Content authored by practitioners with experience at Amazon India, Intuit QuickBooks, and the Tata Group. Meet the team →
Indian subscription merchants and NBFC EMI books quietly absorb the cost of eNACH mandate rejections at approximately fifteen rupees plus eighteen percent GST per failed debit, compounded across two-cycle retry policies, and on a 38,000-mandate book at a 12 percent rejection rate the annual cost reaches roughly forty-one and a half lakh once delay-cost on the receivable is added. The headline rejection fee appears as a single consolidated line on the aggregator's monthly invoice with no per-batch, per-sponsor-bank, or per-return-code breakdown, so finance controllers cannot tune retry policy, identify rejection-trap sponsor banks, or hold the aggregator to the contracted fee schedule. The leakage is invisible because the four-layer cost stack — mandate-creation, first-attempt rejection, retry-rejection, successful-debit — is never assembled in one place.
Build a per-mandate registry keyed by mandate identifier with sponsor bank, destination bank, mandate creation date, status, and current attempt count. For every settlement batch, pull the eNACH return file and attribute every failed debit to its mandate and attempt number. Look up the NPCI return reason code and total the rejection fee plus GST per code, per sponsor bank, and per batch. Compute the first-attempt rejection rate and retry-cycle rejection rate by sponsor bank and rank against the cluster mean to surface rejection-trap candidates. Compute the days-past-due carry-cost on the receivable per failed first-attempt debit using marginal cost of funds and average ticket size. Reconcile the per-batch fee totals across the month against the aggregator's monthly fee invoice within a small rounding tolerance.
Per-mandate registry with sponsor and destination bank metadata. Aggregator settlement and return file ingestion pipeline. NPCI return reason code dictionary covering insufficient balance, account closed, signature mismatch, mandate inactive, technical reject. Per-debit rejection fee schedule by aggregator with GST eighteen percent overlay. Retry policy engine with code-aware first-cycle and T+3 retry rules. Marginal cost of funds parameter and average EMI ticket parameter for delay-cost computation. Monthly aggregator tax invoice reconciliation rules. Variance register feeding the four-layer eNACH cost dashboard.
A four-layer eNACH cost dashboard showing mandate-creation, first-attempt rejection, retry-rejection, and successful-debit costs with GST overlays, totalled per sponsor bank and per return reason code. A rejection-trap report ranking sponsor banks against the cluster mean with thirty-day rolling first-attempt rejection rates. A code-aware retry policy recommendation engine surfacing return codes where retry economics are negative. A monthly delay-cost computation showing carry-cost on the receivable book attributable to rejected first-attempt debits. A monthly fee reconciliation between per-batch totals and the aggregator tax invoice for ITC alignment under Rule 36(4).
A non-banking financial company in Hyderabad runs a personal-loan EMI book with 38,000 active eNACH mandates. The treasury team books the eNACH operational cost as a single monthly accrual against the aggregator’s invoice and moves on. The actual cost picture is four layers deep. The first layer is the small mandate-creation charge the sponsor bank levies when a customer signs up. The second is the rejection fee — approximately ₹15 plus 18 percent GST every time a scheduled debit fails. The third is the retry-cycle cost — a T+3 retry and a T+5 retry, each a fresh fee event when they fail. The fourth is the days-past-due carry-cost the loan book holds while the retry cycle runs. At a 12 percent first-attempt rejection rate, those four layers add up to roughly ₹41.5 lakh a year. The aggregator’s invoice never breaks any of it down.
This article is the per-batch reconciliation framework that surfaces the four-layer cost stack, classifies every failure back to its NPCI return reason code, and detects the sponsor banks whose realised rejection rates are structurally elevated against the cluster mean. The discipline pays for itself inside one quarter on any subscription book above 20,000 active mandates.
Quick reference: the four-layer eNACH cost stack
| Cost layer | Typical rate | When it triggers | Reconciliation source |
|---|---|---|---|
| Mandate-creation charge | One-time, varies by sponsor bank | At eNACH mandate setup | Aggregator setup fee statement |
| First-attempt rejection fee | Approximately ₹15 plus 18% GST per failed debit | Scheduled debit fails on due date | eNACH return file with NPCI code |
| Retry-cycle rejection fee | Approximately ₹15 plus 18% GST per failed retry | T+3 and T+5 retry debits fail | eNACH return file with attempt number |
| Successful-debit cost | Aggregator’s per-debit fee, often passed through | Scheduled debit clears | Aggregator settlement file |
| Delay-cost on receivable | Marginal cost of funds × days-past-due × outstanding | Between failed debit and eventual success | Loan management system days-past-due flag |
| GST on every fee component | 18% on each layer | On every invoiced fee | Aggregator monthly tax invoice |
Why a flat monthly accrual hides the rejection-fee leakage
A typical NBFC books “eNACH operating cost” as a single accrual line, often a fixed-rupee number negotiated annually with the aggregator. The line is calibrated to a budgeted rejection rate that was set at the start of the financial year. When realised rejection rates drift above budget — and they almost always do on personal-loan and small-ticket EMI books — the variance lands in a generic “collections operations” cost centre that no one audits at line-item level. The aggregator’s invoice arrives once a month with a single mandate-rejection fee total and no per-batch breakdown. Finance signs off, the variance is absorbed, and the leakage is structural.
A per-batch reconciliation flips the discipline. Every settlement batch produces a return file with every failed debit, every retry attempt number, and every return reason code. Aggregating those across a month gives the merchant a per-sponsor-bank, per-code, per-cycle fee picture that the aggregator’s invoice will not. The two totals should reconcile within a rounding tolerance. They almost never do on the first pass, and the variance is the leakage.
What is the eNACH mandate-rejection fee actually paying for?
The fee is a return-handling charge on the NACH rail. When a sponsor bank submits a debit instruction in the daily NACH file and the destination bank rejects it, the destination bank issues a return message with one of the published NPCI return reason codes. That return message has to be settled, reconciled, and reported back to the customer through the aggregator’s dispute desk. The sponsor bank charges the aggregator for the return handling, the aggregator passes it through to the merchant, and the merchant sees ₹15 plus 18 percent GST on the line. The fee does not depend on the rupee value of the failed debit — a failed ₹500 EMI and a failed ₹50,000 EMI both attract the same ₹15 plus GST.
That flat structure is what makes the fee bite hardest on small-ticket subscription books. The merchant who sells a ₹149 monthly OTT subscription pays the same ₹15 plus GST on a failed debit as the NBFC collecting a ₹50,000 monthly EMI. As a proportion of the underlying invoice, the small-ticket merchant absorbs an 11.8 percent fee on every failed debit. As a proportion of the EMI value, the NBFC absorbs 0.035 percent. The reconciliation discipline is identical. The economic impact is not.
How do retry cycles compound the rejection fee?
A typical EMI retry policy is a first attempt on the due date, a T+3 retry, and a T+5 second retry. Each retry is a fresh NACH debit message and a fresh fee event if it fails. The compounding depends on the return reason code distribution. Failures caused by insufficient balance have a high retry-recovery rate because the customer’s salary credit usually arrives within seven days. Failures caused by mandate inactive or account closed have near-zero retry recovery — every retry is a guaranteed fee burn against a debit that will not clear.
The reconciliation discipline is code-aware retry policy. For codes with positive retry economics, the full two-cycle retry runs. For codes with negative retry economics, the retry policy is suspended after the first attempt and the case is routed to the collections team for direct customer outreach. A merchant who does not classify codes runs the same two-cycle retry on every failure, pays the rejection fee twice on every dead mandate, and writes off the customer relationship anyway. A merchant who classifies codes pays the fee once on dead mandates, runs the retry only where it will land, and converts the ₹15-times-three saving into a real operating-cost reduction.
Worked example: the 38,000-mandate NBFC EMI book
The Hyderabad NBFC runs the following monthly numbers. Active eNACH mandates: 38,000. First-attempt rejection rate: 12 percent. First-cycle failures: 4,560. Rejection fee per failed debit: approximately ₹15 plus 18 percent GST, which equals ₹17.70. First-cycle monthly rejection-fee burden: 4,560 multiplied by ₹17.70, which equals ₹80,712.
The retry policy is a T+3 retry on all first-cycle failures. Sixty percent of the first-cycle failures recover on the T+3 retry — these are predominantly the insufficient-balance codes where the customer’s salary credit has now arrived. Forty percent persist into a second failure. Second-cycle failure count: 1,824. Second-cycle monthly rejection-fee burden: 1,824 multiplied by ₹17.70, which equals ₹32,285.
Combined monthly mandate-rejection fee: ₹1,12,997, rounded to ₹1.13 lakh. Annual mandate-rejection fee on the rejection layer alone: ₹13.56 lakh.
Now the delay-cost layer. The average EMI ticket is ₹1,200. The 4,560 first-cycle failures hold ₹54.7 lakh of receivable in delayed status. The marginal cost of funds for the NBFC is 9 percent annualised, which equals roughly 0.0247 percent per day. The average delay between the first failed debit and the eventual successful retry, weighted across recovered and unrecovered cases, is 7 days. Daily carry-cost on the delayed receivable: ₹54.7 lakh multiplied by 0.0247 percent, which equals ₹13,510. Across a 7-day average delay: ₹13,510 multiplied by 7, which equals ₹94,570, applied across the cycle of first failure to retry resolution and projected across a full month of overlapping cycles produces a monthly delay-cost in the band of ₹2.34 lakh.
Total monthly four-layer cost: ₹1.13 lakh rejection fee plus ₹2.34 lakh delay-cost, equals ₹3.47 lakh. Annual: ₹41.5 lakh.
The CFO who books eNACH operating cost as a flat ₹1 lakh monthly accrual is under-reserved by 3.5 times. The leakage is the difference, and it is invisible until the four-layer dashboard surfaces it.
Model the eNACH cost stack against UPI AutoPay for your subscription book
For sub-₹2,500 recurring tickets, UPI AutoPay carries zero MDR on the bank-account debit while eNACH carries the ₹15-plus-GST rejection-fee stack. The comparator surfaces the cross-over ticket size where the rail choice matters most.
Open the tool →How do you detect a rejection-trap sponsor bank?
A rejection-trap sponsor bank is one where the realised first-attempt rejection rate is structurally elevated against the cluster mean, controlling for customer profile and mandate vintage. The detection lives in the per-batch return file. For every settlement batch, the reconciliation engine attributes every failed debit to its sponsor bank and computes the rejection rate by sponsor for that batch. Averaging across thirty days and ranking surfaces the candidates.
A book where the cluster-mean first-attempt rejection rate is 10 percent and one sponsor bank consistently runs at 18 percent on the same destination-bank mix is a rejection-trap. The cause is almost always one of three. The first is a stricter destination-bank acceptance check on that sponsor’s mandate-creation messaging — sometimes the sponsor’s mandate file format does not pass the destination’s validation cleanly, and the mandate is accepted on paper but rejected at the first debit. The second is a daily file-submission delay at the sponsor, where the file misses the destination bank’s cut-off and is treated as a same-day reject. The third is a higher rate of legacy paper-mandate conversions on that sponsor that fail signature validation at the destination.
Once detected, the procurement team raises the variance with the aggregator and routes new mandate creation to a different sponsor for the affected destination banks. The reconciliation continues to monitor whether the cluster mean for that sponsor reverts. The recurring add-on layer of the contract — explored in the cluster sibling on recurring add-on and eNACH mandate-rejection fees — is renegotiated against the documented rejection-rate variance.
Reconciliation discipline: per-batch attribution back to the NPCI return code
The per-batch reconciliation is the spine of the discipline. Every NACH settlement batch produces a return file. The fields the reconciliation engine needs from the return file are mandate identifier, original debit amount, sponsor bank, destination bank, attempt number, return reason code, return date, and the per-debit fee. The fields from the per-mandate registry are mandate creation date, status, contract reference, and customer identifier. The join key is the mandate identifier.
The output of the join is one row per failed debit with all four cost layers attributed. Aggregating by NPCI return reason code surfaces the failure distribution. Aggregating by sponsor bank surfaces the rejection-trap candidates. Aggregating by retry-cycle attempt number surfaces the retry-policy effectiveness. Aggregating by destination bank surfaces customer-side patterns — a destination bank where the rejection rate is elevated across all sponsors is signalling a customer-base problem, not a sponsor problem.
The monthly fee invoice from the aggregator should reconcile against the per-batch fee totals for the month within a small rounding tolerance. Any variance above the tolerance is leakage — either the aggregator is billing rejection fees the return file does not document, or the rejection fee per debit has drifted above the contracted rate without notice. Both have happened. Both are recovered through the dispute desk within the platform’s contractual variance window. The same reconciliation discipline applies to NACH return-code economics covered in the cluster sibling NACH return codes: full taxonomy and reconciliation, and the rail-choice question for small-ticket recurring books is the subject of UPI AutoPay vs eNACH for ₹149 to ₹2,499 subscription tickets.
What does the GST 18 percent overlay do to the picture?
The aggregator charges 18 percent GST on the rejection fee, on the mandate-creation charge, and on the per-debit successful-debit cost. The GST is claimable as input tax credit for the NBFC under Rule 36(4) of the CGST Rules, conditional on the aggregator’s monthly tax invoice flowing through to GSTR-2B without variance. The reconciliation issue is the timing mismatch. The per-batch settlement file books the GST daily. The aggregator’s monthly tax invoice consolidates the GST month-end. The GSTR-2B then reflects the aggregator’s invoice. The merchant’s input-tax-credit claim in GSTR-3B is governed by the GSTR-2B amount, not the per-batch sum.
When the per-batch GST sum and the monthly tax invoice GST diverge — they often do by 1 to 2 percent because of rounding and batch-cut-off treatment — the merchant either over-claims ITC on the per-batch number and gets a notice, or under-claims ITC on the monthly invoice and leaves money on the table. The fix is a structured GST reconciliation between the settlement-file GST totals and the aggregator’s monthly tax invoice, run before the GSTR-3B filing window closes. The same reconciliation discipline applies across the broader payment-gateway fee stack — see Payment gateway reconciliation for the full per-statement framework.
What changes for subscription merchants under the broader fee-recon programme?
Subscription merchants need to assemble the full picture across all rails the customer uses, not just eNACH. A customer on card-on-file pays base MDR plus the recurring add-on. A customer on eNACH pays the four-layer stack documented above. A customer on UPI AutoPay pays zero MDR on the bank-account debit but may attract a small per-mandate platform fee from the aggregator. The reconciliation engine has to compute the effective per-customer collection cost across rails and surface the rail-mix that minimises total cost for the segment. For ₹149 to ₹2,499 monthly tickets, UPI AutoPay almost always wins. For ticket sizes where the eNACH four-layer stack is unavoidable, the rejection-trap and code-aware-retry disciplines are the levers. The full software-side framework lives at Reconciliation software India.
Continue reading: merchant-fees cluster
- Recurring add-on and eNACH mandate-rejection fees: stacked costs — the contract-side view of the same fee stack, focused on the Razorpay 0.99 percent recurring add-on
- UPI AutoPay vs eNACH for ₹149 to ₹2,499 subscription tickets — the rail-choice decision for small-ticket recurring books
- NACH return codes: full taxonomy and reconciliation — the NPCI return reason code dictionary that drives code-aware retry policy
- Merchant fees cluster hub — full cluster index across 25-plus articles on MDR, gateway platform fees, leakage detection
- Payment gateway reconciliation money page — the per-statement reconciliation framework across all gateways and rails
External authority
The operative framework, return reason code taxonomy, and product details for NACH and eNACH live with NPCI. The NPCI NACH product overview is the authoritative source for the mandate life cycle, settlement messaging, and return code dictionary referenced throughout this reconciliation framework.