Content authored by practitioners with experience at Amazon India, Intuit QuickBooks, and the Tata Group. Meet the team →
Indian finance teams that suspect revenue leakage rarely have a structured recovery program. Recovery happens ad-hoc — disputes filed when someone notices, escalations raised when a customer is otherwise contacted, write-offs taken at quarter-end. The result is that disputable leakage ages out, structural leakage gets accepted by default, and the audit committee cannot tell whether the leakage trend is improving. A formal recovery playbook with per-class owners, SLA library, Discovered Money register, and quarterly cycle turns recovery from project work into a standing operational capability.
Map each of the seven leakage classes to a named owner, a regulator or contractual anchor, a dispute window or recovery action, a standard SLA, and an escalation route. Operate a Discovered Money register that quantifies recoverable leakage by class with stuck / at-risk / recoverable / recovered / structurally lost status. Run a four-week baseline measurement to size each class. Publish dispute-template and escalation-route libraries. Operate a monthly recovery review and a quarterly audit-committee cycle that reports rupees recovered, rupees in pipeline, rupees structurally lost. Integrate with existing reconciliation engine capability to scale residual handling.
Seven-class owner matrix with named individuals and back-up. Regulator-and-contractual anchor table by class with operative reference. SLA library per class with dispute window, escalation trigger, recovery action. Dispute-template library per class with regulator-aligned language. Discovered Money register with per-class rows, status field, and recovery-state machine. Monthly recovery review cycle. Quarterly audit-committee pack template. Annual playbook review covering owner rotation, SLA tightening, and template updates.
A monthly Discovered Money register decomposing total recoverable leakage by class with recovery probability bands. A weekly dispute-and-escalation operating queue routed to owners. A quarterly audit-committee leakage pack with rupees recovered, rupees in pipeline, rupees structurally lost, and trend. An annual playbook review report covering owner-matrix changes, SLA tightening, recovery-rate improvement, and infrastructure investment recommendations.
A CFO at a Mumbai mid-market business with ₹180 crore of revenue sits with the board chair after a friend’s CFO mentions recovering ₹3.2 crore from a structured leakage program. The board chair asks: how much are we losing? The honest answer is: we don’t know, because we’ve never measured. The next question is harder: what would it take to find out? The answer to that is the subject of this playbook — a six-to-ten-week deployment that takes the business from “we suspect we leak” to “we recover ₹X crore per year and we know which class drives most of it.”
This is the playbook. Step-by-step, with owners, SLAs, regulator anchors, and the Discovered Money register that holds the program together. The seven-class framing is from the Seven Classes guide; this article tells you how to operate it.
Quick reference: the recovery operating model
| Phase | Duration | Primary output | Owner |
|---|---|---|---|
| Phase 1 — Baseline measurement | Weeks 1-4 | Sized leakage per class with confidence band | CFO with class owners |
| Phase 2 — Owner matrix and SLA library | Weeks 5-6 | Published per-class ownership and recovery SLAs | CFO with controllership |
| Phase 3 — Discovered Money register | Weeks 7-10 | Operational register feeding monthly review | Controllership with reconciliation lead |
| Phase 4 — First quarterly audit-committee cycle | Week 12 onwards | Board-track leakage pack | CFO |
| Phase 5 — Quarter-over-quarter improvement | Continuous | Recovery-rate compounding | Class owners |
Phase 1 — Baseline measurement across the seven classes
A four-week effort to size each class with enough confidence to set realistic recovery targets. The output is a one-page summary per class with current rupee leakage estimate, primary driver, and quick-win recovery candidates.
For fee-deduction class. Pull six months of platform settlement files and bank-fee invoices. For each, compute expected fee at contracted rate against actual fee. Aggregate by aggregator and channel. The total over six months annualised is the current rupee leakage. Time required: 8-12 finance-team hours plus engineering support for the data pulls.
For tax-deduction class. Pull books TDS receivable position and aggregate Form 26AS plus Form 168 for the same period. The gap is the standing leakage. Time required: 6-10 finance-team hours.
For discount class. Pull all invoice lines with discount applied over six months. Cross-check against the contracted discount slab table per customer. Anything beyond the slab is unauthorised. Time required: 4-8 finance-team hours.
For rounding class. Pull a 30-day per-transaction settlement sample. Compute per-transaction expected net to the paise and compare to actual credit. The systematic per-transaction delta times monthly volume times 12 is the annual rounding leakage. Time required: 2-4 finance-team hours.
For short-settlement class. Pull every invoice marked settled in the last 90 days. Check cash-allocation against invoice gross. Sum the residuals closed without a credit note. The 90-day total annualised is the standing leakage. Time required: 6-10 finance-team hours.
For penalty-and-interest class. Pull the bank statement for the period and identify all interest-and-charge debits (Section 416 interest on late TDS, Section 50 interest on Rule 36(4) or Rule 37 reversals, NACH bounce charges). Classify each as avoidable or unavoidable. Time required: 4-6 finance-team hours.
For unexplained-variance class. Pull every JV posted to “reconciliation adjustment”, “variance write-off”, or similar generic accounts over six months. Sum the rupee value. This is the floor of unexplained-variance leakage. Time required: 2-4 finance-team hours.
End of Phase 1: a CFO has a defensible per-class rupee estimate and a quick-win candidate list per class.
Phase 2 — The owner matrix and SLA library
A two-week effort to formalise ownership and operating discipline. The output is a published owner matrix that every class has named, with backup, and an SLA library that every class has a dispute window, escalation trigger, and recovery action documented.
Owner matrix. Fee-deduction: AR controller for platform fees, treasury controller for bank fees. Tax-deduction: tax controller (TDS desk). Discount: AR controller plus commercial pricing lead. Rounding: AR controller plus reconciliation engine administrator. Short-settlement: AR controller plus key-account lead for top customers. Penalty-and-interest: tax controller for statutory, AR for NACH. Unexplained-variance: controllership with hard 7-day class-assignment rule.
SLA library. For fee-deduction: dispute window 60-90 days per platform contract, escalation trigger 30 days from filing, recovery target 70-85% within two quarters. For tax-deduction: 30/60/90/180-day deductor ageing per the TDS leakage article. For discount: pre-invoice authorisation gate with retrospective escalation within 30 days of identification. For short-settlement: residual-recovery queue with 60-day customer follow-up cycle. For penalty-and-interest: calendar-driven statutory clock for TDS and GST, NACH recharge within 30 days of bounce identification. For unexplained-variance: 7-day class-assignment SLA with weekly review.
End of Phase 2: every class has an owner, a window, and a written recovery action.
Phase 3 — The Discovered Money register
A four-week operationalisation of the recovery state machine. The output is a live register that quantifies recoverable leakage by class at any point in time and routes it to the right owner.
Register schema. Class. Originating event (invoice, settlement, deduction, dispute). Regulator anchor (Section 393, Rule 36(4), Section 50, contract clause). Rupee value. Date of identification. Recovery owner. SLA window. Current step (logged, dispute filed, escalation raised, recovered, structurally lost). Recovery status (stuck — no dispute yet filed; at-risk — filed but uncertain; recoverable — accepted by counterparty pending settlement; recovered — cash received; structurally lost — no recovery path).
Daily operation for active classes. Fee-deduction and tax-deduction rows updated daily as platform settlements arrive and Form 26AS / Form 168 pulls refresh. ITC leakage rows updated daily as GSTR-2B reflections come through.
Weekly operation for batch classes. NACH bounce-recharge rows updated weekly as the bank statement is reconciled. OEM short-pay rows updated weekly as the cash settlements process.
Monthly review. Class-wise recoverable balance, recovery rate quarter-to-date, ageing trend, top three escalations.
End of Phase 3: the register is operational and produces the weekly operating queue and the monthly review pack.
Phase 4 — The quarterly audit-committee cycle
Twelve weeks into the deployment, the first audit-committee cycle. The output is a four-to-six-page pack that the board chair can read in fifteen minutes and that the audit-committee chair can question deeply.
Standard pack contents. Page one: total recoverable leakage by class at start of quarter, rupees recovered in quarter, rupees in pipeline at end of quarter, rupees moved to structural-loss in quarter. Page two: per-class drill-down with top three counterparties or events. Page three: SLA-adherence statistics and exception report. Page four: recovery-rate trend versus baseline. Page five: working-capital leakage overlay (per the Working capital leakage article). Page six: recommendations for SLA tightening, owner reassignment, or infrastructure investment.
End of Phase 4: leakage recovery is a permanent agenda item on the audit committee.
Size your leakage baseline before the deployment
Plug in revenue, receivable base, and rough class mix to project leakage by class and the recovery upside under a structured playbook.
Open the Revenue Leakage Calculator →Phase 5 — Quarter-over-quarter improvement
The compounding phase. Recovery rate improves as the dispute-template library tightens, the SLA library gets refined to actual counterparty behaviour, the owner matrix gets refined with rotation and back-up coverage, and the reconciliation engine residual queue shrinks. A typical deployment moves from year-one recovery of 40-55% of standing leakage to year-three recovery of 70-80% as the operating discipline matures.
The compounding also flows through the working-capital leakage line. As days-recon-delay falls (from 14-22 days at baseline to 4-7 days at maturity), the working-capital overlay leakage falls correspondingly. On a ₹140 crore receivable base, cycle reduction from 18 days to 6 days releases roughly ₹46 lakh of cost-of-capital saving per year (see the working capital article for the math).
Integrating the playbook with reconciliation infrastructure
The playbook is process; the engine is technical capability. A reconciliation engine operating on the 51% to 88% match-rate band automates the bulk of class-1 fee-deduction matching, class-4 rounding detection, class-5 short-settlement residual queue, and the class-7 unexplained-variance reduction. The playbook routes the engine’s surfaced residuals to the right owner with the right SLA. Together they deliver the 60-80% recovery rate on standing leakage that the playbook targets.
Without the engine, the playbook is operationally heavy at the AR-controller seat and the recovery rate plateaus at 30-50% because manual surfacing of residuals leaves too much in the unexplained pool. Without the playbook, the engine produces clean residuals that still age out because no owner is bound by an SLA to chase them. The deployment sequence in this article assumes both are being put in place in parallel.
Common deployment mistakes to avoid
Mistake one: starting with class 7 (unexplained variance) instead of class 1 (fee deduction) or class 2 (tax deduction). The high-rupee, high-recovery-probability classes are the right place to build the operating muscle. Unexplained variance shrinks as the other classes get classified out.
Mistake two: under-staffing the tax-deduction class. The TDS desk needs sustained engagement with deductors; recovery does not happen with monthly batch follow-up. Daily 26AS / 168 pulls and weekly escalation packs are the minimum cadence.
Mistake three: setting recovery-rate targets above 80% in year one. The structural-loss residual is real for every class; pushing for 90% recovery in year one frustrates owners and devalues the program.
Mistake four: not making the audit-committee cycle the forcing function. Without quarterly external review, the program drifts within 6-9 months. The committee discipline is what sustains the operating cadence.
Continue reading on the leakage backbone
For the umbrella framing, start at Revenue leakage and the Seven Classes framework. For the board-approval companion, see Building the board case for revenue leakage recovery. For the working-capital cost overlay, see Working capital leakage from reconciliation delays. The Stop Revenue Leakage pillar page anchors the broader story.