How do MCA funders classify deposits on bank statements in 2026?

Full answer

Why deposit classification matters in 2026. Bank statements show every credit transaction, but not all credits represent business revenue. A $50K deposit might be a transfer from a sister account (not revenue), a loan proceed (not revenue), a tax refund (not revenue), or a genuine customer payment (revenue). Funders classify each deposit to extract true qualifying revenue — the number that drives advance sizing (typically 80-130% of average monthly qualifying revenue). Misclassification can mean a merchant gets $30K approved instead of $50K, or vice versa.

Classification bucket 1: qualifying revenue (counted). Credits that count toward revenue: (a) customer ACH payments with B2B descriptors, (b) processor settlements (Stripe, Square, Toast, Clover, Worldpay batches), (c) check deposits from named customers (verified against invoices when available), (d) wire transfers from B2B customers, (e) Cash App / Venmo / Zelle business inflows tied to invoiced customers. These flow directly into the funder's monthly revenue calculation.

Classification bucket 2: inter-account transfers (excluded). Credits explicitly flagged as transfers: (a) 'TRANSFER FROM' / 'ZELLE FROM SELF' / 'INTERNAL TRANSFER' descriptors, (b) round-trip patterns (same amount debited from one account and credited to another within 1-2 days), (c) owner-account transfers (savings to checking, sister-LLC transfers). Funders exclude these because they don't represent new revenue — counting them would double-count cash and inflate the advance.

Classification bucket 3: loan and MCA proceeds (excluded). Credits identifiable as financing inflows: (a) prior MCA funding deposits (matched against known funder ACH IDs in DataMerch and Decisionlogic databases), (b) SBA loan disbursements, (c) line-of-credit draws (named bank or fintech), (d) crowdfunding payouts (Kickstarter, GoFundMe), (e) factoring company advances. Funders strip these because they represent debt inflows that will be repaid, not earned revenue.

Classification bucket 4: refunds and chargebacks (netted out). Refund-related transactions: (a) processor chargeback reversals (counted as negative revenue), (b) customer refund returns (netted against the original sale month), (c) ACH return reversals. Funders typically net these against gross revenue to produce 'net qualifying revenue'. High refund/chargeback rates (>3% of gross) trigger underwriter review and may reduce advance amount or decline application entirely.

Classification bucket 5: tax refunds and owner injections (excluded). One-time inflows: (a) IRS / state tax refunds, (b) owner capital contributions (cash injections from personal account), (c) insurance claim payouts, (d) grant disbursements (PPP, EIDL, state grants in some funder policies), (e) settlement proceeds. These are one-time events not representative of ongoing business revenue, so funders exclude to avoid inflating sustainable revenue estimates.

Classification bucket 6: processor settlements (counted with adjustments). Card processor batch deposits (Stripe, Square, Toast, etc.) are counted as revenue but with adjustments: (a) gross vs net settlement awareness — Stripe deposits net of fees, Square deposits gross with separate fee debit, etc. (b) chargeback offsets within batch. (c) processor reserve holdbacks (some processors hold 5-10% as rolling reserve — funders adjust). Sophisticated funders gross-up net deposits to compare apples-to-apples across processors.

Classification methodology 2026. Modern MCA funders use a 4-step classification pipeline: (1) descriptor pattern matching — regex and keyword lookup against descriptor library (10K+ patterns for processors, lenders, payment apps). (2) Amount and frequency analysis — recurring round-number deposits flagged as potential transfers. (3) Counterparty matching — verify deposit sender against business name, owner name, sister-LLC names from secretary-of-state lookups. (4) ML classification model — trained on labeled deposit data to handle edge cases (novel processors, ambiguous descriptors). Output is a labeled deposit with confidence score.

Software vendors and tooling 2026. Decisionlogic and Ocrolus both ship deposit classification engines as part of their analytics layer. Plaid provides raw transaction categorization (Plaid Categories taxonomy) but funders typically apply MCA-specific overrides on top. Sigma (by Plaid) and Heron Data are emerging specialty vendors focused on cash flow underwriting classification. Funders building in-house (Credibly, Forward Financing, OnDeck) layer proprietary descriptor libraries on top of vendor outputs.

Impact on advance amount 2026. Classification accuracy directly translates to advance dollars. Example: merchant reports $80K/mo in gross deposits, but classification breaks down to $50K qualifying revenue + $15K transfers + $10K loan proceeds + $5K tax refund. Approved advance: 100% of qualifying revenue = $50K (not $80K). Merchants who structure operations to maximize qualifying revenue in the business operating account (vs scattered across personal/sister-LLC accounts) get materially larger approvals.

Common merchant mistakes that hurt classification 2026. (a) Running payroll out of business account and re-injecting (looks like loan inflow). (b) Owner Zelle-ing personal cash into business account (excluded as transfer). (c) Multiple checking accounts with frequent transfers (transfers excluded, fragmented revenue picture). (d) Mixing personal and business deposits in one account (personal credits stripped out). (e) Using Cash App / Venmo personal accounts for business (often not even classified as business revenue). Cleaning up account structure 30-90 days before applying improves classification outcomes.

Bottom line. MCA funders in 2026 classify bank deposits into 6 buckets: qualifying revenue (counted), inter-account transfers (excluded), loan/MCA proceeds (excluded), refunds and chargebacks (netted), tax refunds and owner injections (excluded), and processor settlements (counted with adjustments). Classification accuracy directly impacts approved advance amount — misclassification can swing funding by 20-40%. Funders use descriptor pattern matching, frequency analysis, counterparty verification, and ML models via Decisionlogic, Ocrolus, Plaid, Sigma, and Heron Data tooling. Merchants who clean up account structure (single business operating account, minimal transfers, no owner injections in trailing 90 days) get materially larger approvals than those with fragmented or commingled deposit patterns.

Related questions

• MCA funder bank statement software typical
• MCA funder bank statement revenue vs deposit detailed
• MCA funder bank statement trended analysis detailed
• MCA funder decisioning engine typical process