GST Reconciliation Software: Metrics and Thresholds

Most Indian finance teams do not have a GST matching problem in the abstract. They have a specific one: a purchase register that will not agree with GSTR-2B, and a credit ledger that is short by an amount nobody can quantify until the CA asks. It hits finance leads at SMBs and mid-market firms running 500 to 5,000 invoices a month on Tally, Zoho Books or a custom ERP, usually one person owning reconciliations alongside everything else. Software solves part of it; knowing which numbers to watch solves the rest.

What should gst reconciliation software actually measure?

It should measure the value, age and cause of every mismatch, not simply count how many invoices matched. A match percentage on its own tells you nothing you can act on — 94% may hide ₹8 lakh of high-value gaps while 88% may contain nothing but ₹200 rounding differences. The useful output is a rupee figure per vendor, per mismatch type, with an ageing clock on each one. The categories that matter are missing in 2B, GSTIN errors, period mismatches, tax amount differences, and place-of-supply errors. Anything that produces a report you cannot work through line by line is a filing accelerator, not a reconciliation tool. Since the Invoice Management System went live, 2B also carries supplier-side actions you can accept or reject, and those decisions belong in the same measurement.

The five metrics worth tracking every month

Track a small fixed set, so month-on-month comparison is possible. In our experience these five cover most of the ground:

• ITC value at risk (₹). Total input tax credit on purchase invoices with no matching line in GSTR-2B. This is the number the CFO asks for first.
• Auto-match rate (%). Share of 2B line items matched to the purchase register without human touch, split by tax period.
• Ageing of unmatched ITC (days). Bucket into 0-30, 31-90, 91-180 and 180+. Past 180 days it is a Section 16(4) risk, not a chasing problem.
• Vendor mismatch concentration. Top 10 vendors by unmatched value. Three or four suppliers usually explain most of the gap.
• Filing close time (days). From 2B download on the 14th to GSTR-3B filing. A team starting on the 12th is guessing.

What thresholds should trigger action?

Write the thresholds down before you look at results, so the decision is not relitigated every month. Reasonable starting points for a business with ₹50 lakh to ₹2 crore of monthly purchases:

• Auto-match rate below 90% → audit the matching rules before filing; usually a rule is wrong, not your vendors.
• ITC at risk above ₹1 lakh, or above 2% of monthly ITC → hold the disputed portion, chase the top vendors, escalate to the finance head.
• Any unmatched invoice past 150 days → move to a written vendor escalation; past 180 days → decide explicitly whether to write it off.
• The same vendor in the top-10 mismatch list three months running → treat it as a process failure on their side, not bad luck.
• Close time slipping past the 18th → the bottleneck is upstream document collection, not reconciliation.

A worked example on ₹1.2 crore of monthly purchases

A mismatch rate that looks tolerable is often expensive once you value it in rupees. Take a distributor buying ₹1.2 crore a month across roughly 1,800 invoice lines at 18% GST, so monthly ITC is about ₹21.6 lakh. A 12% auto-match failure works out to ₹14.4 lakh of unmatched purchases and roughly ₹2.6 lakh of credit sitting in limbo in a single month. If half of that is never recovered — a fair assumption for credit nobody chases — the annual leak is around ₹15-16 lakh. That is usually more than the software and the person running it cost combined, and it is avoidable with a top-10 chase list and a 30-day ageing review.

What to automate and what to keep manual

Automate extraction, matching and flagging; keep write-offs and vendor disputes with a human name against them. Pulling GSTR-2B and the purchase register into one schema and matching on GSTIN, invoice number and a date tolerance is mechanical work software does better than people. That layer is usually the first thing GST billing and reconciliation automation covers, because everything downstream depends on it. What you should not automate is the judgement calls: accepting a credit note, tolerating a small rounding mismatch, or treating a GSTIN typo as closed. Before committing to a build, compare the three-year cost of the tool against the ITC you expect to recover — the pricing page is a reasonable place to frame that comparison internally rather than inside a vendor meeting.

Where to start

Start with one month of data, not a full-year clean-up. Take the most recent month with a 2B file, put the five metrics on a single page, and fix the thresholds in writing before you see the results — otherwise you will rationalise whatever you find. Do it twice and you will know whether the problem sits with your vendors, your master data, or your process. Maysan Labs builds these reconciliations for finance teams as part of its GST and billing automation work, and we are happy to talk through the methodology whether or not anything gets built.