +33% year over year (Jan–Apr 2026; vendor claim, global)

  • Signifyd’s own network of thousands of online merchants and 950 million digital wallets (p. 7). Global network, including Europe, the UK and Latin America. Ecommerce only, not small-business-specific; no merchant count or size mix given. “Fraud pressure” is the share of very risky orders on the network (p. 4).
  • First-party fraud and abuse rose 9% over the same four months (p. 4). Account takeover attempts rose 78% year over year in the first third of 2026 (p. 9). Card-testing attacks rose 175% year over year in the first four months of 2026 (p. 13).
  • These are changes in attack counts or shares, with no base figures, and they do not say how much money was lost.
  • Signifyd, State of Fraud Report 2026, copy supplied 25 Oct 2026; page numbers are the report’s own.
UK data, not USUK Finance, 2025

UK payment fraud £1.28B, up 4%

  • Losses were £1,279.8M in 2025, against £1,228.2M in 2024, the highest since 2021 (UK Finance Fraud Report 2026, published Jun 2026).
  • Remote card purchase fraud (card not present) lost £423.5M, up 3% by value on £412.5M in 2024. The number of cases rose 13%. A trade-press summary (Fintech Futures, 15 Jun 2026) gave 13% as the rise in value; the report gives 13% for cases.
  • This is a UK total in pounds, not a rate per $100, so it cannot be set against the US figures above. It shows the same direction: online card fraud is where card losses sit.
Figure 1 · card fraud, cents per $100 of card volume · worldwide and US estimate · 2023–2024 · newest published (Nilson, Jan 2026); 2025 due Jan 2027

Worldwide card fraud fell to 6.43¢ per $100 in 2024 from 6.58¢ in 2023; the US estimate is near 10.2¢.

0¢4¢8¢12¢ForeGaze · 2023: 6.58¢ per $100, worldwide (Nilson, 6 Jan 2025)6.58US 2023: 11.0¢ per $100 (ForeGaze division of Nilson shares)11.02024: 6.43¢ per $100, worldwide (Nilson, 7 Jan 2026)6.43US 2024: 10.2¢ per $100 (ForeGaze division of Nilson shares)10.220232024
Worldwide (Nilson)US (ForeGaze estimate from Nilson’s US shares)

One cent per $100 is one basis point (bp), or 0.01%. The rest of the world: 5.1¢ in 2024. Nilson’s volume includes ATM cash. Nilson’s 2025 figure is not out yet; it usually releases a full year the following January, so 2024 is the newest published year and 2025 is due in January 2027. The US figure is our division of Nilson’s published US shares of losses and volume, not a number Nilson publishes. Sources: Nilson releases of 7 Jan 2026 and 6 Jan 2025, read 23 Oct 2026.

Figure 2 · fraud loss, basis points of value · US debit and cards, one worldwide row · Federal Reserve and Nilson

Card fraud lost 6.43 bp of value worldwide in 2024; US debit lost 17.6 bp in 2023.

US debit cards (2023)Debit cards, US, all parties (2023): 17.6 bp of value (Fed Reg II report)17.6 bpUS all cards (2024)All cards, US (2024): about 10.2 bp of value (ForeGaze division of Nilson shares)10.2 bpWorld, all cards (2024)All cards worldwide (2024): 6.43 bp of value (Nilson)6.43 bp

There is no current official fraud rate for ACH or checks. The Fed’s last cross-rail comparison covers 2015–16, so we left it out; its 2025 triennial payments study (1 Jul 2026) reports 2024 volumes but no fraud findings yet. The nearest 2025 evidence is a survey, not a loss rate: AFP’s 2026 survey of 465 treasury staff (organizations of varying sizes, self-reported) found 58% of organizations saw check fraud and 30% saw ACH debit fraud, attempted or actual. A Fed survey of 403 banks and credit unions (Q4 2025; primary, institutions not businesses) found debit cards the most-hit rail: 75% saw attempts and 56% had losses. Neither gives a rate per dollar. Sources: Fed debit report, 19 Dec 2025; Nilson; Fed, 1 Jul 2026; Fed Financial Services, 14 May 2026; AFP 2026 survey highlights (Truist copy).

Where it fails. The rates above count unauthorized payments: stolen cards, forged checks. They miss payments a firm sends itself after being fooled. Those are rising, and the next section counts them.

2 · The official counts · US

Reported losses keep climbing

The FBI’s Internet Crime Complaint Center (IC3) and the FTC’s Consumer Sentinel count what victims report. Both rose again in 2025.

Figure 3 · losses reported to the FBI IC3, US$ billions · mostly US complainants · 2024–2025

Reported losses rose 26% to $20.9 billion in 2025; business email compromise was 15% of the total.

$8B$16B$24B2024: $16.6B reported lost$16.6B2024Business email compromise 2024: $2.77B2025: $20.9B reported lost$20.9B2025Business email compromise 2025: $3.05B
All reported lossesBusiness email compromise

FBI IC3 (US, complaints reported by victims): $20.877 billion lost in 2025, up 26%, from 1,008,597 complaints. Business email compromise: 24,768 complaints and $3,046,598,558 in 2025, against $2,770,151,146 in 2024. Source: FBI Internet Crime Report 2025, copy supplied 25 Oct 2026 (primary; PDF pages 4, 6, 7–9, 25–26; public link ic3.gov). Cross-check only: a law-firm summary gives the same 2025 totals (McDonald Hopkins, 7 Apr 2026).

FTCConsumer reports, 2025

$15.9 billion reported lost, up 27%

  • Up from $12.5 billion in 2024, from 3 million reports (FTC, 25 Mar 2026). Consumers, not businesses.
  • Bank payments carried the highest losses, then cryptocurrency; credit cards were the most often named (FTC testimony). The FTC gave a ranking, not dollar amounts, for 2025.
  • Bank payments are pushed, so they are hard to reverse. Firms meet the same scams as fake invoices, fake buyers and overpayment tricks.
Type (IC3)Complaints 2024Complaints 2025Lost 2024Lost 2025Average 2025
Business email compromise21,44224,768$2,770M$3,047M$123,005
Credit card or check fraud12,87618,774$200M$283M$15,056
Non-payment or non-delivery49,57256,478$785M$503M$8,913
Tech or customer support scams36,00247,794$1,465M$2,135M$44,664
Phishing or spoofing193,407191,561$70M$216M$1,127
Reported to the FBI, US. Card and check complaints rose 46% in number. Average = loss ÷ complaints, computed by ForeGaze; a few large cases pull it up, and the FBI doesn’t publish medians.

Shaded row: business email compromise has the largest dollar loss in this table ($3,047M in 2025, an average of $123,005 a complaint), so it is the row to read first.

FBI IC3California

First in losses: $3.67 billion

  • 116,414 complaints from California in 2025, the most of any state.
  • That is 18% of all reported losses. Texas ($1.83B) and Florida ($1.60B) follow.
  • We found no California rule on payment fraud that differs from federal rules.
FBI IC3Speed matters

58% frozen by the FBI Recovery Asset Team

  • The FBI’s process froze $679 million of $1.16 billion in 3,900 incidents in 2025, including 326 international ones. The report gives no split by how fast the report was made.
  • Call your bank first to recall the payment, then file at ic3.gov.
  • The report says time is of the essence.
FBI IC3AI-related

22,364 complaints, $893 million (2025; FBI report)

  • Complaints that mention AI averaged $39,946 each in 2025.
  • They overlap the other types, so they don’t add to the total.
  • The FBI says AI can be used to write boss-style emails and clone voices.
  • Complaints filed at IC3.gov by the public and by businesses (it includes more than $30M of AI-related business email compromise losses, p. 39); not a loss rate. Exact adjusted losses $893,346,472 from 22,364 complaints in 2025 (pp. 7, 8, 39). A complaint counts as AI-related when it contains a reference to AI (p. 61), so AI’s role is not verified.
  • FBI, Internet Crime Report 2025 (primary, US, self-reported by complainants; copy supplied 25 Oct 2026).
3 · How it happens · US

Seven ways money leaves a small firm

The Association for Financial Professionals (AFP) surveys corporate treasury staff each year. Its 2026 survey covers 2025, and checks still top the list.

Figure 4 · % of US organizations hit by attempted or actual fraud, by payment type · 2025

Checks were hit at 58% of organizations, nearly twice ACH debits at 30%.

ChecksChecks: 58% of organizations hit in 202558%ACH debitsACH debits: 30% of organizations hit in 202530%Wire transfersWire transfers: 25% of organizations hit in 202525%Corporate cardsCorporate cards: 21% of organizations hit in 202521%ACH creditsACH credits: 18% of organizations hit in 202518%Mobile walletsMobile wallets: 2% of organizations hit in 20252%Faster paymentsFaster payments: 1% of organizations hit in 20251%

AFP 2026 Payments Fraud and Control Survey: 465 treasury staff, fielded Jan 2026. Mostly larger firms, not a small-business sample. 76% saw some payments fraud; 74% saw business email compromise. Sources: AFP release, 14 Apr 2026; key highlights.

FBI, AFPBusiness email compromise

A real-looking request to pay a new account

  • A thief takes over or copies a vendor’s or boss’s email, then asks you to change bank details.
  • Of firms hit by business email compromise (BEC), 49% saw it via wires in 2025, down from 63%. Via ACH credits: 31%, down from 50%.
  • Nacha calls this paying under “False Pretenses”; its 2026 rules target it.
Fed surveyUSPISCheck fraud

63% of banks saw check fraud attempts (Q4 2025)

  • 31% of those banks and credit unions had losses from it (Federal Reserve Financial Services, 2026 Risk Officer Report, p. 4; 403 institutions, not businesses).
  • 58% of US organizations in the AFP 2026 survey saw check fraud in 2025, the most of any payment type (organizations of varying sizes).
  • Mail checks inside a Post Office or before the last pickup (US Postal Inspection Service).
Fed surveyAccount takeover

Someone signs in as you

SurveyFinCENDeepfake voice and video

7% of firms hit by fraud saw deepfake attempts (2025)

  • Share of AFP 2026 survey firms with fraud that saw deepfake attempts in 2025 (organizations of varying sizes, self-reported).
  • The FBI counted 22,364 complaints that mention AI in 2025 (see the FBI cards in section 2).
FTCPayment-app and transfer scams

Bank payments top 2025 losses

  • Consumers reported the highest 2025 losses on bank payments, then cryptocurrency (FTC; ranking only, no dollar amounts).
  • These are pushed payments, so they are hard to reverse.
  • Firms meet the same scams as fake invoices, fake buyers and overpayment tricks.
Networks, surveysChargebacks and friendly fraud

A real buyer disputes a real sale

FedCard-not-present fraud

Stolen card numbers used online

  • 54.3% of US debit fraud losses in 2023 were card-not-present.
  • Online orders carry the most risk of any type in the Fed’s 2023 split.
  • Your processor’s address and security-code checks and 3-D Secure (an extra check by the card’s bank) cut it.
4 · Cards and chargebacks · US

On debit cards, merchants now carry half the loss

The Fed’s debit card report shows who paid for fraud in 2023. Merchants paid more than banks and cardholders.

Figure 5 · US debit card fraud losses, 2023 · large issuers · bp of value and % of losses

Merchants bore 8.75 of 17.6 bp in 2023; card-not-present fraud was more than half the loss.

Who bore the loss, bp of value (17.6 bp in all)Merchants: 8.75 bp8.75Issuers: 4.96 bp4.96Cardholders: 3.83 bp3.83Merchants 8.75 bp · Issuers 4.96 bpCardholders 3.83 bpType of fraud, % of lossesCard not present: 54.3%54.3%Lost or stolen: 22.2%22.2%Counterfeit: 17.5%17.5%Other: 5.9%Card not present 54.3% · Lost or stolen 22.2%Counterfeit 17.5% · Other 5.9%

The Fed’s 2023 data are the latest available. The three shares add to 17.54 bp in the Fed’s table; the gap is rounding. Source: Fed, 2023 Interchange Fee Revenue … report, 19 Dec 2025, read 23 Oct 2026.

What this means for a small seller
  • In person, tap or dip. Counterfeit cards were 17.5% of US debit fraud losses in 2023, against 54.3% for card-not-present (Fed, large issuers).
  • Online, you prove it. For remote sales, the burden of proof sits with the merchant (Mastercard, 18 Mar 2024). Keep delivery and login records.
  • Repeat buyers help. Visa’s Compelling Evidence 3.0 lets you answer a fraud dispute with two earlier undisputed sales on the same card, 120 to 365 days old, where two details match, such as IP address, device or delivery address (industry guide; Visa’s rule text not read).
  • Watch your dispute ratio. From 1 Apr 2026 Visa flags a merchant once fraud reports plus disputes reach 1.5% (150 bp) of its online card transactions, counted by number, down from 2.2%. It applies only once a merchant has at least 1,500 fraud reports and disputes in a month (Visa fact sheet).
  • Your processor may set a lower line. Few small firms reach 1,500 disputes a month, but processors set their own limits and often do not publish them. unconfirmed
Dispute and chargeback fees in The Small Print →

Fees count too. A lost dispute costs the sale, the goods and a dispute fee. Merchants put the average cost to resolve a first-party dispute above $80 (MRC and Visa Acceptance 2026 report, p. 8; global, merchant-reported). Holds, reserves and payouts in The Small Print →

5 · Wires, ACH and the 2026 rules · US

Nacha now requires fraud monitoring from every business that sends ACH

Nacha writes the rules for ACH payments. In 2026 it made every party, down to a one-person business that sends payroll, watch for fraud.

  • 20 Mar 2026Phase 1: all sending banks (ODFIs), and businesses and processors that sent 6 million or more ACH entries in 2023. Big receiving banks must watch incoming credits (Nacha).
  • 19 Jun 2026Phase 2: every non-consumer originator, whatever its volume, and every receiving bank. The date fell on a holiday, so it ran from 22 Jun (Nacha).
  • What counts“False Pretenses”: payments induced by someone lying about who they are, who they act for, or whose account it is. Fake goods are not covered.
  • What you doKeep risk-based procedures and review them once a year. Which controls count is for you and your bank to decide; Nacha’s page does not list examples (not confirmed).
  • What you don’tYou need not screen every entry or check before sending. Whether you can rely on your bank’s or payroll provider’s monitoring depends on your agreement; the Nacha page we read does not say (not confirmed).
AFPWires

Fast and final

  • Wires were hit at 25% of organizations in 2025.
  • Once sent, a wire is very hard to pull back; speed of reporting decides it.
  • Call the bank within hours, not days.
AFPACH debits

30% hit; blocks help

  • An ACH debit pulls money out, so a thief needs only your account number.
  • Debit blocks and filters let only named companies debit you.
  • ACH positive pay asks you to approve unknown debits.
Fed surveyInstant payments

A relative bright spot

  • Banks saw few attempts and losses on instant payments (Fed Financial Services, Q4 2025).
  • Scams still work on any rail you push yourself.
  • Verify before you send, not after.
6 · By firm size

Where the risk sits as a firm grows

No official source breaks fraud down for firms under 50 people. These profiles combine the data above with the calculator’s example firms.

SoloOne person pays every bill, so no second pair of eyes.
  • Main risks: email takeover, fake invoices, payment-app scams.
  • Do: multi-factor sign-in, call-back on any bank change, card reader for in-person sales.
  • Example expected loss: $281 a year.
  • Inputs (our assumptions, ForeGaze calculation): card sales $150,000; 1% chance of a $10,000 email-fraud loss; 1% chance of a $5,000 check loss.
About 5 peopleA bookkeeper pays vendors; the owner approves sometimes.
  • Main risks: vendor impersonation, washed checks, keyed-in card sales.
  • Do: written call-back rule, ACH debit block, fewer paper checks.
  • Example expected loss: $1,906 a year.
  • Inputs (our assumptions, ForeGaze calculation): card sales $750,000; 2% chance of a $40,000 email-fraud loss; 3% chance of a $15,000 check loss.
About 20 peoplePayroll, many vendors, maybe online sales.
  • Main risks: payroll redirection, larger BEC wires, chargebacks.
  • Do: two-person approval over a threshold, positive pay, dispute evidence on file.
  • Example expected loss: $6,975 a year.
  • Inputs (our assumptions, ForeGaze calculation): card sales $3,000,000; 3% chance of a $120,000 email-fraud loss; 5% chance of a $15,000 check loss.
About 100 peopleSeveral people handle payments; payroll and the vendor list are large.
  • Main risks: the ~20-person risks at a larger scale: payroll redirection, large email-fraud wires, check fraud across more accounts, chargebacks.
  • Do: one person sets up a payee and another approves it, positive pay on every account that writes checks, dispute evidence on file.
  • Example expected loss: $19,099 a year ($13,125 cards, $4,920 email fraud, $1,054 checks).
  • Inputs (our assumptions, ForeGaze calculation): card sales $15,000,000 ($150,000 a person, as on the other cards); 4% chance of a $123,005 email-fraud loss; 7% chance of a $15,056 check loss. The two losses are the FBI 2025 averages per complaint; the sales and both chances are our own assumptions, not findings.
Survey, sponsoredUS small businesses, 0 to 500 employees

72% hit by fraud, scams or ransomware (2025)

  • Average losses of about $60,000 for payment fraud and over $90,000 for email compromise, among businesses that were hit.
  • 77% of those hit reported it; 54% of them said it was fully resolved and 33% partly.
  • Larger small businesses said they felt better prepared than smaller ones; the page gives no figures by size.
  • Method: 506 owners, fielded 19–29 Dec 2025 by Morning Consult for the Public Private Strategies Institute, weighted by size, industry and region. The institute names no funders on its page. A survey of owners, not records; no official source splits fraud by firm size. PPSI survey page; press release, 9 Apr 2026. The page and the release give different shares who feel prepared for payment fraud (40% and 44%), so we do not use that figure.

Example firms are illustrations, not data. Card losses use 8.75 bp, merchants’ share of debit fraud in 2023. Solo $131, ~20 staff $2,625, ~100 staff $13,125 a year in card losses. The chances of a loss, the card sales and the loss sizes on the three smaller cards are our own assumptions too, not findings.

100 to 500 people. We stop at about 100: we found no source for card sales, fraud odds or loss sizes above that, so the calculator has no preset beyond it. For a firm of 100 to 500 people, start from the ~100 preset and replace the card sales, the chances and the loss sizes with your own; the same controls (call-back rule, positive pay, two-person approval) apply to more accounts and payees. The FBI averages are per complaint, from all complainants, not by firm size.

7 · Spot it

Red flags, and a checklist you can tick

The red flags come from the FBI, FinCEN and the US Postal Inspection Service (USPIS). Tick the controls you already have; nothing is stored.

Red flags: stop and verify
  • A vendor or boss emails new bank details, especially with urgency or secrecy.
  • A request to pay by wire, gift card, crypto or a payment app you don’t normally use.
  • A voice or video call that won’t switch to your known number, or claims a “glitch.”
  • An invoice that matches a real one except the account number.
  • A buyer who overpays and asks for a refund to a different account.
  • A big online order with mismatched billing and shipping, or many tries on one card.
  • Checks you mailed that never arrived, or cleared for a different payee.
  • Sign-in alerts you didn’t trigger, or new email forwarding rules.
Protect your payments: tick what you have

0 of 12 in place.

8 · Your numbers

Expected fraud loss, and how fast controls pay back

Start from an example firm, then put in your own numbers. The chance of a loss is your estimate: no official source gives a small firm’s odds. The $25, $45 and $80 control prices come from two credit unions’ schedules; the When Fraud Hits post lists eight banks at $0–$75. The samples differ and are not combined.

Start from an example firm

Card fraud and dispute loss rate you bear

Positive pay fee a month

How much a control cuts the risk (your assumption)

Assumptions. Expected loss = chance × loss if it happens; card loss = sales × rate. Call-back and multi-factor sign-in cut email and impersonation risk; positive pay cuts check risk. Card losses are left to your processor’s tools. Positive pay fees from two credit unions’ 2025–26 schedules. Expected values understate a single large loss; the last line shows that risk. Where the inputs come from. Sourced: the 8.75 bp card rate (Fed, merchants’ share of 2023 debit fraud), 17.6 bp (Fed, all parties), $123,005 and $15,056 (FBI 2025 averages per complaint), and the $25, $45 and $80 positive pay fees (two credit unions). Our own assumptions, not findings: each example firm’s card sales, chances of a loss and loss sizes, the call-backs a month, 10 minutes and $40 an hour, the $0 multi-factor cost, and the 25%, 50% and 75% risk cuts. The card rate is a debit rate applied to all card sales, which is also our simplification.

At the ~5-staff example: expected loss $1,906 a year ($656 cards, $800 email fraud, $450 checks). Call-backs cost $240 a year and avoid $400, paying back in 7.2 months. Positive pay at $45 a month costs $540 and avoids $225. It pays only once expected check losses top $1,080 a year. One $40,000 loss equals 167 years of call-backs.

9 · In Canada

Canada: C$704 million reported, most of it never counted

The Canadian Anti-Fraud Centre (CAFC) collects reports nationwide. It says only 5% to 10% of frauds get reported.

CAFC2025

Over C$704 million, 112,000 reports

  • Investment fraud led at C$351 million (CAFC, Feb 2026).
  • Reported figures, not estimates of total fraud.
CAFCSpear phishing

C$67.9 million, second by loss

  • 813 reports and 571 victims in 2025, about C$118,914 per victim.
  • The CAFC defines spear phishing as a fraudster posing as a trusted source, such as a boss, supplier, client or landlord, with a spoofed sender address, to get a business or person to send money (CAFC).
  • The same call-back rule applies on both sides of the border.
10 · What to do

Six steps, cheapest first

  1. Write a call-back rule. No change to anyone’s bank details without a call to a number you already had.
  2. Turn on multi-factor sign-in for email, bank and payment accounts. Security keys and passkeys resist phishing best (CISA).
  3. Cut paper checks. Pay vendors by ACH. Mail any checks you still write inside a Post Office.
  4. Ask your bank about debit blocks and positive pay. Weigh the fee against your check and debit exposure in the calculator.
  5. Take cards by chip or tap. Online, turn on your processor’s address checks and 3-D Secure, and keep delivery proof.
  6. Write a one-page Nacha procedure if you send ACH payroll or vendor payments. Review it every year.
When it happens

Who pays when fraud hits, what proof to keep and what fraud tools cost are in the companion post, Part 1b. In most cases the business carries the loss; the exceptions and the time limits are there.

11 · Gaps

What this research did not cover

Open, not answered. Other parts of the Trust series pick some of these up.

  • Official fraud data for firms under 50 people. One sponsored survey covers firms of 0 to 500 employees but gives no figures by size.
  • A small firm’s yearly odds of a BEC or check loss.
  • The FTC’s 2025 dollar losses by payment method (only a ranking is public).
  • Median losses from the FBI; it publishes totals only.
  • Official approval-rate data for US card payments.
  • Processors’ own dispute-ratio limits for small merchants.
  • Fraud rates for instant payments and wires by value.
  • Payments Canada and Interac fraud rates.
  • Fraud findings from the Fed’s 2025 triennial payments study; its 1 Jul 2026 release gave 2024 volumes only. Until then there is no current official fraud rate for ACH or checks.

The Trust series. This is Part 1. Next: The Compliance Stack (Part 2); Built to Be Trusted: High-Trust Architecture for a Small Firm (Part 3); What AI Can Learn From Payments (Part 4); Safe, Private and Personal AI: a New Small-Business Type (Part 5); Robots and Your Business (Part 6).

Your next step

Protect your payments: tick the 12 controls

Most losses here start with one changed bank detail or one stolen check. Go back to the checklist, tick what you already have, and start with the cheapest gap: a call-back rule.

Open the protect-your-payments checklist

Rules and fees change; check with your bank and processor before you act. Provider names are trademarks of their owners. See Disclaimers above.