Manual reconciliation teams that adopt a manufacturing cause taxonomy — Ishikawa's 4M (Man, Machine, Material, Method) or its 6M extension (adding Measurement and Milieu) — find that four of the six categories are either empty or ambiguous when applied to a finance function. There is no Machine in reconciliation, no Material, the Measurement is the reconciliation itself, and the Milieu category is forced to absorb portal downtime, cutoff drift, cross-era payment-code confusion, and counterparty behaviour as a single environmental bucket. The residual 'Method' category then has to hold SOP gaps, hand-off failures, cutoff discipline, sign-off matrix ambiguity, and period-boundary drift. A 4M or 6M walkthrough on a Section 16(4) permanent-loss failure typically produces the conclusion 'method error' — a diagnosis that is technically true but too imprecise to design a prevention control against.
Publish the 6P cause taxonomy — People, Policy, Process, Portal, Period, Partner — as the finance-team replacement for Ishikawa 4M or 6M. People covers the finance team's own analyst error, cross-training gap, and seasonal load. Policy covers the SOP and treatment questions the written documentation does not answer. Process covers the finance team's own cutoff discipline, hand-off between AR and AP and tax teams, and sign-off matrix. Portal covers CBIC, CBDT, GSTN, TRACES, and bank-side timing the finance team does not control — including the IMS 15-day action window, the Rule 88C DRC-01B seven-day reply clock, and the TRACES lag between challan deposit and Form 168 reflection. Period covers the boundary between two reporting eras — the Section 393 payment code framework from 1 April 2026, the Section 16(4) 30 November cutoff, the Section 39(9) amendment window, quarter-end drift. Partner covers counterparty behaviour — supplier GSTR-1 delay triggering Rule 37A cascading reversal, deductor Form 168 delay, aggregator file-format changes, DTAA Tax Residency Certificate delays. Every failure mode on the register carries at least one P; many carry two.
One-column addition to the reconciliation control plan template — the cause column carries the P (or two P's) that produced the failure mode. The 6P is a walkthrough discipline, not a scoring input — the Severity, Occurrence, and Detection ratings sit downstream. The register is walked by the controller monthly for Occurrence and Detection re-rating and re-opened on any process change, portal change (Form 168 switchover, IMS live-cutover, GSTR-2B date shift), or after every field incident that reveals a new failure mode. A new P is not added — the taxonomy is fixed at six by design so that walkthroughs remain comparable across streams and across reporting periods.
A reconciliation risk register whose cause column separates Portal and Partner from Process — the two most systematically under-diagnosed categories in an Ishikawa-driven walkthrough — and whose prevention and detection controls are calibrated to the actual origin of the failure rather than to the residual 'Method' bucket. A register configured against the 6P routinely surfaces failure modes whose true cause is Partner (Rule 37A supplier default) or Portal (IMS 15-day action window, Rule 88C seven-day DRC-01B clock) that a 4M or 6M walkthrough had collapsed into a generic 'Method' or 'Milieu' entry and left without a specific control lineage.
A reconciliation failure that a manufacturing engineer would classify as a “Machine” problem — the ERP posted a wrong value — often has a finance-side cause the Ishikawa 4M category cannot see. The supplier’s own GSTR-1 was filed four months late and triggered a Rule 37A cascading reversal. The TRACES portal had not yet reflected the challan. The SOP was silent on the cross-era payment code that landed in Q1 of FY 2026-27. Ishikawa’s 6M — Man, Machine, Material, Method, Measurement, Milieu — was built for a physical production line and carries no category for any of these causes; applied to a finance function, it collapses six or seven distinct Indian causes into a single “Method” bucket.
This method article publishes Terra Insight’s 6P cause taxonomy — People, Policy, Process, Portal, Period, Partner — the finance-team replacement for Ishikawa 4M or 6M. It sits inside the reconciliation process design method as the cause column of the register and works alongside the anchored SOD scale that rates each failure mode.
Why the manufacturing Ishikawa 4M or 6M does not fit finance
Ishikawa’s cause taxonomy was built for a defect on a production line. The four Ms — Man, Machine, Material, Method — divide the causes into operator, tooling, input material, and procedure; the 6M extension adds Measurement (the inspection instrument) and Milieu (ambient environment). Applied to a finance reconciliation function, four of the six categories are either empty or forced.
- Machine. A reconciliation has no machine — the ERP, the GSTN portal, and TRACES are instruments the process uses. What manufacturing calls Machine is almost always a Portal cause in reconciliation.
- Material. A reconciliation does not consume raw material — its inputs are documents produced by counterparties or portals, and they belong to Partner or Portal.
- Measurement. The reconciliation itself is the measurement — the category is circular.
- Milieu. Portal downtime, cutoff drift, and cross-era rule changes each deserve their own category rather than a single ambient bucket.
The residual Method category then has to hold every SOP gap, every hand-off failure between AR and AP and tax, every cutoff-calendar drift, every sign-off ambiguity, and every period-boundary drift. A 4M walkthrough on a Section 16(4) permanent-loss failure typically produces the conclusion “method error” — technically true, but too imprecise to design a prevention control against.
The 6P replacement
The 6P splits Method into its three distinct finance-team causes (Policy, Process, Period), promotes portal-side timing to a first-class category (Portal), retains People for the analyst-error slot that maps to Man, and adds Partner — a category the manufacturing taxonomy has no analogue for because the recipient’s exposure is entirely a function of the counterparty’s own filing behaviour.
People — analyst error
Analyst turnover, cross-training gaps, seasonal load. A skipped row, a Section 194J invoice deducted at the Section 194C rate on manpower supply, a Section 393 payment code from the 1001-1092 range applied to a legacy FY 2025-26 transaction that should have carried the old Section 194x identifier, the analyst on the desk during the October-to-November window that overlaps the Section 16(4) 30 November cutoff, the Section 39(9) amendment window closing on the same date, the Rule 37A 30 November reversal deadline, and the year-end audit fieldwork. Reduced by training, checklists, and independent peer review.
Policy — written-documentation gap
The SOP is silent on the treatment question. Materiality thresholds are not defined at Rs 5,000 per invoice or Rs 5 lakh at aggregate. The sign-off matrix does not distinguish preparer from reviewer on the Rs 5 lakh, Rs 25 lakh, and Rs 1 crore rows. Foreign OTA commission RCM policy is undocumented. TDS on GST-inclusive amount is not explicitly forbidden. Blocked ITC under Section 17(5) is not enumerated. The Section 195 non-resident payment path — payment code 1057 versus the residual 1027 domestic code — is not published to vendor onboarding. Reduced by written treatment memos, annual policy review, and signed analyst acknowledgement.
Process — finance-team-controlled cadence
The bank statement is pulled before cutoff. The GSTR-2B is pulled before the supplier’s late filings settle. The monthly close is signed before the challan is confirmed on the CBDT portal. The exception queue is reviewed by the same analyst who created it. The hand-off between the AR analyst running invoice-to-bank and the tax head running GSTR-1 carries no documented protocol. Reduced by a redesigned SOP, a published cutoff calendar, and hard preparer-versus-reviewer separation.
Portal — CBIC, CBDT, GSTN, TRACES, and bank-side timing
The TRACES portal has not yet reflected the deductor’s Q3 challan on the day of the reconciliation. The GSTN portal is in a maintenance window. The bank’s net-banking export capped the pull at 999 rows and the analyst missed the pagination. The Invoice Management System introduced on the GSTN portal in October 2024 has a 15-day action window from GSTR-2B availability that closes regardless of whether the reviewer is on leave. The Rule 88C DRC-01B intimation carries a seven-day reply clock that starts the moment the notice lands on the portal. Form 168, the quarterly deductor statement replacing Form 26Q from 1 April 2026, has its own filing cadence the recipient’s reconciliation depends on. Portal-P failures cannot be prevented by any change to the finance team’s own SOP — they are reduced by a monitoring cadence, a portal-status log, and a rerun-if-stale rule.
Period — reporting-era boundary
The Section 393 payment codes 1001 to 1092 in effect from 1 April 2026 are a Period surface — a receivable booked in FY 2025-26 that settles in Q1 of FY 2026-27 must be reconciled with the legacy Section 194x code on the invoice side and the new payment code on the challan side (cross-era TDS matching is a distinct failure class). The Section 39(9) GST amendment window closing on 30 November, the Section 16(4) 30 November permanent-loss cutoff, quarter-boundary drift on TDS receivable at Q4-to-Q1 transitions, and the FY-end cutoff on Section 43B(h) MSME payables that flips a payment beyond 45 days into a taxable-income adjustment are all Period surfaces. Reduced by an ageing queue keyed to the specific statutory cutoff date.
Partner — counterparty behaviour
The supplier files GSTR-1 four months late and triggers a Rule 37A cascading reversal by 30 November. The deductor files Form 168 for the quarter three weeks after the recipient’s audit fieldwork begins. The aggregator changes the reconciliation-file column layout without notice. The non-resident vendor’s Tax Residency Certificate under the applicable DTAA arrives three weeks late, exposing the recipient to a Section 206AA higher-rate 20 percent deduction on missing PAN documentation. The Section 194Q buyer’s TDS obligation at payment code 1031, triggered by aggregate purchases above Rs 50 lakh from a single supplier, depends on the supplier’s own turnover disclosure. Reduced by a supplier watchlist keyed to filing history, a vendor-onboarding PAN validation status, and a DTAA-partner documentation checklist.
The failure-mode mapping
Every reconciliation failure mode traces to at least one P; many trace to two. The mapping below is the walkthrough the controller runs across the twelve-class failure-mode taxonomy for each function on the register.
| Failure class | Typical primary P | Typical secondary P |
|---|---|---|
| Data extraction | Portal | Process |
| Data classification | People | Policy |
| Data completeness | Partner | Portal |
| Matching logic | People | Policy |
| Timing and period | Period | Portal |
| Counterparty behaviour | Partner | — |
| Precision and numeric | People | Policy |
| Policy and interpretation | Policy | People |
| Ageing and escalation | Process | Period |
| Cutoff and sign-off | Process | Period |
| Documentation and evidence | Process | Policy |
| Portal and system | Portal | Process |
A failure class that a 4M or 6M walkthrough would collapse into “Method” — timing, ageing, cutoff, documentation, portal — splits across four different P’s (Portal, Process, Period, Policy), each requiring a different prevention control. That granularity is why the 6P surfaces failure modes an Ishikawa-driven walkthrough had systematically missed.
Worked walkthrough — a GSTR-2B ITC failure through the 6P
A mid-sized Indian distributor with a Rs 240 crore annual GST-eligible purchase base runs the GSTR-2B ITC reconciliation monthly. In September 2026 the analyst clears a Rs 3.2 lakh ITC variance on Supplier A as “supplier will file late, roll forward”. Supplier A files no GSTR-1 for October through December and no GSTR-3B by 30 September 2027 — the Rule 37A cutoff. On 30 November 2027 the recipient reverses Rs 3.2 lakh with interest under Section 50 at 18 percent per annum from the September 2026 availment date, adding Rs 68,220 on top of the reversal. The 6P walkthrough surfaces four P’s on this single failure mode:
- Partner. Supplier A’s own non-filing is the root cause. Prevention: a supplier watchlist flagging any counterparty with a filing lag greater than 45 days on the previous cycle.
- Portal. GSTR-2B refreshes on the 14th; the IMS 15-day action window closes on the 29th. Prevention: a monthly IMS-review cadence pegged to the 14th-to-29th window.
- Period. The Rule 37A 30 September and Section 16(4) 30 November cutoffs are two distinct Period surfaces. Detection: an ageing queue keyed to both dates — the invoice ages against Rule 37A first, then against Section 16(4).
- Policy. The written treatment for “supplier will file late, roll forward” was permissive. Prevention: a policy rewrite that forbids roll-forward without a Rule 37A ageing note and a Section 16(4) exposure calculation.
A 4M walkthrough would have flagged only “Method — supplier late” and left the prevention control undesigned. The 6P produces four distinct controls, each pinned to the P that generated the cause.
When the 6P outgrows the manual detection layer
A well-walked 6P register does not remove failure modes — it names them and pins each to a specific control. Those controls — a supplier watchlist refreshed daily, a portal-cadence monitor pegged to the IMS 15-day window and the Rule 88C DRC-01B seven-day reply clock, an ageing queue keyed to both the Rule 37A and Section 16(4) cutoffs, a Section 393 cross-era payment-code translation table — sit inside a manual spreadsheet as long as the counterparty base and the portal event volume remain tractable. Above roughly 200 suppliers, above roughly 3,000 monthly purchase invoices, or on a multi-GSTIN structure carrying more than three GSTINs, the manual detection layer for the Portal and Partner P’s stops being economically viable across the September-to-November window that compresses the Rule 37 ageing walk, the Rule 37A annual walk, and the Section 16(4) November lockdown into a single quarter. Terra Insight’s reconciliation software carries the continuously-refreshed detection layer the 6P register demands on Partner and Portal rows — the GST reconciliation software closes the Rule 37A and Section 16(4) queues, and the TDS reconciliation software closes the cross-era Section 393 queue and the Form 168 portal-cadence queue on the same operating rhythm.
Where this fits
- Reconciliation process design — the methodology pillar
- The anchored SOD rating scale for Indian reconciliation
- Action Priority for reconciliation — severity-first prioritisation
- Manual detection techniques for reconciliation
- Reconciliation control plan template
- Monthly close reconciliation playbook
Frequently Asked Questions
Why does the manufacturing Ishikawa 4M or 6M cause taxonomy not fit finance reconciliation?
Ishikawa’s 4M — Man, Machine, Material, Method — and its 6M extension adding Measurement and Milieu were built for a physical production line where a defect can be traced to raw material, tooling, operator, or environmental condition. Applied to a finance reconciliation function, four of the six categories are either empty or ambiguous. There is no “Machine” in reconciliation — the ERP, the GSTN portal, the TRACES portal, and the spreadsheet are all instruments the process uses rather than machines that produce a physical output. There is no “Material” — a reconciliation does not consume raw material. “Measurement” is circular because the reconciliation itself is the measurement. “Milieu” collapses portal downtime, cutoff drift, and cross-era payment-code confusion into a single environmental bucket. The residual real category — “Method” — has to absorb SOP gaps, sign-off matrix ambiguity, cutoff-calendar drift, hand-off gaps between AR and AP and tax teams, and cross-era period boundaries. That single-bucket compression is why a 4M or 6M walkthrough on a Section 16(4) permanent-loss failure typically produces the useless conclusion “method error”. The 6P — People, Policy, Process, Portal, Period, Partner — splits Method into its three distinct finance-team causes (Policy, Process, Period), promotes portal-side timing to a first-class category (Portal), and adds a Partner category the manufacturing taxonomy has no analogue for because the recipient’s exposure is entirely a function of the supplier’s or deductor’s own filing behaviour.
How is Portal different from Process in the 6P taxonomy?
Process is what the finance team controls — the SOP step, the cutoff discipline, the hand-off between the AR analyst and the tax head, the sign-off matrix at each Rs threshold. If the process is redesigned tomorrow the failure mode goes away. Portal is what the finance team does not control — the CBIC or CBDT publishing cadence, the GSTN downtime window, the TRACES lag between challan deposit and Form 168 reflection, the Invoice Management System 15-day action window that closes on the 30th of the following month regardless of whether the reviewer is on leave, the DRC-01B seven-day reply clock that starts the moment the intimation lands on the portal. A Portal-P failure cannot be prevented by redesigning the finance team’s own process — it can only be detected earlier and reacted to faster. The distinction matters because the prevention control for a Process failure is a written SOP change, and the prevention control for a Portal failure is a monitoring cadence that catches the portal event within the portal’s own reply window.
What makes Partner a distinct P from People?
People is the finance team’s own analyst error — a skipped row, a wrong period, a transposed digit, an unfamiliar Section 393 payment code applied to a legacy Section 194J case. Partner is the counterparty’s behaviour — the supplier who files GSTR-1 four months late and triggers a Rule 37A cascading reversal on the recipient’s ITC, the deductor who files Form 168 late and leaves the recipient’s TDS receivable un-credited, the aggregator who changes the reconciliation-file column layout without notice, the non-resident vendor whose Tax Residency Certificate under the applicable DTAA arrives three weeks after the payment is due. A Partner-P failure has a distinct signature — the finance team is doing everything correctly and the failure still lands, because the counterparty’s own filing or documentation is the input on which the reconciliation depends. The prevention control for Partner failures is a supplier watchlist keyed to filing history, a vendor onboarding discipline that requires a PAN validation status under Section 206AA, and an escalation protocol that starts the moment the counterparty’s ageing crosses a threshold — not any change to the finance team’s own SOP.
Where does Period fit — is it not the same as the cutoff discipline under Process?
The cutoff discipline is a Process failure — the finance team’s own SOP is silent on which invoice booked on 31 March at 23:47 belongs to which financial year, or the SOP is clear but was not followed. Period is different. Period is the failure mode that lives on the boundary between two reporting eras when the rule itself changes. The Section 393 payment codes 1001 to 1092 that took effect from 1 April 2026 are a Period-P failure surface because a receivable booked in FY 2025-26 that settles in Q1 of FY 2026-27 has to be reconciled with the legacy Section 194x code on the invoice side and the new payment code on the challan side. The Section 39(9) amendment window that closes on 30 November following the financial year of the underlying invoice is a Period-P failure surface because the correction opportunity does not exist after that date. The Section 16(4) 30 November cutoff is a Period-P failure surface for the same reason. Period causes look like Process causes at first glance, but they are not fixable by a better internal SOP — they are fixable only by an ageing queue keyed to the specific statutory cutoff date.
How is the 6P used in an actual failure-mode walkthrough?
The 6P sits on top of the twelve-class failure-mode taxonomy — data extraction, data classification, data completeness, matching logic, timing and period, counterparty and partner, precision and numeric, policy and interpretation, ageing and escalation, cutoff and sign-off, documentation and evidence, and portal and system. For every function on the register the analyst asks, for each of the twelve failure classes, which of the six P’s could produce this class of failure for this specific function. A single failure mode often has two P’s — for example, “IMS Default-Accept on a wrongly-issued invoice from a compromised supplier” is a Partner cause (the supplier issued the wrong invoice) compounded by a Portal cause (the 15-day IMS action window closed before the two-eyes reviewer returned from leave). The prevention and detection controls are then designed against both P’s — a supplier-whitelist gate for the Partner side, a daily IMS-review cadence for the Portal side. The walkthrough is disciplined by the reconciliation control plan template which carries the 6P as the cause column.
- ▸ Section 16(4), Central Goods and Services Tax Act 2017 — Time limit for availing Input Tax Credit. A registered person shall not be entitled to take ITC in respect of any invoice or debit note after the 30th day of November following the end of the financial year to which such invoice pertains, or furnishing of the relevant annual return, whichever is earlier. Section 16(4) is the Period-P anchor for the GSTR-2B stream — a permanent statutory loss with no rectification, no condonation, and no revival path once the November 30 cutoff crosses.
- ▸ Rule 37A, Central Goods and Services Tax Rules 2017 — Where a supplier has furnished the statement of outward supplies in FORM GSTR-1 but has not furnished the return in FORM GSTR-3B for the said tax period by 30 September following the end of the financial year in which the ITC was availed, the recipient shall reverse the ITC availed on such invoices, along with applicable interest, by 30 November of that financial year. This is the Partner-P anchor — the failure originates entirely with the counterparty, but the cash outflow lands on the recipient's books.
- ▸ Section 200A read with Section 201(1A) and Section 234E, Income-tax Act 1961 (retained in Income-tax Act 2025) — Processing of the deductor's TDS statement. Any short-deduction or short-payment identified triggers a demand notice, with interest under Section 201(1A) accruing at 1 percent per month for short-deduction and 1.5 percent per month for short-payment, and Section 234E late-filing fee at Rs 200 per day capped at the tax deductible amount. Section 200A is the Portal-P anchor for the TDS stream — the demand originates from CPC-TDS processing rather than from any book-side visibility.
- ▸ Section 393, Income-tax Act 2025 and CBDT payment code notification 1001-1092 — With effect from 1 April 2026, tax deduction and collection at source is administered through a four-digit payment code framework in the 1001-1092 range that replaces the legacy Section 194x identifiers. Form 168 becomes the deductor's quarterly statement of tax deducted. The cross-era mapping between the legacy section identifier and the new payment code is the Period-P anchor for cross-era TDS reconciliation — a failure the 6P walkthrough surfaces because it originates in the year-of-transaction versus year-of-filing boundary rather than in any single analyst error.
- ▸ Rule 88C and Form DRC-01B, Central Goods and Services Tax Rules 2017 — Where the tax liability declared in GSTR-1 for a tax period exceeds the tax paid in GSTR-3B by the prescribed amount and percentage, an intimation in Form GST DRC-01B is auto-generated. The recipient shall pay the differential with interest under Section 50 through Form GST DRC-03, or furnish a reply within seven days, failing which recovery under Section 79 may follow. Rule 88C is a Portal-P failure surface — the notice is produced by the GSTN portal on a cadence the finance team does not control.