thumb
  • FrenzoCollect

  • 28-09-26

SMA-0, SMA-1, SMA-2: why the most valuable 90 days in lending get the least attention

SMA stands for Special Mention Account - the RBI's classification for loans showing early stress, before they cross the 90-day line into NPA. The buckets are simple: SMA-0 covers accounts overdue by 1 to 30 days (and accounts showing incipient stress even without overdue payments), SMA-1 covers 31 to 60 days, and SMA-2 covers 61 to 90 days. Day 91, the account becomes an NPA, provisioning kicks in, and everything gets harder.


That's the regulatory definition. The operational meaning is more interesting: SMA is a 90-day countdown timer, and what a lender does while it runs determines most of what the recovery numbers look like later.


The regulator built the window; most lenders waste it

The SMA framework exists because the RBI wanted stress made visible early - lenders must identify and report these accounts, and the framework's whole premise is that early identification enables early correction. Daily stamping rules make the clock precise: an account's SMA status is marked as part of day-end processes, so slippage between buckets is recorded the day it happens, not at month-end.


What the framework can't mandate is what lenders do with the information. In practice, most collections operations treat SMA-0 as a reporting category rather than a working queue. Serious effort begins somewhere in SMA-1 or SMA-2, when the account is 45–75 days old, the borrower has stopped answering unknown numbers, and the recovery probability has already fallen from above 80% to somewhere near half of that.


The perverse result: the window with the highest recovery odds and the lowest intervention cost gets the least attention, and the buckets where every rupee is expensive get the most.


What working the SMA window actually means

Working SMA-0 properly is mostly a data and automation problem, which is why manual operations skip it - the volumes are too high for agents and the accounts don't feel urgent yet.


It starts before day 1. Incipient-stress signals - a bounced payment elsewhere visible in bureau data, a salary credit that didn't arrive, utilisation spiking on other lines - can flag an account while it's still current. On the platform side this is what predictive scoring is for: producing a default probability per account so that outreach begins in the days before the due date for the accounts that need it.


Days 1 through 15 belong to automation. A borrower one week overdue usually needs a reminder, a payment link, and possibly a small accommodation - not a hard conversation. SMS and WhatsApp sequences with self-serve payment resolve a large fraction of SMA-0 at nearly zero marginal cost, which is the only economical way to cover a bucket this large.


By days 15 to 30, differentiation matters. Accounts that ignored two automated touches and carry rising risk scores route to agents; accounts that engaged but didn't pay may need a restructure conversation. The point of intelligent routing is that agent hours land only where automation has already failed, which is a small minority of the bucket.


Handled this way, SMA-1 and SMA-2 become small buckets of genuinely hard cases rather than a broad river of neglected ones. On FrenzoFinserv, lenders see the largest single recovery uplift - around 38% - in exactly these pre-due and early buckets, because the platform is doing work that previously wasn't being done at all.


The Q2 FY27 angle

September 30 closes the quarter. Every account sitting in SMA-2 today either resolves in the next few weeks or reports as a fresh NPA slippage in the Q2 numbers - and every account in SMA-0 today is deciding, quietly, which side of the FY27 second-half story it joins. A lender who can see the whole SMA population live, scored and routed, is managing that outcome. A lender reading batch reports is watching it.


If your SMA visibility today is a weekly Excel export, that's fixable in weeks, not quarters. FrenzoFinserv puts live DPD and SMA tracking, predictive scores and automated pre-due workflows on top of your existing LMS in 4–6 weeks. The FAQ has a fuller primer on DPD buckets if you want the foundations first.


FAQ

What is the difference between SMA and NPA?

SMA covers the first 90 days of overdue status (SMA-0: 1–30 days, SMA-1: 31–60, SMA-2: 61–90). At 90+ days overdue the account is classified NPA, triggering provisioning requirements.


Can an account be SMA-0 without a missed payment?

Yes. SMA-0 also covers accounts showing signs of incipient stress even where payments are not yet overdue.


Why does SMA-0 matter so much for recovery?

Recovery probability in the earliest window exceeds 80% and interventions cost the least. Every bucket the account rolls through roughly halves the odds and multiplies the cost.