Status: Methodology open for public review; data publication
scheduled for Q3 2026. Every provider named below has been or will be
contacted for formal comment before publishing results. We will not publish
numbers we can’t reproduce from a public sandbox or explicit provider consent.
Why we’re publishing this
Fintech teams picking a payout API today have no credible, reproducible,
free benchmark to work from. Vendor-produced comparisons predictably favour
the vendor publishing them. Independent reviews are either paywalled (analyst
firms) or out-of-date. We run payout flows in production ourselves, so we have
skin in the game — and motive to be transparent about methodology. This page
locks the methodology before any numbers are collected.
What we measure
- P50 / P95 / P99 latency — wall-clock time from API call to terminal status (SUCCESS / FAILED).
- Decline rate — failed / total, broken down by NPCI / RBI reason code.
- Webhook lag — time from terminal internal state → webhook delivered to our endpoint.
- Reconciliation accuracy — T+1 settlement-file ledger match vs API-reported status. Discrepancies are failures.
- Pricing transparency — qualitative score on how easy the pricing page makes it to compute true cost-per-transaction.
Sampling plan
- 10,000 real payouts per provider across a 30-day window.
- Stratified across 6 banks (HDFC, SBI, ICICI, Axis, Kotak, PNB).
- Amount buckets: ₹1–100, ₹100–1K, ₹1K–10K, ₹10K–1L, >1L.
- Time-of-day coverage: morning peak (10:00–12:00 IST), afternoon, evening peak (18:00–21:00 IST), overnight (00:00–06:00 IST).
Providers in scope
Every RBI-licensed or RBI-authorised payout API provider with publicly
documented API + sandbox access, as of 1 March 2026. If you are a provider
and want to be included or excluded, email
benchmark@nxtbanking.com by 30 April 2026.
Methodology integrity
- Clock sync via NTP; latency measured server-side at our edge, not in the browser.
- No synthetic “best time to test” — we accept the natural distribution of peak/off-peak.
- Results published per-provider with raw CSV of measurement timestamps.
- Methodology + raw code on GitHub (link will appear here at publication).
- 24-h embargo to each named provider before publication for factual-accuracy response.
Conflict-of-interest disclosure
NxtBanking operates its own payout API. We will publish our own numbers
under the same methodology, in the same report, with the same scrutiny. Our
goal is credibility with developer readers and journalists — not a vanity
benchmark.
Contribute
- Provider? Email benchmark@nxtbanking.com for sandbox credentials or opt-out.
- Journalist? The CSV + raw timestamps will be available under an embargo; email the same address to be on the list.
- Developer? Methodology feedback is welcome — file an issue on the GitHub repo (link pending).
*Last updated: 23 April 2026. Methodology frozen as of this date; any
subsequent changes will be ChangeLog’d inline below.*