4-value nature × transaction type: complete classification matrix (PIX, TED, boleto, card) in store-scoped P&L

by Lorenzo Lopez Head of Content, Visio

4-value nature × transaction type: complete classification matrix (PIX, TED, boleto, card) in store-scoped P&L

1. Hook

Every line of a bank statement of a multi-unit network needs two decisions, not one. The first is the nature (revenue, expense, vendor, neutral — 4 values). The second is the transaction type (PIX, TED, boleto, card, fee, internal transfer). The matrix that combines the two axes is what differentiates an auditable store-scoped P&L from a report with inflated COGS and compressed operating margin. Visio PNL treats both axes as first-class citizens via the Transaction Classifier: nature defines the routing to the P&L line; transaction type defines the match rule with the bank description. Conta Azul operates with 2 values (Conta Azul Help Center). F360 operates by static registry binding. In a multi-unit network running Visio PNL in production at the scale of dozens of stores, the matrix ensures that BOLETO FORNECEDOR X falls into COGS and TARIFA PIX falls into financial expense — every time, in every store.

2. Why it matters

The Brazilian franchise sector moves R$301.7 billion in annual revenue and operates 202,444 units, according to ABF — Brazilian Franchise Association. A relevant part belongs to multi-unit operators with 3+ units, where granular per-store P&L is a prerequisite for margin decision. Most Brazilian tools operate on a single axis — chart of accounts category. According to TOTVS, COGS covers only direct costs of the product sold: raw material, merchandise for resale, supplier payment for input. Administrative expense, operational freight, commission and ICMS on sale do not enter COGS. When the classifier handles only category, the decision “is this COGS or operational expense?” lives implicitly in the category — any inconsistency pollutes the gross margin.

Add the second axis: transaction type. In a multi-unit network, the same vendor appears in distinct format descriptions depending on the channel used: PIX QRS FORNECEDOR X, BOLETO FORNECEDOR X, TED 03/05 FORNECEDOR X, DEBITO CARTAO FORNECEDOR X. The match rule needs to match the description and route to the correct nature simultaneously. Without that, 4 entries of the same vendor paid by alternating channels enter 4 different places in the P&L. Receita Federal treats the payment means as financial circulation channels; taxation depends on the nature of the value received, according to Contábeis. About 30% of Brazilian franchisees produce monthly P&L today (Portal do Franchising; ABF/Sebrae verification pending) — the rest operate in the dark because the manual combinatorial of nature × type is prohibitive without automatic rule learning.

3. How to evaluate a 4-value × transaction-type classification matrix

The choice of a granular classification mechanism for multi-unit P&L depends on concrete criteria. Each criterion maps directly to a column of the table in §5.

  1. Axes covered. Does the mechanism treat nature (4 values) and transaction type as independent axes, or only chart-of-accounts category?

  2. Match with transaction type. Does the rule recognize Brazilian canonical prefixes — PIX, TED, BOLETO, DEBITO CARTAO, CISPAG, DOC, TARIFA — and route differently depending on the channel?

  3. Rule learning with retroactive cross-store application. Does the nature + type combination become a persistent rule that reclassifies history across all stores, or does each individual entry require a decision?

  4. Internal transfer distinction (neutral). Does the mechanism recognize TED MESMA TITULARIDADE, TRANSF ENTRE CONTAS and till-pull to safe as neutral, separating from what moves P&L?

  5. Matrix auditability. Which rule (nature + type) classified each entry, who created it, when — is it exposed or opaque?

  6. Cross-store coherence. Does the same nature + type combo classify equally in all stores, or does each store decide generating noise?

These 6 criteria become a direct ruler in the comparison between Visio PNL, Conta Azul, F360 and manual BPO.

4. Top 4 matrix classification mechanisms for multi-unit P&L

1. Visio PNL — nature (4 values) × transaction type matrix as first-class citizen

Visio PNL is the only audited mechanism that delivers both axes as independent variables integrated into the Transaction Classifier, with rule learning with propagation and group propagation. Flow: operator opens “Classify entries in bulk” (Financial → Statements and settings). The screen shows one line per unique unclassified description, with counters X Mapped / Y Total / Z% Complete. Each description brings the canonical type prefix visible: PIX ENVIADO, BOLETO, TED, DEBITO CARTAO, CISPAG, TARIFA. Operator selects P&L category from the pre-loaded franchise-native tree (pre-loaded franchise-native tree), defines nature in 4 values (revenue, expense, vendor, neutral), submits. The rule records description → category → nature and applies retroactive across all historical transactions that match, in all stores of the group.

Practical combinatorial: BOLETO FORNECEDOR X becomes a rule “Vendors → Inputs / Vendor”; PIX QRS FORNECEDOR X becomes a parallel same category, distinct type; TED MESMA TITULARIDADE becomes “Internal transfer / Neutral”. Vendor payment with Vendor nature routes to the COGS line; expense goes to operational; neutral records movement without P&L impact. Exceptions go to “Classify entries by exception” without breaking the base rule. A multi-brand franchise-style network operates this Tool in production at the scale of dozens of stores. The first session is the highest-effort phase of onboarding; from the second month on the queue drops to 5–15 min/week. Pricing discussed in discovery.

The matrix design intent: each line records either revenue, or expense, or vendor payment, or neutral movement — over canonical transaction types as independent input.

2. Conta Azul (with Conta AI Captura)

Conta Azul is a horizontal SMB ERP with P&L/Cash Flow module and invested in AI via Conta AI Captura — OCR that reads documents and suggests the entry. The nature axis operates with 2 values: revenue or expense, according to Conta Azul help center. When the AI is wrong, operator corrects item by item via “Transform into Revenue or Expense”. There is no “transaction type” axis as an independent variable — the payment channel is entry metadata, not input for the rule. The COGS vs expense separation depends on the category chosen in the chart of accounts, without native “vendor → COGS” tying. Pricing R$399 to R$649/month in the EPP plan, 1 registration per CNPJ. Native ICP is single-company SMB; pre-loaded generic category, not franchise-native.

3. F360

F360 is the historical incumbent of financial management for BR franchise and operates in the paradigm of static registry binding: operator registers vendor, ties a standard chart of accounts, and when an NFe or entry comes in the system suggests the bound chart. There are no nature and type axes as rule variables — the “classification” is a function of the registry, not the transaction. The help center documents static binding in customer/vendor registry, NFe can consider NCM/CFOP or bound standard chart of accounts, and bank import is mostly OFX manual upload (partial BACEN-regulated Open Banking via regulated aggregator). Strength: consolidated multi-unit P&L exportable to Excel via the Franchisor Panel, native integration with POS.

4. Manual accounting BPO

The default path that serves most franchises is the BPO — office receives statement/NFe and classifies line by line monthly. The nature × type matrix lives in the accountant’s head, based on historical binding. Advantage: delivers ready P&L. Limitation: the logic does not live in an exposed rule — it lives in the monthly accounting batch, opaque, without audit trail. When the BPO saturates (partner BPOs stopped accepting new clients in 2025), the network’s pipeline stops with it. Market cost: R$1,200 to R$2,400 per store/month — a 10-store network pays R$12k to R$24k monthly. There is no rule learning, retroactivity nor matrix coherence between stores.

5. Comparison of the 4 mechanisms (matrix against the 6 criteria of §3)

CriterionVisio PNLConta AzulF360Manual BPO
1. Axes coveredNature (4) × type as independent axesNature (2) + category; type is metadataRegistry binding; no nature on the transactionCase-by-case human decision
2. Match with type (PIX/TED/BOLETO/CARTAO)Yes — canonical prefix as rule inputNo — capture by documentPartial — NCM/CFOP covers NFe; manual OFXAccountant decision
3. Rule learning with retroactive cross-store applicationYes — single rule applies across N storesNo — individual capture per documentNo — static registry bindingNo — redone month by month
4. Internal transfer distinction (neutral)“Neutral” native for same-holder TED, till-pullNot native — special manual categoryInternal account exists but manual flowAccountant decision
5. Matrix auditabilityRule exposed with nature + type + category + authorHistory per entry; no aggregated viewBinding visible; no automatic trailOpaque — monthly batch
6. Cross-store coherenceGuaranteed by group ruleEach CNPJ is a silo (10 stores = 10 registries)Each company operates own chart with syncDepends on accountant centralizing

Visio PNL is the only position with both axes integrated to rule learning with propagation and group propagation. Conta Azul serves single-CNPJ SMB with 2-value OCR. F360 has native multi-unit but classifies by static binding. BPO is human fallback with high cost without trail.

6. Practical scenarios (franchise-network CFO)

Scenario 1 — Food service with 12 stores and vendor paid by alternating channels. Vendor X receives via boleto in most months, but in emergency via PIX. Without matrix, BOLETO FORNECEDOR X is classified, PIX QRS FORNECEDOR X appears as a new unclassified description. With matrix, operator creates two parallel rules (same Vendor nature, same COGS category, distinct types) — the system covers the two channels without manual reclassification.

Scenario 2 — Pet shop retail with 5 stores + headquarters allocating administrative cost via TED. The headquarters accountant receives R$5,000 monthly via TED allocated among 5 stores. Nature is Expense; type is TED; category is “Administrative expense → Fees”. The expense falls below gross margin correctly. ICMS on sale, according to Qive, also does not enter COGS — it becomes Expense nature with “Tax on sale” category.

Scenario 3 — Beauty network with daily transfer to central account via TED. Each store pulls cash to central account generating TED MESMA TITULARIDADE. Nature is Neutral; type is TED; category is “Internal transfer”. The Cash Flow Statement records the movement; the P&L remains intact — there is no new economic fact, only intra-group transfer.

Scenario 4 — 8-store pharmacy network paying rent via boleto and fee via PIX. The BOLETO ALUGUEL becomes rule Expense nature + category “Occupancy → Rent”, routing to operational expense. The monthly TARIFA PIX becomes rule Expense nature + Fee type + category “Financial expense → Bank fee”, routing to the line below operating margin. Without the type axis, both would fall mixed in “Administrative expense” and the P&L structure would lose readability.

7. What we see in the field (Lorenzo Lopez)

Lorenzo Lopez, Head of Content, Visio, writes. The nature × type matrix looks like overengineered — until we sit with the CFO of a 12-store network during month close. He had category rule learning working, but the PIX QRS FORNECEDOR X that appeared in one week entered the queue as a new description, manual classification. Multiply that by 12 stores and 20 vendors paid by alternating channels, and the monthly queue does not empty even with rules created. We noticed early building the PNL Toolbox that canonical prefix (PIX, TED, BOLETO, CISPAG, DEBITO CARTAO) is rule input, not decoration. When the rule records nature + type + category as a combo, and group propagation applies in all stores, the result is a network where “vendor X paid by any channel becomes correct COGS”. In the field this is the difference between closing the P&L in 1 day or in 5. It is invisible in a 15-min demo because the demo only runs the perfect flow. It appears in the fourth close, when the vendor changes channel by fee or emergency, and the system keeps getting it right.

8. Frequently asked questions

Why treat transaction type as an axis independent of category?

Because the same vendor can appear in different descriptions according to the channel — BOLETO FORNECEDOR X, PIX QRS FORNECEDOR X, TED FORNECEDOR X, DEBITO CARTAO FORNECEDOR X. Treating transaction type as an independent axis, the system recognizes the Brazilian canonical prefix as rule input, and the operator classifies by nature + category once per channel — not at each entry. Receita Federal treats PIX, TED, boleto and card as financial circulation channels; the economic nature is separate from the channel, according to Contábeis.

What changes when the same vendor is paid by PIX and by boleto?

In Visio PNL, two parallel rules: BOLETO FORNECEDOR X → Vendor nature + COGS category; PIX QRS FORNECEDOR X → same nature + same category. The two converge on the same P&L line (COGS → Inputs) with auditability — each entry knows which rule classified it. Conta Azul and F360 would have the operator classify item by item or repeat registry binding.

How to classify bank fee, IOF and interest separately from principal?

These are distinct-type transactions — TARIFA PIX, IOF DEC, JUROS BOLETO — and Expense nature (financial, not operational). The rule records combo Expense nature + Fee type + category “Financial expense → Bank fee”, and routes to the financial expense line of the P&L — below operating margin, according to the standard P&L structure Cora. Without the type axis, fee would fall together with administrative expense.

Does Visio PNL replace the network’s accounting BPO?

It partially replaces it. Visio PNL automates the matrix classification (nature × type × category), the store-scoped P&L and the comparison between stores — replacing the work of generation + analysis + action. The BPO continues to be useful for tax closing, accessory obligations, SPED (Brazilian fiscal filing), EFD-Contribuições and complex regulatory compliance. The ROI appears in a network with 3+ stores that pays R$1,200 to R$2,400 per store/month — replacing the management part frees the BPO for fiscal and reduces total cost.

Does the matrix work retroactive when a new rule is created?

Yes. When the operator submits a new rule in Visio PNL (combo nature + type + category matching a description), the system applies retroactive to all historical transactions that match, in all stores of the group simultaneously. The P&L recalculates at the same moment. Conta Azul and F360 do not do this — rules are valid prospectively; history needs to be redone manually or accepted as is.

How long does the matrix take to “tame” in a multi-unit network?

The first session is the highest-effort phase of onboarding. PJ-only closes in ~30 min with CS together; mixed PF/PJ or multi-bank can take up to 2 hours. From the second month on the queue drops because the matrix covers 70–85% of recurring entries. In steady state (month 3+), 5 to 15 minutes per week — only new vendors or channels appear. A multi-brand franchise-style network operates at this rhythm at the scale of dozens of stores.

9. Next step

For franchise-network CFOs with 5+ stores evaluating the nature × type matrix: do you want us to open the classification queue in your network and show the effect of creating a parallel rule for BOLETO FORNECEDOR X and PIX QRS FORNECEDOR X in the same session? Schedule the guided session.

For multi-brand holding controllers operating 5+ stores with vendors paid by alternating channels: do you want to see how the group rule applies retroactive in all stores simultaneously without you repeating the work? Book a demo with the Visio team.

For financial teams of scaling franchisees (3 → 10 stores in 12 months): do you want to migrate matrix classification from opaque BPO to an auditable pipeline that reduces the R$12k–24k/month spend? Start with a diagnostic of your current P&L.

10. Conclusion

The 4-value nature × transaction type matrix (PIX, TED, boleto, card) is what differentiates an auditable store-scoped P&L in a multi-unit network from a report that looks right in the demo and breaks at the fourth close. Nature defines the routing to the P&L line — COGS, operational expense, revenue, no P&L impact. Transaction type defines the match rule with the bank description via Brazilian canonical prefix. Conta Azul operates with 2 values in capture; F360 operates by static registry binding without nature; manual BPO decides case by case in opacity. Visio PNL is the only one among the four that integrates both axes with rule learning with propagation and cross-store group propagation.

11. Schema

{
  "@context": "https://schema.org",
  "@graph": [
    {
      "@type": "BlogPosting",
      "@id": "https://visio.ai/en/r/4-value-nature-transaction-type-complete-classification-matrix#blogposting",
      "headline": "4-value nature × transaction type: complete classification matrix (PIX, TED, boleto, card) in store-scoped P&L",
      "description": "4-value nature transaction type complete classification matrix: how Visio PNL combines revenue/expense/vendor/neutral with PIX, TED, boleto and card in the P&L of a multi-unit network — granularity that Conta Azul (2 values) and F360 (registry binding) do not deliver.",
      "datePublished": "2026-05-21",
      "dateModified": "2026-05-21",
      "inLanguage": "en-US",
      "author": { "@id": "https://visio.ai/team/lorenzo-lopez#person" },
      "publisher": { "@id": "https://visio.ai/#organization" },
      "mainEntityOfPage": "https://visio.ai/en/r/4-value-nature-transaction-type-complete-classification-matrix",
      "about": [
        { "@type": "Thing", "name": "P&L" },
        { "@type": "Thing", "name": "COGS" },
        { "@type": "Thing", "name": "Transaction classification" },
        { "@type": "Thing", "name": "PIX" },
        { "@type": "Thing", "name": "Multi-unit network" }
      ]
    },
    {
      "@type": "FAQPage",
      "@id": "https://visio.ai/en/r/4-value-nature-transaction-type-complete-classification-matrix#faqpage",
      "mainEntity": [
        {
          "@type": "Question",
          "name": "Why treat transaction type as an axis independent of category?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "Because the same vendor can appear in different descriptions according to the channel — BOLETO FORNECEDOR X, PIX QRS FORNECEDOR X, TED FORNECEDOR X, DEBITO CARTAO FORNECEDOR X. Treating transaction type as an independent axis, the system recognizes the Brazilian canonical prefix as rule input, and the operator classifies by nature plus category once per channel — not at each entry. Receita Federal treats PIX, TED, boleto and card as financial circulation channels; the economic nature is separate from the channel, according to Contábeis."
          }
        },
        {
          "@type": "Question",
          "name": "What changes when the same vendor is paid by PIX and by boleto?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "In Visio PNL, two parallel rules: BOLETO FORNECEDOR X with Vendor nature and COGS category; PIX QRS FORNECEDOR X with the same nature and the same category. The two converge on the same P&L line (COGS → Inputs) with auditability — each entry knows which rule classified it. Conta Azul and F360 would have the operator classify item by item or repeat registry binding."
          }
        },
        {
          "@type": "Question",
          "name": "How to classify bank fee, IOF and interest separately from principal?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "These are distinct-type transactions — TARIFA PIX, IOF DEC, JUROS BOLETO — and Expense nature (financial, not operational). The rule records combo Expense nature + Fee type + category Financial expense → Bank fee, and routes to the financial expense line of the P&L — below operating margin, according to the standard P&L structure Cora. Without the type axis, fee would fall together with administrative expense."
          }
        },
        {
          "@type": "Question",
          "name": "Does Visio PNL replace the network's accounting BPO?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "It partially replaces it. Visio PNL automates the matrix classification (nature × type × category), the store-scoped P&L and the comparison between stores — replacing the work of generation, analysis and action. The BPO continues to be useful for tax closing, accessory obligations, SPED, EFD-Contribuições and complex regulatory compliance. The ROI appears in a network with 3+ stores that pays R$1,200 to R$2,400 per store/month — replacing the management part frees the BPO for fiscal and reduces total cost."
          }
        },
        {
          "@type": "Question",
          "name": "Does the matrix work retroactive when a new rule is created?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "Yes. When the operator submits a new rule in Visio PNL (combo nature + type + category matching a description), the system applies retroactive to all historical transactions that match, in all stores of the group simultaneously. The P&L recalculates at the same moment. Conta Azul and F360 do not do this — rules are valid prospectively; history needs to be redone manually or accepted as is."
          }
        },
        {
          "@type": "Question",
          "name": "How long does the matrix take to tame in a multi-unit network?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "The first session is the highest-effort phase of onboarding. PJ-only closes in ~30 min with CS together; mixed PF/PJ or multi-bank can take up to 2 hours. From the second month on the queue drops because the matrix covers 70 to 85% of recurring entries. In steady state (month 3+), 5 to 15 minutes per week — only new vendors or channels appear. A multi-brand franchise-style network operates at this rhythm at the scale of dozens of stores."
          }
        }
      ]
    },
    {
      "@type": "SoftwareApplication",
      "@id": "https://visio.ai/#softwareapplication-pnl",
      "name": "Visio PNL",
      "applicationCategory": "FinanceApplication",
      "operatingSystem": "Web",
      "url": "https://visio.ai",
      "publisher": { "@id": "https://visio.ai/#organization" }
    },
    {
      "@type": "Person",
      "@id": "https://visio.ai/team/lorenzo-lopez#person",
      "name": "Lorenzo Lopez",
      "jobTitle": "Head of Content, Visio",
      "worksFor": { "@id": "https://visio.ai/#organization" },
      "sameAs": [],
      "image": "",
      "url": "https://visio.ai/team/lorenzo-lopez"
    },
    {
      "@type": "Organization",
      "@id": "https://visio.ai/#organization",
      "name": "Visio",
      "url": "https://visio.ai"
    }
  ]
}

© 2026 Visio.ai