DPD Is a Clock, Not a Diagnosis: Sequencing Collections by Capacity and Intent

Almost every collections book in India is sequenced by days past due. It is the default because it is available, universal and easy to bucket. It is also a measure of how long someone has not paid, which is not the same as a measure of whether they can.

Two borrowers sit at 45 DPD. The first has ₹80,000 moving through their account every month and a salary that lands three days after your presentation date. The second has an account that has been at zero for six weeks. Same bucket, same queue position, same script — and the correct treatments are not merely different, they are opposite. The first needs a date change and no conversation at all. The second cannot be called into solvency.

The variable that separates them is visible in the account, and it is not days.

The two questions that determine treatment

Every delinquent borrower sits somewhere on two axes.

Capacity. Does the money exist? Is there any point in the month when the obligation amount was available?

Intent. Given capacity, is the borrower choosing to pay you?

DPD collapses both into a single number and loses both. Reading them separately produces four groups requiring four different responses.

Capacity present, intent present — timing or mandate failure. The money was there and the debit still failed. This is a presentation date misaligned with the borrower's inflow, an expired or amount-mismatched mandate, or a bank-side limit. The cure is mechanical and costs nothing: re-present after the inflow, or fix the mandate. Calling this borrower spends money to solve a problem that is not theirs.

Capacity present, intent absent — priority decision or dispute. The money exists, other things are being paid, yours is not. This is a dispute, a service failure, or a deliberate ranking of creditors. It is the group where a skilled human conversation has the highest return, and where pressure tactics have the lowest — because the borrower has already demonstrated they can pay and chosen not to.

Capacity absent, intent present — genuine distress. The borrower wants to pay and cannot. Calling harder produces nothing except complaints. This is the restructuring queue.

Capacity absent, intent absent — terminal. Legal and write-off track.

Most collections stacks treat all four as one queue ordered by days, which means the cheapest cures are delayed behind expensive ones and the hopeless cases absorb the same effort as the recoverable ones.

Reading capacity from the account

Capacity is not a score. It is a set of specific, checkable facts.

Was the amount ever available? The single most useful test. Across the month, on how many days did the end-of-day balance exceed the obligation? If the answer is more than zero, capacity existed and the failure was one of timing or authority — not of money. If the answer is zero across the full month, the diagnosis is different and no amount of contact changes it.

Where is the inflow date? Most borrowers have a recurring credit — salary, a settlement, a dominant buyer — landing on a stable date. If your presentation sits before it, you are guaranteed to fail regardless of the borrower's condition.

What is current throughput against baseline? Monthly credit value and transaction count against the borrower's own trailing median, adjusted for known seasonality. A borrower at 60% of baseline throughput has reduced capacity. One at 95% does not.

Is capacity internal or borrowed? If the obligation is being funded by related-party inflows or fresh disbursals from other lenders, capacity exists this month and is fragile. That distinction matters for whether you settle now or restructure.

Has the account gone dormant? An active borrower whose collection account has gone quiet has moved their banking. Capacity may be entirely intact and simply invisible. This is a visibility problem being misread as a capacity problem, and it is common.

Reading intent from the account

Intent sounds unobservable. It largely is not.

Competing obligations are the strongest signal available. If a borrower is servicing three other EMIs on time and failing only yours, that is not a capacity problem. It is a ranking, and you have been ranked. This is the single most actionable piece of information in a collections file and it sits in plain view in the account, requiring only that recurring debits be correctly identified as obligations and attributed to counterparties.

Bounce reason codes distinguish refusal from shortfall. A stopped payment, a customer-requested return or a cancelled mandate is a decision. Insufficient funds is a condition. As set out in our NACH return code reference, these are routinely filed together and should not be — a cancelled mandate on a live loan belongs at the top of the queue, not in an administrative exception report.

Healthy account, failing obligation. Strong throughput, stable balances, and your EMI is the only thing not clearing. Same conclusion as competing obligations, arrived at differently.

Sudden reduction in banked activity with no business explanation. Money may be being deliberately routed away from an account the lender can see, which is an intent signal rather than a capacity one.

Presentation timing is the cheapest recovery in the book

Worth isolating because it requires no contact, no agent, no negotiation and no regulatory exposure.

A large share of first-presentation failures in retail lending are date-alignment problems. The EMI is presented on a fixed calendar date; the borrower's money arrives on a different one. The account holds the amount for twenty days of the month and not on the one day you asked.

Re-presentation strategies in most institutions run on a fixed schedule — day three, day seven, day fifteen — unrelated to anything about the borrower. Sequencing re-presentation against the observed inflow date instead converts a portion of the failed book at zero marginal cost.

This is also the highest-yield fix available to a collections function that cannot expand headcount, because it moves recoveries out of the contact channel entirely.

Why sequencing matters more than it used to

Contact is a constrained resource, and the constraint is tightening.

Recovery contact has been restricted to the hours between 8:00 AM and 7:00 PM since instructions issued in August 2022, applicable across commercial banks, cooperative banks, NBFCs, ARCs and financial institutions. In February 2026 the RBI issued draft Responsible Business Conduct amendment directions for each class of regulated entity, proposed to take effect from 1 July 2026, which would strengthen agent conduct and escalation requirements further — including prior notice for physical visits and recording of recovery calls.

The consequence for a collections operation is arithmetic rather than philosophical. Each borrower interaction is becoming more expensive, more auditable, and available in smaller quantity. When touches were cheap and unlimited, poor sequencing could be absorbed by volume. It cannot now.

There is also a direct cost to getting it wrong. Recovery-related grievances remain among the largest complaint categories the RBI ombudsman receives, and the ombudsman can award compensation for proven harassment. Calling a genuinely distressed borrower repeatedly is not merely ineffective — it generates complaint risk against a recovery that was never available.

Sequencing correctly is therefore not an efficiency project. It is how a shrinking contact budget gets spent on the borrowers where contact is the right instrument at all.

Building the sequence

A workable order, applied within each product and ticket band rather than across the whole book:

1. Route mechanical failures out of the contact queue entirely. Mandate issues, amount mismatches and date misalignment go to re-presentation, not to an agent.
2. Rank capacity-present, intent-absent cases first for human contact. Highest recovery per touch, and the group most responsive to a conversation.
3. Route capacity-absent, intent-present cases to restructuring. Contact here should be about terms, not payment.
4. Hold terminal cases for the legal track rather than cycling them through calling.
5. Re-derive the quadrant monthly. Borrowers move between them. A distressed borrower whose throughput recovers should leave the restructuring queue; one whose account goes to zero should leave the calling queue.

That last point is where most implementations decay. The quadrant is assigned once at delinquency and never revisited, so the book slowly fills with borrowers being treated according to a condition they were in four months ago.

What has to be right underneath

All of this depends on the account being read correctly.

Competing obligations cannot be detected unless recurring debits are identified as EMIs and attributed to the lenders receiving them. Capacity cannot be assessed if loan disbursals are counted as income, because a borrower funding themselves with fresh credit will read as a borrower with healthy inflows. Related-party funding cannot be separated from genuine receipts unless circular flows are netted. The inflow date cannot be established unless the recurring credit is distinguished from one-off ones.

Each of those is a categorisation task, and each fails in the same specific ways we measured across 200,001 transactions. A collections model built on incorrect categorisation does not fail visibly — it produces a confidently wrong sequence, and the book quietly underperforms without anyone able to say why.

Frequently asked questions

Why is DPD bucketing insufficient for collections prioritisation? Days past due measures elapsed time, not recoverability. Two borrowers in the same bucket may differ completely in whether the money exists and whether they are choosing to pay, and those two facts determine which treatment works. Sequencing by DPD alone gives identical treatment to cases requiring opposite responses.

How can you tell if a borrower can afford to pay? Check whether the end-of-day balance exceeded the obligation amount on any day in the month, compare current credit throughput against the borrower's own trailing baseline, and identify whether inflows are operational or funded by related parties and fresh borrowing.

How do you distinguish inability to pay from unwillingness? The clearest indicator is competing obligations: a borrower servicing other lenders' EMIs on time while failing yours has capacity and has made a ranking decision. Bounce reason codes also separate the two, since stopped payments and cancelled mandates indicate refusal while insufficient funds indicates shortfall.

What is the most cost-effective collections intervention? Aligning presentation and re-presentation dates to the borrower's observed inflow date. A significant share of failures are timing mismatches rather than capacity failures, and correcting them recovers money without any borrower contact at all.

How often should collections segmentation be recalculated? Monthly at minimum. Borrowers move between capacity and intent states, and a segmentation assigned once at delinquency progressively misdirects effort as the underlying situation changes.

---

Fiscus reads capacity, competing obligations, inflow timing and conduct from banking data across the life of a loan, so collections sequencing rests on what the account shows rather than on days elapsed. Book a parallel evaluation on a segment of your current book.