State of SMB Accounts Receivable 2026 — payment delays, DSO benchmarks, and what the data shows

The Atradius Payment Practices Barometer, published annually across major markets, is one of the most-cited datasets on B2B payment behavior. In recent US editions, Atradius reports that around 50 percent of B2B invoice value is overdue at any point in time, with small businesses disproportionately affected because they tend to have lower pricing power with large customers and thinner reserves to absorb slow pay. The NACM Credit Managers' Index, which surveys credit professionals monthly, shows receivables-beyond-terms as a persistent concern even in years with otherwise healthy economic conditions.

The Federal Reserve's Small Business Credit Survey, which covers non-employer and employer firms with fewer than 500 employees, consistently ranks cash flow and revenue volatility as the top financial challenge. In the 2023 edition, roughly 43 percent of firms reported financial challenges in the prior year, with cash flow as the most commonly cited category. For service businesses with irregular project cycles — HVAC, roofing, plumbing, general contracting — the cash-flow challenge is structurally linked to how quickly invoices convert to received payment, not to whether the underlying business was profitable.

Payment delay patterns are not uniform across industries. Construction and professional services carry the highest DSO of any service-sector category. The US Bureau of Labor Statistics and Dun & Bradstreet trade-credit data both show construction receivables aging at 60 to 90 days as a norm rather than an exception. Residential trades (HVAC, plumbing, electrical) trend faster: the customer relationship is B2C or small-B2B, the invoice amounts are smaller, and the follow-up friction is lower. The typical DSO for residential trades runs 30 to 45 days when no active follow-up is occurring; businesses with structured day-three follow-up routinely report DSO below 20 days on their own accounting.

Published DSO benchmarks for SMBs are harder to find than enterprise benchmarks because smaller businesses are underrepresented in public filings and credit surveys. The ranges below are drawn from commercial credit industry sources and should be read as central tendencies, not hard floors or ceilings.

Residential trades (HVAC, plumbing, electrical, roofing, landscaping, painting, pest control, pool service): DSO typically 20 to 50 days. The wide range reflects follow-up behavior more than customer type. Businesses that make proactive calls in the first 10 days consistently report DSO at the low end; businesses that rely on email reminders alone cluster at the high end.

Professional services (law, accounting, consulting, marketing agencies, design): DSO typically 35 to 65 days. Client relationships are longer-term, invoicing is often on retainer or milestone, and the follow-up call feels more fraught to principals than in trades contexts. The discomfort of making the call is higher in professional services, which is one reason DSO is higher.

General contracting and construction: DSO typically 60 to 100 days. Owner-to-GC payment delays cascade into GC-to-sub delays. Retainage rules in most states allow owners to withhold up to 10 percent of contract value until substantial completion, structurally extending DSO on the largest piece of each project's AR.

Healthcare (dental, physical therapy, chiropractic, occupational therapy): DSO varies widely based on payer mix. Patient-pay balances mirror residential trades behavior; insurance-billed amounts can stretch 90 to 180 days waiting for adjudication. The data does not cleanly separate these pools, which makes published DSO benchmarks for healthcare largely uninformative without knowing payer mix.

B2B SaaS and digital services: DSO typically 20 to 40 days. Invoice amounts are often smaller, billing is often automated through Stripe or similar, and dunning automation is more common than in trade businesses. The average is lower, but the businesses that fall off the low end often do so because of complex multi-stakeholder enterprise deals where invoice approval requires three sign-offs.

The NACM and commercial credit industry publish write-off benchmarks for B2B trade credit at roughly 1 to 2 percent of net sales as a normal range for well-managed receivables. Businesses without active follow-up processes frequently report write-off rates of 3 to 5 percent, with some trade categories running higher.

The predictors of write-off outcomes, based on the aggregate credit-management literature, are: age of the receivable at first contact, method of first contact, and number of follow-up attempts made before day 30. Age is the single largest predictor. An invoice contacted at day 5 has a materially different outcome distribution than the same invoice contacted at day 35. This is the core empirical justification for day-3 calling as a practice.

Method matters too. Phone contact substantially outperforms email contact in closing overdue balances in the under-90-day window, based on aggregate reporting from credit managers and collection industry sources. The mechanism is conversational: a phone call creates a commitment device that email cannot. When a customer says 'I'll send payment Thursday' on a call, they are more likely to send it Thursday than if they read the same request in an email they marked for later.

Number of attempts is less predictive than age and method, but the data supports at least three contact attempts before writing off, across multiple contact types (email, phone, postal notice). Most businesses that write off at or before day 90 report having made fewer than two contact attempts, suggesting the write-off was a giving-up decision rather than an inability-to-collect outcome.

The regulatory environment for businesses that automate invoice follow-up — including AI voice calls — changed materially in 2024 and continues to tighten in 2026. The key developments:

The FCC's February 2024 Declaratory Ruling placed AI-generated voice calls inside the definition of 'artificial or prerecorded voice' under TCPA. This closed the argument that AI-generated voice was categorically different from traditional robocalls. As of 2024, every AI voice call is a TCPA call: strict-liability framing, $500 per inadvertent violation, $1,500 per willful violation, private right of action, and state enforcement authority.

The FCC's July 2024 Notice of Proposed Rulemaking on AI-generated calls proposed specific consent requirements and in-call disclosure obligations tailored to AI-generated calls. The NPRM was not finalized as of this writing, but the direction of travel is clear: more disclosure requirements, tighter consent standards, and explicit identification of AI callers. Businesses deploying AI voice agents for invoice follow-up in 2026 should expect final rules that look broadly similar to the NPRM's proposed framework.

State-level regulation continues to diverge from the federal floor. California's Rosenthal Act extends FDCPA-style restrictions to some first-party callers, which is unusual compared to other states. Florida's FCCPA caps consumer call windows at 8 PM local time, an hour earlier than TCPA's 9 PM floor. New York City's DCWP imposes additional consumer-protection requirements for invoice-related contacts with NYC consumers. Businesses operating across multiple states need per-state compliance logic, not a single national call window.

First-party callers remain outside FDCPA's scope. The FTC's FDCPA interpretation has consistently held that a creditor calling its own customer about its own debt falls outside the statute. This is a meaningful structural advantage for service businesses doing their own follow-up: the call can be framed as customer service, can acknowledge specific transaction details, and is not subject to the mini-Miranda disclosure requirement that applies to third-party collectors. That framing advantage erodes at day 90 when customers shift from 'I need to pay this' to 'someone is trying to collect from me.'

Three practical observations follow from the data above.

First, the write-off risk is front-loaded. Most of the recoverable value on any overdue invoice is in the first 30 days. Businesses that do not make a first-party contact in that window are not just delaying recovery — they are foregoing most of it. A 2-percent write-off rate on $500,000 in annual revenue is $10,000 per year. A 1-percent write-off rate is $5,000. The gap between 'structured day-3 follow-up' and 'no structured follow-up' is measurable in thousands of dollars per year for a business doing $500k in revenue, and the dollar figure scales with revenue.

Second, the regulatory environment makes automated follow-up more valuable and more technically demanding at the same time. More valuable because the manual call takes time a business owner does not have. More technically demanding because TCPA strict-liability framing means a compliance mistake in automated calling is an expensive mistake. The product that solves this correctly enforces compliance at the architecture level — hardcoded disclosures, deterministic call windows, database-sourced amounts — rather than relying on prompt instructions to a language model.

Third, DSO is a lagging indicator of follow-up behavior, not a leading indicator of customer quality. Businesses with high DSO often attribute it to 'slow customers' when the data suggests the proximate cause is no structured follow-up. The same customer pool, contacted three days past due rather than thirty, resolves at a materially different rate. DSO improvement from 50 days to 25 days is achievable for most residential-trade businesses with structured day-3 follow-up; it is a process change, not a customer-quality change.

This document draws on four primary public sources: the Atradius Payment Practices Barometer (US edition, most recent available), the Federal Reserve Small Business Credit Survey (2023 edition), the NACM Credit Managers' Index, and commercial credit industry benchmarks from Dun & Bradstreet and similar sources. All are aggregate datasets with their own sampling and methodology limitations.

SMB-specific data is underrepresented in most financial surveys because small businesses are underrepresented in public filings. The DSO ranges cited are drawn from credit industry reporting and should be treated as order-of-magnitude guidance rather than precise benchmarks. Industry-specific ranges vary by 10 to 20 percentage points from the figures above; a specific business's experience can vary further.

Syntharra has not yet accumulated a statistically representative dataset of recovery outcomes across customer businesses and invoice ages sufficient to publish Syntharra-specific benchmarks with confidence intervals. This document does not contain Syntharra-specific recovery claims. When Syntharra-specific data reaches a publishable threshold, a separate data note will be published with the methodology, sample characteristics, and limitations.

This document will be updated annually. The regulatory section reflects the state as of May 2026 and may be outdated for events occurring after that date.