How to Handle Non-Registered Supplier Invoices in Japan AP

A non-registered Japanese supplier is a tax-exempt enterprise (免税事業者, menzei jigyōsha) that has not registered as a qualified invoice issuer under Japan's Qualified Invoice System. Their invoices remain legally valid. The buyer's purchase-tax credit on those invoices is capped by the revised transitional schedule: 70% of the Japanese Consumption Tax (JCT) is creditable from October 2026, 50% from October 2028, 30% from October 2030, and 0% from October 2031. Purchases above ¥100 million per tax-exempt supplier per taxable period lose even the transitional credit on the excess.

If you run AP or procurement at a Japan-subsidiary or Japan-facing firm, you have probably already identified that some of your suppliers sit in this bucket. What you need now is the operating playbook, not another explainer on how the invoice system works. This article covers how to handle non-registered supplier invoices in Japan AP end to end — from the vendor-master flag that identifies them through the decisions you make about whether to keep working with them. If you want the regulatory first-principles view before continuing, Japan's Qualified Invoice System compliance requirements walks the framework itself.

The seven steps that follow run in workflow order:

• Step 1 — identify non-registered suppliers and flag them in the vendor master.
• Step 2 — apply the transitional credit to each invoice using the transaction date.
• Step 3 — book the uncreditable JCT portion in the ledger.
• Step 4 — track the ¥100 million per-supplier annual cap.
• Step 5 — run the supplier registration conversation.
• Step 6 — decide whether to continue or switch.
• Step 7 — monitor vendor status over time.

Each step is an operational control with concrete AP deliverables. Together they cover identification, per-invoice treatment, accounting, cap tracking, supplier dialogue, the continue-or-switch decision, and the ongoing maintenance that keeps the workflow honest.

---

Step 1: Identify Non-Registered Suppliers and Flag Them in Your Vendor Master

Given a Japanese vendor list, how does AP know which suppliers are qualified invoice issuers and which are not? The answer sits in one field on the invoice: the T-number (T番号), the 13-digit registration identifier each qualified issuer is assigned by the National Tax Agency (国税庁). A T-number check on each active supplier produces one of three outcomes, and the outcome drives everything downstream.

Outcome 1: T-number present on the invoice and verifies against the NTA registry. The supplier is a qualified invoice issuer (適格請求書発行事業者). Standard input tax credit applies. The vendor master record carries their T-number and a qualified-issuer status.

Outcome 2: T-number present but fails verification. The digits are wrong, the number was withdrawn, or the record was entered incorrectly at either end. Hold the invoice and investigate before posting. Do not assume the supplier is non-registered on a failed check; equally, do not assume a typo without confirming. A brief email to the supplier's accounts-receivable contact resolves most of these quickly.

Outcome 3: T-number absent. The supplier is either a tax-exempt enterprise (免税事業者) or a qualified issuer who has failed to disclose their registration on the invoice. Treat as non-registered until confirmed otherwise. A follow-up with the supplier either surfaces a missing number (update the vendor master and the outcome flips to qualified-issuer) or confirms tax-exempt status.

Carry the result as a concrete field on the vendor master — a JCT-status flag with three values: qualified-issuer, tax-exempt, unknown-pending-verification. This flag is the key that drives downstream transitional-credit calculation and the per-supplier ¥100 million cap tracking covered in Step 4. Without it, the subsequent steps are guesswork at invoice-post time; with it, they are mechanical.

Populating the flag at scale depends on getting the T-number out of every incoming Japanese invoice reliably. Manual capture works at low volume but breaks down once a Japan entity is processing hundreds or thousands of invoices a month, and handwritten spreadsheets of vendor T-numbers fall out of sync with invoices the first time a supplier's registration changes. AP platforms that read Japanese invoices programmatically fold the T-number capture into routine processing — an extraction prompt asks for the T-number alongside the invoice number, date, vendor name, and totals, and the output carries the value through to the vendor-master update. AI-powered invoice data extraction for Japanese AP handles that capture step: the prompt specifies the fields, the system returns them from every invoice in a batch, and the NTA registry lookup against the returned T-number is a separate call the buyer (or the buyer's AP platform) runs against the extracted data. The field-level mechanics of where the T-number appears on a Japanese invoice — its position, formatting, and common variations — are covered in reading the fields on a Japanese supplier invoice.

Qualified-issuer status is not permanent — suppliers deregister, and a supplier who was qualified when onboarded six months ago may not be one today. Step 7 covers the re-verification cadence and the deregistration-handling rules that keep the JCT-status flag accurate over time.

The operational payoff of this step is narrow and load-bearing: without a JCT-status flag on every Japanese vendor-master record, the rest of the playbook cannot run cleanly. The flag is what turns an ambiguous vendor list into a workable segmentation.

Step 2: Apply the Transitional Credit on Each Invoice Using the Transaction Date

The transitional measure (経過措置, keika sochi) on purchases from non-registered suppliers runs in four tranches. Under the revised schedule the FY2026 tax reform (令和8年度税制改正) put in place:

• October 2023 through September 2026: 80% of the JCT on a non-registered supplier's invoice is creditable as input tax.
• October 2026 through September 2028: 70% creditable.
• October 2028 through September 2030: 50% creditable.
• October 2030 through September 2031: 30% creditable.
• October 2031 onward: 0% — no transitional credit, the full JCT on these invoices is non-creditable.

The reform extended the final taper by two years against the original schedule and revised the intermediate percentages. If you are reading older reference material that shows the credit falling from 80% straight to 50% at October 2026, it predates the reform.

The applicable rate on any given invoice is determined by the transaction date — the date the goods or services were supplied — not the invoice-post date and not the payment date. This is the kind of specific operational rule that is easy to miss until a quarter-close audit surfaces a discrepancy. An invoice dated October 5 2026 for services rendered in September 2026 still receives the 80% rate because the transaction fell in the old tranche; an invoice dated September 29 2028 for services rendered October 3 2028 falls into the 50% tranche. The AP system should read the transaction date from the invoice and compute the credit rate from it, not from the header date.

A worked example makes the math concrete. A ¥1,100,000 purchase from a tax-exempt supplier with a transaction date in FY2027 breaks down as ¥1,000,000 pre-tax plus ¥100,000 JCT at the standard 10% rate. Under the 70% transitional rate, ¥70,000 of the JCT qualifies as purchase tax credit (仕入税額控除, shiire zeigaku kōjo) — this is the amount the buyer can claim against output JCT on its own sales. The remaining ¥30,000 is not creditable and has to land somewhere else in the books.

Where the buyer's ledger posts JCT as a separate line, the split follows the math:

• Creditable ¥70,000 routes to the JCT-receivable account (仮払消費税 / input JCT), available to offset output JCT at return time.
• Uncreditable ¥30,000 routes into the cost of the purchased goods or services — the accounting-treatment options for how that lands are covered in Step 3.

Where the buyer records JCT inclusive in the expense line (a tax-inclusive accounting policy, 税込経理), the whole ¥100,000 sits in the expense total and the creditable portion is only separated out at return preparation. Both policies are in active use; what matters is that the workflow applies the same policy consistently across non-registered suppliers and carries the ¥30,000 uncreditable portion to the right home.

The reduced JCT rate (8%, applied to food items, subscription periodicals, and certain other categories) works the same way. An invoice subject to the 8% rate in FY2027 carries ¥80 of JCT per ¥1,000 of pre-tax purchase; ¥56 is creditable, ¥24 is not. Only the base rate differs; the transitional percentages and the transaction-date rule apply identically.

Step 3: Book the Uncreditable JCT Portion in Your Ledger

Once Step 2's credit math has split the JCT on a non-registered supplier's invoice into a creditable portion and an uncreditable one, the uncreditable portion needs a home in the ledger. Practitioners use one of three approaches, and each carries a real trade-off against the others. None is "the right answer" in isolation — the fit depends on the accounting system in use and on how much per-supplier JCT-leakage visibility the finance team actually wants.

The simplest approach is gross-in-expense — book the entire purchase, including the uncreditable JCT, into the expense or COGS line. Under the ¥1,000,000 / ¥100,000 example from Step 2, the buyer posts ¥70,000 to the JCT-receivable account and ¥1,030,000 to expense (the ¥1,000,000 pre-tax amount plus the ¥30,000 non-creditable JCT absorbed into cost). The method requires no chart-of-accounts changes and aligns naturally with a tax-exclusive books policy. Its cost is visibility: the expense line mixes real cost and JCT leakage, and per-supplier leakage analysis later on has to be reconstructed from invoice-level detail rather than read off the ledger.

The visibility-preserving alternative is a separate JCT-leakage sub-account: route the creditable portion to JCT-receivable as before and the uncreditable portion to a distinct expense sub-account — a line often named something like "JCT non-creditable — tax-exempt supplier" (仕入税額控除対象外消費税 or a company-specific equivalent). In the same example, ¥70,000 goes to JCT-receivable, ¥1,000,000 to the purchase or expense account, and ¥30,000 to the dedicated leakage sub-account. Per-supplier and per-period JCT leakage becomes a ledger query rather than a reconstruction. The cost is setup: the accounting system has to support the sub-account, and the posting rule needs to branch on the vendor master's JCT-status flag so that tax-exempt supplier invoices route differently from qualified-issuer invoices automatically. Not every smaller accounting package handles that branching cleanly.

A middle path is memo-tracking by supplier — post gross-in-expense in the main ledger but maintain a per-supplier memo total of JCT leakage in a separate schedule, whether a spreadsheet, a reporting view, or an AP-platform field. In the example, the ledger treatment matches the gross-in-expense case, but a memo schedule records ¥30,000 of JCT leakage against this supplier in this period. This splits the difference: visibility without chart-of-accounts changes. It depends on the discipline to maintain the memo schedule, and on anyone who reads it understanding that the ledger expense line includes the leakage while the memo breaks it out separately. Memo tracking often earns its place as a bridge — the buyer starts here, then migrates to the dedicated sub-account once the finance team has confirmed per-supplier visibility is worth the chart-of-accounts work.

Choosing between them comes down to two questions. What does the accounting system actually support? If the general ledger cannot branch posting on a vendor-master flag, the dedicated sub-account needs a workaround (manual journal entry per supplier, or a purchase-journal pre-processor) that may not be worth the effort for the volume of tax-exempt purchases involved. How granular does the finance team's reporting need to be? A team that plans to run the continue-or-switch decision in Step 6 on real per-supplier leakage data needs the sub-account or, at minimum, disciplined memo tracking. A team treating the leakage as undifferentiated cost of doing business in Japan can stay with gross-in-expense and pick the leakage out of invoice-level detail if a specific question comes up later.

One operational note cuts across all three: whichever method is chosen needs to be documented in the accounting manual and applied consistently across suppliers and across periods. Switching methods part-way through a tax period, or applying the sub-account treatment to some tax-exempt suppliers and gross-in-expense to others, makes the tax return harder to prepare and undermines any per-supplier analysis that depends on comparable data. Pick the method that fits the system and the reporting need, write it down, and apply it uniformly.

The uncreditable JCT portion is not the only Japan-specific ledger complication these invoices might carry. Services that fall under the withholding-tax regime — Japan withholding tax (gensen choshu) on supplier invoices is a separate treatment AP needs to apply to certain professional-services payments — interact with this workflow at the invoice-line level without overlapping the JCT split above.

Step 4: Track the ¥100 Million Per-Supplier Cap

The FY2026 tax reform introduced a hard ceiling on how much of a tax-exempt supplier's volume can even qualify for the transitional credit. According to EY Japan's 2026 tax reform alert on consumption tax changes, when purchases from one non-qualified invoice issuer exceed ¥100 million in a taxable period, the exceeding portion cannot apply the input tax credit transitional measures — effective for tax periods starting October 1, 2026.

Two points of scope matter before walking the math. The ¥100 million threshold is tracked per tax-exempt supplier — aggregate purchases across all non-registered suppliers do not add up against a single cap; each supplier has its own. And it is tracked per taxable period (課税期間) — the counter resets at the start of each new period. For companies whose fiscal year aligns to the calendar, the first period in which the cap applies is the one beginning January 1, 2027. For companies whose fiscal year ends in a different month (a March 31 fiscal close is common in Japan), the effective date is "the first taxable period beginning on or after October 1, 2026," which may be a period that started before most people think about it. Read the effective date as a period-start condition, not a calendar date.

A ¥150 million annual spend on a single tax-exempt supplier in a taxable period inside FY2027 works out as follows. The buyer pays ¥150 million pre-tax plus ¥15 million JCT at the standard 10% rate — ¥165 million total. The cap slices that into two tranches:

• First ¥100 million pre-tax (below the cap). The ¥10 million JCT on this tranche is subject to the 70% transitional rate that applies in FY2027. ¥7 million is creditable as input tax; ¥3 million is not creditable.
• Remaining ¥50 million pre-tax (above the cap). The ¥5 million JCT on this tranche cannot apply the transitional measure at all. Zero is creditable; the full ¥5 million is not creditable.

Totalling across the period: ¥15 million JCT paid, ¥7 million creditable, ¥8 million of JCT leakage against this one supplier. The same purchase from a qualified invoice issuer would have produced ¥15 million creditable and zero leakage — the cap's impact on a concentrated relationship is a concrete ¥8 million annual cost at these volumes, not a schedule-percentage abstraction.

The cap is a step function, not a gradient. The transitional rate applies fully up to the cap; above the cap, it falls immediately to zero on the excess. There is no partial credit on purchases over the threshold and no proration. This matters for the Step 6 decision framework: the cost of continuing with a non-registered supplier who falls below the cap ramps gradually through the taper (70% to 50% to 30% to 0%), while the cost of continuing with one who crosses the cap jumps sharply at the crossover point and stays elevated.

Running the cap correctly requires a control that earlier AP setups did not need: a YTD pre-tax-purchase counter for every tax-exempt supplier, tied to the vendor-master JCT-status flag from Step 1, resetting to zero at each taxable-period boundary. Each posted invoice on a tax-exempt flagged supplier adds its pre-tax amount to that supplier's counter. When the counter reaches ¥100 million, subsequent postings in the same period apply the 0% rate on the excess portion of each invoice — the portion below the counter's crossover retains the scheduled transitional rate; the portion above gets no credit. Invoices that straddle the crossover need to be split mechanically at the cap boundary.

Feeding the counter reliably depends on extracting the pre-tax amount, vendor identity, transaction date, and JCT status from every Japanese invoice and routing those into the AP system or the downstream tracker that holds the running total. Prompt-based extraction tooling — the kind of system where the user specifies "extract invoice number, transaction date, vendor name, vendor T-number, pre-tax amount, JCT amount, and total, one row per invoice" in a single prompt and gets a structured output across the whole batch — covers the data-capture side cleanly. The AP ledger or a reporting layer on top of it holds and resets the counter; the extraction layer ensures no tax-exempt-supplier invoice is booked without the amount landing in that counter. For buyers where concentrated tax-exempt spend exists, this is an operational workflow decision to take before the first taxable period under the cap, not after.

Buyers with no single-supplier exposure anywhere near ¥100 million can still track the counter as a hygiene control — the cost of running the control is low, and suppliers grow. Buyers with concentrated relationships (large single-category spend, long-term subcontractors, major recurring service providers) have to run it from day one of the first capped period or the period's return will be wrong.

Step 5: Run the Supplier Registration Conversation

At some point with a non-registered supplier — usually when a quarterly review surfaces the JCT leakage, or when an annual price negotiation is coming up, or when a counter crossing toward the ¥100 million cap makes the economics sharper — AP or procurement has the conversation about registration. The point of the conversation is to share information and understand the supplier's position, not to demand registration. Handled well, the call either opens a path to the supplier becoming a qualified invoice issuer (適格請求書発行事業者) or confirms the reasons they will remain tax-exempt, both of which inform the decisions in Step 6.

A usable skeleton runs through five moves.

1. Explain the credit-position shift. State the current creditable percentage on their invoices and the next step-down plainly — "in our current fiscal year we can claim 70% of the consumption tax on your invoices; from October 2028 that drops to 50%, and from October 2030 to 30%." Many small suppliers have not been told what the buyer-side math looks like on their invoices. Starting with the facts, without a pitch, sets the tone of the call as informational.

2. Ask whether they have considered registering. This is an information-gathering question, not a persuasion opener. Many suppliers — especially ones that have dealt with other corporate buyers since 2023 — have read NTA material, talked to their tax adviser, and arrived at a considered position. Some have decided not to register because the economics of becoming JCT-taxable (課税事業者, kazei jigyōsha) outweigh the buyer-relationship cost for their particular revenue mix. Knowing their position before proposing anything saves time on both sides.

3. Acknowledge the threshold context. If the supplier's annual taxable sales are below ¥10 million (the menzei-ten, 免税点 — the default threshold below which Japanese enterprises are exempt from JCT-taxable status), registering as a qualified invoice issuer also elects them into JCT-taxable status. That is a real cost: they take on JCT filing obligations and lose the exemption on their own output tax. For a small single-proprietor or family business, that trade-off is not trivial. Naming the context explicitly — "we understand registering would mean electing into JCT-taxable status, and we know that is a real change for a business your size" — shows the buyer has done the work to understand the other side's position.

4. Offer practical support. Point them to the NTA registration portal for the actual filing; offer to share what paperwork the process requires based on the buyer's experience with other suppliers; ask whether an introduction to a tax professional who handles qualified-issuer registrations would help. Practical support is often the difference between a supplier who considers the idea and one who acts on it.

5. Be candid about the business-impact dimension. The supplier needs to understand what the buyer will do if they remain non-registered, and vague language here wastes the conversation. The realistic options are absorbing the JCT leakage for now, renegotiating the unit price to offset some of the leakage, or — when the economics become material and alternatives exist — considering a shift to a qualified-issuer supplier. State which options are on the table honestly. "We plan to keep working together at current terms" is a perfectly good answer if it is the truth; "we will need to revisit pricing at year-end given the 50% step-down" is another. Euphemistic softening ("we may need to explore various options") leaves the supplier unable to plan.

The conversation plays out very differently across supplier contexts. For a small single-proprietor or family business whose revenue genuinely sits below the ¥10 million menzei-ten, the call is a real negotiation about whether voluntary JCT-taxable registration preserves the buyer relationship at acceptable cost to the supplier. For a mid-size supplier whose revenue is already above the threshold but who simply has not registered yet, the ask is much simpler — they are already JCT-taxable, registration adds no new tax burden, and the only question is why they have not done it. Knowing which context applies before the call determines how much of the skeleton above is load-bearing versus ceremonial.

One practical opening line that adapts to most situations — in practice the call will often be in Japanese, but the structure translates directly — runs something like: "I wanted to check in with you about how we are handling your invoices under the invoice system. The transitional credit on non-registered supplier invoices is stepping down again in October, and before that happens I wanted to make sure we are aligned on what that looks like on our side and understand where you are on the registration question." That gets to the subject cleanly, frames the purpose as information-sharing, and opens the floor to the supplier's view without leading.

Nothing in this skeleton is meant to be read verbatim. Adapt the specifics to the relationship, the language of the call, and the history between the firms. The structure is what carries — a call that skips the credit-position explanation or the menzei-ten acknowledgment typically fails to produce a useful supplier response, regardless of wording.

Step 6: Decide Whether to Continue or Switch

With the leakage visible on the ledger, the cap math modelled, and the supplier's registration position known from Step 5, the continue-or-switch decision comes down to a comparison: the total cost of continuing with the non-registered supplier through the taper and beyond versus the total cost of switching — including switching costs. The framework below is about where to spend judgment, not about mechanizing the decision.

The cost-of-continuing inputs are concrete. For a given supplier, project:

• Annual pre-tax spend through each tranche the article is still in — typically the current period's volume held roughly constant, adjusted for any known business-driven change.
• The scheduled transitional credit rate in each period: 70% through September 2028, 50% through September 2030, 30% through September 2031, 0% thereafter.
• The ¥100 million per-supplier cap check per period, applying zero credit on the excess above the cap where it applies.

Multiplying spend by (1 − credit rate) × JCT rate gives the leakage per period; summing through the taper gives the cumulative cost of staying with this supplier as-is.

Compare against the cost of switching. A serious switching-cost estimate covers onboarding a new supplier (time, integration, any one-time setup), any price differential between the non-registered supplier and qualified-issuer alternatives (which may cut either way — qualified issuers are not automatically more expensive, but a non-registered supplier in a specialized category may have been cheaper for non-trivial reasons), quality and reliability risk during the transition, and the relationship cost with the incumbent supplier including any downstream effects on the broader vendor relationship.

The decision shape differs sharply by relationship size, and the framework's usefulness depends on treating the three cases differently:

• Long-tail small-vendor noise. Occasional contractors, utilities, a taxi receipt that arrived on a corporate card, a one-off repair. The annual JCT leakage across the whole taper on a ¥200,000-a-year supplier is tens of thousands of yen — less than the hour of analyst time it would take to seriously evaluate alternatives. Switching costs dwarf any realistic saving. The right treatment is to carry the leakage, review these suppliers as a bucket at the annual portfolio review (Step 7), and not let them consume cost-management attention. Do not overthink them.

• Mid-size vendor relationships with genuine switching options. Annual spend of ¥10 million to ¥50 million on a non-registered supplier in a category where qualified-issuer alternatives exist. Leakage over the taper runs into real money; switching costs are bounded because alternatives are known; the decision turns on the relative economics after modelling, the quality implications, and what the supplier said in the Step 5 conversation. This is where a proper comparison earns its keep, and where decisions can move either way with equal credibility depending on the inputs.

• Concentrated vendors crossing the ¥100 million cap. The Step 4 worked example showed an ¥8 million annual leakage on a ¥150 million spend in the 70% tranche; the same volume in the 50% tranche produces materially more. The math is worse than the schedule-percentage alone suggests, and the decision is sharper. The realistic paths are that the supplier registers, the buyer absorbs the growing leakage, the buyer renegotiates price to offset some of it, or — where category economics allow — the buyer splits volume across multiple qualified-issuer suppliers to bring each below the cap. The comparison to qualified vs non-qualified supplier cost in Japan is not abstract for these relationships; it is the central decision.

Cost is a major input, not the only one. A strategic supplier with a strong quality track record at a material cost penalty may still be the right answer. Supply-chain resilience, contractual commitments, and relationship weight all have legitimate roles in the decision. The framework's job is to surface the cost clearly so the decision is informed rather than silent — teams that skip the leakage modelling often end up continuing with expensive non-registered suppliers not because they have judged the cost acceptable but because nobody has quantified it.

Review cadence matches relationship weight. Long-tail suppliers get an annual look as a bucket. Mid-size relationships get a semi-annual review so that transitional-rate step-downs do not silently degrade the economics between decisions. Concentrated suppliers approaching or above the ¥100 million cap get reviewed at each taxable-period close — the cap's step-function effect makes material movements between periods possible, and the decision on what to do about a large tax-exempt supplier is not one to leave twelve months between reviews.

The framework produces different outputs for different relationships, and that is the point. A playbook that recommends the same treatment for a ¥50,000 occasional contractor and a ¥150 million key supplier is not giving anyone useful guidance; the discipline here is applying the right depth of analysis to the right tier and moving on.

Step 7: Monitor Vendor Status Over Time

The setup and decision work in Steps 1–6 produces a workable state. Without ongoing monitoring, that state silently decays — suppliers deregister, counters roll incorrectly across period boundaries, vendor lists drift, and the transitional schedule keeps stepping down behind decisions that were made against an earlier rate. Four controls keep the workflow current.

Re-verification cadence. Qualified-issuer status is not permanent. Match the NTA registry re-check cadence to supplier volume and margin weight: quarterly batch checks are enough for most of the vendor list; on-invoice-arrival verification is warranted for high-volume or thin-margin suppliers where a missed deregistration would quietly cost meaningful amounts across even a few weeks of postings. Domestic Japanese AP platforms typically fold this lookup into the invoice intake workflow — Money Forward Cloud Accounts Payable qualified-invoice number checks is one example for buyers running that stack — while ledger-only workflows rely on a scheduled batch job to reconcile each active tax-exempt and qualified-issuer vendor against the current registry state.

Deregistration watch. When a supplier transitions from qualified-issuer to tax-exempt, record the deregistration date on the vendor master and route invoices by transaction date, not by posting date. An invoice for services delivered before the deregistration date retains its qualified-issuer treatment; an invoice for services delivered on or after the deregistration date moves into the non-registered workflow from Step 2 onward. Mistakes here typically run in the same direction — an AP clerk sees "qualified" in the vendor history, posts as qualified, and claims full credit on invoices whose transaction dates fall after the deregistration. Reviewing invoices against the deregistration date at posting time is the cheapest way to catch these before they reach the return.

YTD counter reset at taxable-period boundaries. The ¥100 million per-supplier counter from Step 4 resets to zero at the start of each new taxable period. Build the reset into the AP-system workflow so it happens automatically at period close; manual resets are error-prone and quietly under-credit or over-credit across the period boundary. Confirm the reset actually happened at the first posting of each new period — a counter that silently failed to reset shows up as disproportionate leakage in the early weeks of the new period.

Annual tax-exempt supplier review. Once a year — typically at fiscal-year planning, separately from the workflow-level cadences above — walk the full tax-exempt supplier list. For each supplier, check three things: is the JCT-status flag still correct (some will have registered since the last review), has the spend weight changed enough to shift the supplier between the long-tail, mid-size, and concentrated tiers from Step 6, and does the continue-or-switch judgment still hold at the current transitional rate. The review also picks up suppliers who have fallen out of active use — dormant vendor records accumulate otherwise and muddy the per-period analysis.

Treat monitoring as the half of the workflow that keeps the other half honest. A well-designed Step 1 flag, a correct Step 2 rate, a disciplined Step 3 posting rule, and a running Step 4 counter all fail silently if the ongoing controls above do not run — and silent failures are the kind that surface at tax-return preparation, which is the worst time to find them.