The 57 Reconciliation Errors + Detection Envelope — CA One-Pager
The definitive one-page reference on the 57 human-error patterns Terra Insight tests against, structured for the accountant who decides. Print it, circulate it, hand it to a partner review or a client CFO.
Reconciliation is a discovery discipline, not a matching discipline
Reconciliation software is usually pitched as matching software. But matching clean data is easy — the real job is surviving the data finance teams actually produce, files touched by tired clerks at month-end, exported from systems configured years ago, maintained by people who know their business but not every section of the Income-tax Act. And here is the reframe that changes the pitch: many of these errors are not noise the software must tolerate, they are exactly what reconciliation exists to find — a payment recorded twice, an entry posted to the wrong month, a receipt that bounced but was never reversed. So Terra Insight built a catalogue of every human error we could enumerate — 57 of them, in five families — and we manufacture each one, deliberately, into realistic Indian financial data before any release, each with a pre-computed expected outcome the product is graded against.
The full catalogue
Every pattern below is a publicly known failure mode of real Indian financial data. What Terra Insight adds is the discipline of testing against all 57 on every release, on realistic data of the shape your firm actually produces.
Family 1 — Carelessness 16 errors
The month-end-at-9pm class. Nobody is careless on purpose.
| # | Pattern | What it looks like in the data |
|---|---|---|
| 1 | The same row entered twice | byte-identical duplicate in one file |
| 2 | The same invoice re-keyed | same economics under a new voucher number |
| 3 | Transposed digits | Rs 45,000 keyed as Rs 54,000, parses cleanly |
| 4 | The fat-finger zero | one amount x10 or ÷10 of the true value |
| 5 | Day and month swapped | 05/03 read as 03/05, invisible above day 12 |
| 6 | Last month's block pasted in | a copied set of rows from the prior period |
| 7 | The half-filled row | key cells left blank mid-entry |
| 8 | Invisible characters in key fields | trailing space or zero-width char on GSTIN, UTR |
| 9 | A return typed as a positive | one reversal with the sign flipped |
| 10 | Hand-rounded paise | one side rounded to whole rupees, other not |
| 11 | The same file uploaded twice | in two slots, or twice in one upload |
| 12 | Overlapping statement files | two exports whose date windows overlap |
| 13 | One debit posted as a credit | a single dr/cr slip on the wrong side |
| 14 | A hand-mangled amount | 1.00.000 or 1,00,00.00 typed by hand |
| 15 | A year typo | one row dated a year off |
| 16 | The counterparty spelled differently | Nimbus Traders vs Nimbus Trader's, same client |
Family 2 — Knowledge gaps 11 errors
A wrong belief, applied consistently. Consistency looks like correctness.
| # | Pattern | What it looks like in the data |
|---|---|---|
| 1 | The wrong GST rate, applied everywhere | 12% where 18% is due, across all documents |
| 2 | TDS under the wrong section | contract deducted as 194J instead of 194C |
| 3 | Gross where net belongs | GST-inclusive value in a net-of-tax field |
| 4 | Debit/credit convention inverted for a block | a contiguous run with the convention flipped |
| 5 | Tax computed on the wrong base | TDS deducted on the GST-inclusive invoice |
| 6 | TDS not deducted near a threshold | Section 194Q Rs 50-lakh boundary judged wrong |
| 7 | Line-level vs invoice-level rounding | sides round at different levels, totals off by paise |
| 8 | The advance never linked | money received in advance never tied to invoice |
| 9 | The 1% vs 2% TDS step | individual rate applied to a company or vice versa |
| 10 | CGST+SGST where IGST belongs | intra-state treatment on an inter-state supply |
| 11 | Reverse charge missed entirely | RCM purchase with no self-invoice, no liability |
Family 3 — Misconfigured systems 11 errors
Nobody typed anything wrong. The software was set up wrong once, years ago.
| # | Pattern | What it looks like in the data |
|---|---|---|
| 1 | The fiscal year set wrong | every FY label in the export shifted by one |
| 2 | The timezone date-shift | UTC export moves month-boundary rows back a day |
| 3 | A stale opening balance | brought-forward figure never updated |
| 4 | Voucher numbering reset mid-year | same voucher numbers appear in both halves |
| 5 | Amounts exported in lakhs | every figure silently divided by 1,00,000 |
| 6 | A foreign-currency or quantity column drift | some rows carry the wrong unit in amount |
| 7 | Account codes remapped mid-period | same ledger account under two codes |
| 8 | The export cut at a row cap | the system's page limit silently drops the tail |
| 9 | A stale saved-view header | header says Debit while column carries something else |
| 10 | Decimal precision drift | paise pre-truncated or three-decimal exports |
| 11 | Narrations truncated | description field cut mid-UTR by the source system |
Family 4 — Gaps in the data 9 errors
What's missing is invisible unless the arithmetic is forced to account for it.
| # | Pattern | What it looks like in the data |
|---|---|---|
| 1 | A missing fortnight | a hole in the middle of the statement |
| 2 | An entire bank account absent | one of the company accounts simply not supplied |
| 3 | A sidecar covering half the period | register ends mid-month, ledger runs full |
| 4 | The books side missing entirely | a statement with nothing to reconcile against |
| 5 | Random rows lost | 5% of one stream gone, scattered |
| 6 | A header with zero rows | a valid, empty file |
| 7 | The wrong month in the slot | well-formed March file uploaded as April |
| 8 | A required register not uploaded | the module is on, its input never arrived |
| 9 | The window-dressed truncation | tail cut AND closing balance restated to match |
Family 5 — Missing and mistimed entries 10 errors
The quarry. Finding these is what reconciliation is FOR.
| # | Pattern | What it looks like in the data |
|---|---|---|
| 1 | Money in the bank, not in the books | an unrecorded credit |
| 2 | Money in the books, not in the bank | an unrecorded or unpaid obligation |
| 3 | The cutoff straddle | books say March, bank says April |
| 4 | The late posting | entry lands one period behind |
| 5 | A reversal with no original | the undo exists, the do doesn't |
| 6 | A bounce never reversed | bank shows the return, books still show the receipt |
| 7 | A cheque issued, never presented | outstanding beyond its window |
| 8 | A deposit that never cleared | booked, but no credit arrived |
| 9 | The bounce pair, unbooked Most dangerous | credit and return both in bank, books certify money already gone |
| 10 | The vendor paid twice | two bank debits, one books entry |
The Detection Envelope — what it is
- What. A written per-customer report Terra Insight delivers before go-live, structured against this 57-pattern catalogue.
- Content. Every one of the 57 patterns is classified as catch, catch — conditional, or structural miss for the customer's specific data shape, presets, and companion registers.
- Delivery. Signed off inside the 2 to 4 week configuration and preset-tuning window that every deployment runs through — before any statutory reporting cycle depends on it, and before contract locks in production capacity.
- Purpose. The trust artefact your statutory auditor reads. An honest statement of what we can and cannot promise, so the exception plan is designed around what automation cannot cover, rather than under a false assumption that it does.
A test only passes when the product surfaces the error as the right exception, with the right rupees attached. Silently absorbing an error, or worse, letting an error improve the match rate, is an automatic failure. Overclaiming would make us the thing we built this product to kill.
What no reconciliation can structurally catch
An error consistent on both sides is invisible to any cross-check ever built.
Cross-checking is defined as comparing two sides and finding where they disagree. If both sides carry the same wrong information, they agree perfectly and no reconciliation surfaces an exception. Two illustrations from the catalogue above:
- Family 2, wrong GST rate applied consistently. If the invoice carries 12% where 18% is due, and the return filed against it also carries 12%, invoice-to-return reconciliation will match cleanly — the two sides agree. Catching this requires a rate authority external to both sides (the statutory rate schedule, an HSN classification review, or a statutory audit sample).
- Family 5 #9, the bounce pair where the return notification never reached the books. Books show one receipt; bank shows one credit and one debit (net zero). If the book side never received the bounce notification, no cross-check of book-to-bank matches the pair — because the pair is present on one side and absent on the other. Mitigable by ingesting a companion bank feed that tags returns explicitly, hence classified catch conditional rather than structural miss.
The Detection Envelope names each structural miss explicitly so the control layer around it — data-source design, four-eyes review, statutory sampling, audit inquiry — is the layer designed to catch it, not silently absent under the false assumption that automation is on the job.
Three ways this earns its shelf
- Print it. Circulate it. Ask the question. Print this page and give it to the client CFO or the engagement partner. Ask which of the 57 they have lived through in the last twelve months. The list is empirically complete enough that the conversation converges on real, concrete failure modes rather than on abstract "control gaps".
- Pair it with the Detection Envelope. The per-customer Detection Envelope report Terra Insight delivers before go-live is the operational mirror of this catalogue — with the classification of each pattern (catch, catch-conditional, structural miss) written against your client's specific data shape, presets, and companion registers. This one-pager is the framework; the Detection Envelope is the filled-in copy.
- Combine with the three-layer trust stack. The Reconciliation Process Design cluster is the design authority (how to build the control taxonomy), the Reconciliation Playbook cluster is the operations authority (the 20-day monthly cadence), and this 57-Errors Catalogue is the test-honesty authority (what we test against and what we honestly catch). Together with the buyer's-guide pillar, they give the accountant every framework needed to sign off.
Extend this reference with the anchor page, the clusters, and the product
The definitive statement of what the Detection Envelope is, how it is delivered, and why honesty about coverage limits is the trust asset that closes the auditor decision.
Hub for the five family articles that expand every error pattern with Indian regulatory context, worked examples, and Detection Envelope classification framing.
The 12-class failure-mode taxonomy — the framework buckets the 57 concrete instances collapse into when you design the control layer around them.
The 20-day monthly-close cadence that gives every one of the 57 patterns a discovery window inside a working, repeatable close.
The economic-scale companion to this catalogue — how much revenue Indian mid-market finance functions leak, cluster by cluster, and what a control-first response returns.
The product that carries this catalogue into production — 24 industry presets, 40+ engineered signals, and a written Detection Envelope every deployment receives before go-live.
Download-ready PDF is on the way
A single-sheet A4 PDF version of this reference, formatted for partner-review circulation inside CA firms and controllers' offices, is in preparation as a follow-up asset. In the meantime, use the Print / Save as PDF button below — it prints cleanly on A4 without any header, footer, or navigation clutter.