benchmarks 2026-04-18 15 min read the underwriting desk

Chargeback ratio benchmarks by vertical, 2026

3-minute scan
  • Median chargeback ratio across 18 verticals sampled: 0.42%. Range from 0.11% (B2B SaaS) to 1.27% (kratom).
  • The VAMP threshold matters more than the legacy Visa 1.0% threshold in 2026. p75 of VAMP-exposed verticals is now uncomfortably close to the 0.65% tier.
  • Refund rate and chargeback ratio are not substitutes — a 4% refund rate does not necessarily reduce chargeback rate, contra popular advice.
  • Representment win rate was more predictive of acquirer good standing than raw chargeback ratio in our sample.
  • Subscription verticals had the highest "friendly fraud" share of chargebacks: 71% of subscription chargebacks were reason code 4853 or 10.4.
On this page

    Every operator wants to know: "am I normal?" Chargeback ratio is the number acquirers actually look at, and operators routinely over-index on rumors and under-index on their own vertical's real distribution. This benchmark report is built from 410 operators across 18 verticals, covering $3.1B in trailing-twelve-month processed volume between April 2025 and March 2026.

    We report median, 75th percentile, and 90th percentile per vertical. We report "counts-based" ratios (Visa / VAMP methodology) and "dollars-based" ratios where we had the data. We flag where VAMP thresholds bite in 2026 and where the legacy Visa Dispute Monitoring Program (VDMP) thresholds still apply.

    Methodology: respondents supplied processor-level chargeback reports or gateway export dumps. We excluded fraud-assistance fraud (3rd-party fraud that never produced a dispute) and pre-arb/pre-comp events that did not progress to formal chargebacks. Ratios are count-of-chargebacks / count-of-authorized-transactions on a rolling 30-day window unless otherwise noted.

    The headline table: 2026 benchmarks by vertical

    VerticalMedianp75p90VAMP-exposed?
    B2B SaaS0.11%0.19%0.34%No
    B2C SaaS (prosumer)0.27%0.44%0.72%Yes
    DTC apparel0.31%0.54%0.81%Yes
    DTC beauty0.38%0.62%0.94%Yes
    Supplements (mainstream)0.44%0.71%1.08%Yes
    Nutra (high-claim)0.58%0.91%1.41%Yes
    CBD (ecom)0.47%0.79%1.22%Yes
    Kratom0.69%1.04%1.63%Yes
    Peptides (research)0.61%0.97%1.48%Yes
    SARMs0.74%1.15%1.82%Yes
    Telehealth / TRT0.39%0.63%0.96%Yes
    Vape / ENDS0.54%0.84%1.29%Yes
    Subscription boxes0.62%0.93%1.37%Yes
    Marketplaces (3P)0.48%0.78%1.19%Yes
    Digital goods / ebooks / courses0.81%1.21%1.79%Yes
    Travel / hospitality0.42%0.67%1.02%Yes
    Adult / 18+0.89%1.31%1.94%Yes
    Crypto onramp1.04%1.52%2.24%Yes

    Two quick reads. First, the median operator in most DTC and supplement verticals is well below 0.5% — the freeze-trigger point. But the p75 and p90 creep into uncomfortable territory. Second, digital goods, adult, and crypto onramp sit structurally high — if you are in those verticals and you are at the median, you are already above the VAMP Excessive threshold and your acquirer knows it.

    VAMP changed the benchmark, not just the ceiling

    In 2026, the dominant compliance program is no longer Visa's legacy Dispute Monitoring Program (VDMP) — it is VAMP. For a full explainer, see VAMP explained for operators. For this benchmark's purposes, the key numbers are:

    ProgramTierThreshold (count)Threshold (dollars)Practical acquirer action
    VAMPAbove Standard≥ 0.50% dispute ration/aAcquirer coaching, fine exposure
    VAMPExcessive≥ 0.65% dispute ration/aAcquirer fines, often reserve increase
    VDMP (legacy)Early Warning≥ 0.65%≥ 75 disputes / moWarning letter
    VDMP (legacy)Standard≥ 0.90%≥ 100 disputes / moMonitoring, higher fees
    VDMP (legacy)Excessive≥ 1.80%≥ 1,000 disputes / moAcquirer fines, MATCH risk
    Mastercard ECPExcessive≥ 1.50%≥ 100 disputes / moSimilar escalation path

    The critical number for 2026 operators is 0.65% under VAMP. That is the line at which acquirers are getting their knuckles rapped by Visa, and they pass the pain directly through to merchants. Crossing 0.65% routinely — which our p90 shows is the baseline reality for most high-risk verticals — is the dominant driver of reserve increases in our 2026 data.

    How to interpret the percentile columns

    The median is the "normal" operator. The p75 is the operator your acquirer starts to notice. The p90 is the operator on the risk list. Use them as follows:

    1. Look up your vertical.
    2. Compare your own trailing-30-day chargeback ratio to the column that matches your operator type.
    3. If you are above p75 in a VAMP-exposed vertical, you are one bad month from Above Standard.
    4. If you are above p90, you are already being scored as a higher-risk merchant by your acquirer's risk model whether or not they have communicated it to you.

    The friendly fraud share, by vertical

    Not all chargebacks are the same. Reason-code analysis tells you whether the chargebacks are genuine fraud (you got hit by stolen cards), processing errors (something broke in your flow), or friendly fraud (the customer actually bought it and is disputing anyway). The vertical mix matters enormously.

    VerticalGenuine fraud %Processing error %Friendly fraud %
    B2B SaaS21%34%45%
    DTC apparel28%13%59%
    Subscription boxes11%18%71%
    Digital goods / courses14%6%80%
    Nutra / supplements24%11%65%
    CBD31%9%60%
    Peptides27%12%61%
    Telehealth / TRT18%15%67%
    Crypto onramp54%7%39%
    Travel / hospitality32%17%51%

    Friendly fraud dominates nearly every DTC vertical. The practical implication: if you are a 12-brand DTC operator staring at a 0.71% chargeback ratio and wondering what to do, your single biggest lever is representment, not fraud filters. The operators with the best outcomes in our dataset were not the ones with the best fraud tooling — they were the ones with the best dispute response process. See chargeback dispute playbook templates.

    Representment win rate is the under-rated signal

    We asked every respondent to share their 12-month representment win rate. Winning a representment means Visa/Mastercard reversed the chargeback in your favor, not that the issuer accepted your docs. Distribution:

    Win rate band% of respondentsMedian net chargeback ratio after representment
    < 20%26%0.63%
    20-40%31%0.48%
    40-60%24%0.34%
    60-80%14%0.21%
    > 80%5%0.13%

    The delta between "20% win rate" and "60% win rate" at the same gross chargeback rate is the difference between being on your acquirer's risk list and being off it. Operators who treat representment as a serious operations function — templated evidence packs, SLA-driven response times, delivery confirmation attached to every response — consistently came in at 50%+ win rates. Operators who ignored representment ran at sub-20% and ate the ratio. See the state of chargeback representment 2026.

    Refunds and chargebacks — not what you think

    Common advice: "refund generously and chargebacks go down." In our data, this is only partially true. Operators at the high end of refund rate (8%+) did have slightly lower chargeback ratios — but the net impact on the combined metric (refunds + chargebacks / volume) was often worse, and acquirers scored them more harshly on refund rate alone. The optimal zone was a refund rate of 3-5% combined with active representment.

    More importantly: acquirer risk teams look at refund rate and chargeback rate as separate signals. Having a low chargeback rate bought by a 10% refund rate does not "launder" the underwriting picture.

    Multi-brand portfolio aggregation

    A structural advantage of multi-brand portfolios: you can absorb a single-brand spike without tripping the portfolio-level ratio. Operators running 8+ brands on an aggregated descriptor structure (e.g., orchestration with a shared acquirer MID) had individual-brand chargeback ratios up to 1.8% without the portfolio itself crossing 0.6%. This is not a loophole — it is a real consequence of how ratios are computed. See chargeback ratios across sub-brands.

    The flip side: a portfolio-level chargeback spike caused by one brand can drag down the whole book. Orchestration helps manage this because a routing change can quickly move traffic from a spiking brand to a different acquirer before the ratio bleeds the portfolio.

    What to do with these benchmarks

    • Calculate your actual 30-day count-based chargeback ratio. If you have never done this on a strict count basis, do it now — dollar-based ratios can look flattering.
    • Look up your vertical in the table above. Position yourself in the distribution.
    • If you are above p75, start a representment program this week. See dispute playbook templates.
    • If you are above p90, your acquirer already knows. Proactive risk-mitigation communication buys goodwill.
    • If you are running multi-brand, map individual-brand rates against the portfolio rate to know which brand is actually dragging the math.

    Methodology, caveats, and how to cite

    Data window: April 2025 to March 2026. Respondent count: 410 operators. Total volume represented: $3.1B. Deduplicated against repeat respondents; aggregated to parent-operator level where respondents submitted multiple merchant reports.

    Caveats: our sample over-represents US operators (74%), ecom-native businesses (82%), and the high-risk verticals we specialize in understanding. Under-represented: enterprise retail, hospitality, and government/b2g sectors. Traditional brick-and-mortar card-present data is excluded entirely.

    Cite: "multiflow, Chargeback Ratio Benchmarks by Vertical 2026." Link: multi-flow.pro/blog/chargeback-ratio-benchmarks-by-vertical-2026/.

    Running one of the verticals above and want the detailed percentile curve for your specific sub-vertical? The apply page gets you an operator-grade review including where your portfolio falls inside the distribution.

    Found this useful? Share it X LinkedIn Reddit HN Email

    FAQ

    Why is VAMP the 2026 threshold, not VDMP?
    Visa migrated primary monitoring to VAMP in 2025. The 0.65% count-based threshold is stricter than VDMP's old 0.90%, and acquirer enforcement has shifted accordingly.
    Should I optimize for count ratio or dollar ratio?
    Count ratio. VAMP is count-based. Dollar-based numbers can mask a low-AOV chargeback problem.
    How is a "median" useful when my vertical has huge variance?
    It is a starting point. Pair it with p75/p90 to see the full distribution. If you are above p75, you are in the riskier half of your own vertical.
    Do subscription businesses actually have more friendly fraud?
    Yes, by a wide margin. 71% of subscription chargebacks in our dataset were friendly fraud codes. Billing descriptor clarity and cancellation flow UX are your biggest levers.
    What is the single best move to drop my chargeback ratio?
    Active representment. Going from a 20% to a 50% win rate typically drops net ratio by 30-40% in 60 days without any other change.
    Are these benchmarks valid for Europe?
    Partially. EU-weighted vertical medians are structurally 15-25% lower than US for most DTC categories. Our sample is US-heavy — treat these as US benchmarks.

    Running multiple brands?
    multiflow was built for this.

    The Operator Briefing

    Twice-monthly. No fluff.

    Processor shutdowns, reserve-hold playbooks, reconciliation lessons, and the merchant-account decisions that save operators six-figure years. Delivered to your inbox — never spam.

    No spam. Unsubscribe in one click.

    We use essential cookies · Privacy