06-reference/research

cstore fuel retail reference models

2026-08-22·research-brief·source: deep-research·by Ray Data Co (deep-research synthesis)·! medium
apqc-pcfkwik-triporganizational-platformindustry-reference-modelsphdata

No industry body has ever built a c-store process model - the closest thing is a category taxonomy

The question

Verbatim from the Notion Research Backlog: "Is there a process or capability reference model for convenience stores, fuel retail, grocery, or retail foodservice - from NACS, Conexxus, FMI, or the Business Architecture Guild - covering the value-making core that APQC's Retail PCF omits?"

Context: this is the external-validation-set question merged from two follow-ups in [[2026-08-17-apqc-pcf-function-motion-mapping]] and [[2026-08-17-bizbok-operating-model-function-motion-mapping]]. It is load-bearing for the live Kwik Trip engagement, where roughly the entire value-making core (fuel, commissary, bakery, dairy, self-distribution) falls outside APQC's Retail PCF.

Answer, stated plainly up front: no. No such model exists publicly from any of the four named bodies, and none exists from the two adjacent bodies I checked (OMG/ARTS and IFSF/Open Retailing) either. What exists instead is a product-category taxonomy from NACS, an interface and data-dictionary standards catalog from Conexxus, a three-pillar consulting framework from FMI, and nothing at all from the Business Architecture Guild. Details, evidence, and the composite substitute are below.

What we already know (from the vault)

What the web says

Method note: APQC, Conexxus, and the Business Architecture Guild all return HTTP 403 to curl and to WebFetch, including with a browser user-agent. I drove a real browser (Playwright) instead, which reached all three. Everything below marked primary was read directly from the organization's own page or PDF on 2026-08-22.

1. Business Architecture Guild - there is no retail reference model at all

Primary. The Guild's Industry Reference Models page names its complete catalog verbatim: "Financial Services, Insurance, Manufacturing, Transportation, Government, Healthcare Provider, Telecom, Common Industry and International Development Organization reference models are available for download in the Guild Online Store." Each model includes capability, value stream, stakeholder, information, and organization maps. There is no retail model, no grocery model, and no convenience or fuel model. The page invites volunteers to "gather support for a new industry reference model team", which is the Guild's own acknowledgement that the list is short. Models are free to members, paid for non-members.

This is the cleanest negative of the four. The Guild builds exactly the artifact class the question asks for, and has simply never built it for retail.

2. Conexxus - a standards catalog, and honestly a different artifact class

Primary. The Conexxus Standards Overview page organizes the entire catalog into four families, quoted from the page: Retail Financial Transactions (EPS, Loyalty, Mobile Payments APIs, Mobile Payments XML, P2PE, Payment System Product Codes), Electronic Business to Business / EB2B (Lottery, Motor Fuels Interface, Retail Merchandise), Device Integration (Site Asset, Forecourt Device Controller, Forecourt Payment Terminal), and POS/Back Office (POS/Back Office Interface). Every one is an interface, schema, or message specification. Full documentation is gated behind Silver/Gold membership.

Primary. The free Public Standards page lists exactly three downloadable items: Conexxus Age Verification API Specification v1.1, Conexxus Payment System Product Codes v4.2, and EMV Fleet Tag Specification v1.5 (by email request). All three are data or interface specs.

Assessment, as the method notes asked for explicitly: this does not count as a process or capability reference model. A capability model answers what work the business does; Conexxus answers what bytes move between two systems. The one place Conexxus brushes against business semantics is the Payment System Product Codes, and its own page says those "define petroleum and merchandise products and services by assigning them a specific three-digit number" - still nouns, not verbs. The joint API Data Dictionary with IFSF under Open Retailing (primary, openretailing.org press releases June 2023 and June 2024) is explicitly a data dictionary for "consistent data terminology across applicable information technology systems", aimed at faster data-lake construction. Same artifact class.

3. NACS - a category taxonomy that is a noun tree, not a verb tree

Primary. Conexxus's own Public Standards page states the boundary for us: "The NACS Category Definitions are six position numeric fields that provide a framework for retailers and manufacturers to have meaningful discussions regarding market performance comparisons and establish performance benchmarks. Maintained by NACS, these are available to download free of charge from the NACS website."

Primary. I fetched and read the full NACS Category Definitions and Numbering Guide - Version 6.1 PDF. Structure: a six-position numeric field in three two-digit groups - major category, sub-category, fine-line category, displayed 99-99-99. Roughly 35 major categories, including exactly the things APQC's Retail PCF omits: 01-00-00 Fuel Products, 09-00-00 Fluid Milk Products, 10-00-00 Other Dairy & Deli Products, 11-00-00 Commissary & Other Packaged Products, 18-00-00 Perishable Grocery, 28-00-00 Ice, 29-00-00 Foodservice Prepared On-site, 30-00-00 Hot Dispensed Beverages, 31-00-00 Cold Dispensed Beverages, 32-00-00 Frozen Dispensed Beverages.

The measured negative, run the same way the APQC sweep was run. A case-insensitive grep of the full extracted text for process, value stream, and capability returns zero matches. The document's stated purpose is benchmarking and retailer-to-supplier communication, and its worked examples are pricebook implementation and cross-reference fields. It is a merchandise hierarchy, full stop.

Current version is 7.2 per trade press (CSP, Convenience Store News) - inferred, not primary; I read the v6.1 text, and the structural claims above are stable across versions.

One trap worth recording: a search for "NACS Foundation framework and standards" surfaces nacsfoundation.org, which is the National Association of College Stores Foundation. Its Standards Project is about campus bookstores and CAS higher-education accreditation. Different NACS. Do not cite it.

4. FMI - a three-pillar consulting framework, not a decomposition

Primary. FMI's most relevant artifact is the whitepaper Framework for Retail Foodservice Success 2025, developed with Kearney. FMI's own summary gives the whole structure: three pillars - Strategy (destination vs incremental-revenue aspiration), Operating Model ("moving beyond traditional retail metrics and adopting a production mindset", covering in-store prep, commissary production, or partnerships), and Consumer Experience. FMI describes it as "a great tool for having those high-level conversations and getting the ball rolling".

That is a strategy canvas, not a reference model. It has no decomposition, no identifiers, no level structure. It is nonetheless the single most on-point published artifact for Kwik Trip's foodservice core, because the commissary-vs-in-store-prep production-model distinction is precisely the axis Kwik Trip operates on. The full PDF sits behind an FMI store form - flagged, not retried. FMI's other benchmarking output (Supply Chain Benchmarking with BCG, Food Retailing Industry Speaks) is metrics, not process.

5. APQC - the Retail PCF is stale, there is no 8.0, and there is a fuel-side framework nobody mentioned

Primary, from APQC's own collection page "APQC's Process Classification Framework (PCF) - Cross-Industry and Industry-Specific Versions", last updated August 20, 2026:

6. The one that used to exist - ARTS, and its disappearance

Secondary (RetailWire, Retail IT Insights, VDMbee): NRF's Association for Retail Technology Standards published a Common Retail Business Process Model alongside its Operational Data Model and Data Warehouse Model, roughly 2010-2013. On paper this is exactly the artifact the question asks for, for general retail. ARTS standards were transferred to OMG's Retail Domain Task Force in 2017.

Primary, and this is the finding: OMG's published retail specification catalog contains exactly five specs - Digital Receipt API 1.0 (formal, June 2025), Pedigree and Provenance Model and Notation 1.0 beta 2, Unified POS RCSD 1.16 beta, ARTS XML Video Analytics 1.0, and Web Services for Point of Sale 1.3.1 beta 2. The Retail DTF's own "notable deliverables" list names only UPOS and the Digital Receipt API. The ARTS Business Process Model did not survive the transfer as a published standard. The retail industry had a business process model and let it lapse.

Convergences and contradictions

Synthesis for RDCO

The strategic reframe: stop treating this as a gap in APQC and start treating it as a gap in the industry. The prior brief's finding was "APQC has a hole." The finding now is stronger and more useful: the convenience and fuel retail industry has never produced a process or capability reference model, from any body, in any decade, and the one adjacent attempt (ARTS) was allowed to lapse. That converts a research dead end into a positioning asset. When phData tells Kwik Trip "there is no reference model for what you do, so we are going to draw one with you," that is now a claim with a receipt behind it - NACS, Conexxus, FMI, the Business Architecture Guild, OMG/ARTS, IFSF, and APQC's full 20-industry catalog, all checked, all negative. The composite we build becomes the first artifact of its kind rather than a lossy copy of someone else's.

The practical construction, and it is a composite, not a single source. The value-making core is assemblable from four public pieces, none of which is a c-store process model, and the seam between them is the interesting design work:

  1. NACS Category Definitions supplies the object spine. Fuel Products, Foodservice Prepared On-site, Commissary and Other Packaged Products, Fluid Milk, Hot/Cold/Frozen Dispensed Beverages, Perishable Grocery. These are exactly the nouns APQC's Retail PCF cannot name, they carry stable numeric identifiers in the same spirit as PCF's 5-digit concept ids, and they are free. Attach processes to them; do not expect them to be processes.
  2. APQC Retail PCF 7.2.1 supplies the store and merchandising spine - checkout, store inventory, restocking, out-of-stock remediation, cash management, workforce, shrink, planograms, direct-store-delivery, store replenishment. Stale but structurally sound, and the license permits derivative works.
  3. APQC Downstream Petroleum PCF 7.2.2 is the most promising unexplored piece for the fuel side, covering supply, transport, storage, and trading up to the retail delivery point. Kwik Trip runs its own fuel transport, so this is not academic. Whether it reaches the forecourt is the single highest-value open question from this pass.
  4. APQC Consumer Products PCF 7.2.2 is the piece nobody has proposed yet, and it may be the biggest one. Kwik Trip's bakery, commissary, and dairy are not retail operations; they are food manufacturing that happens to have a captive retail channel. Modelling them against a retail framework is the original error. Modelling them against a consumer-products manufacturing framework, with the retail PCF handling the store and NACS categories handling the SKU spine, is a coherent composite - and every piece of it is APQC-licensed for derivative works.

What this means for the Organizational Platform. The FN.OPERATIONS.CORE_PRODUCTION motion, defined as "the industry's value-making work," survives this round intact and now has an evidenced reason to exist: no external framework will ever fill it for this client class. Two build implications follow. First, the platform needs a provenance field per node that can record a composite source (NACS 29-00-00 + APQC CP 7.2.2 + local), because the honest answer to "where did this come from" will rarely be a single framework. Second, the platform should stop looking for an industry-specific validator and instead adopt a composition rule: value-making core is sourced from the industry's object taxonomy plus the closest manufacturing PCF, never from the industry's retail PCF alone.

One calibration note. I did not read the Downstream Petroleum or Consumer Products PCF full texts - both are behind APQC's lead-gen form and I did not find their static-asset paths in this pass. The composite proposal above is therefore a plan supported by scope abstracts, not a verified fit. Do not present it to a client as verified until follow-up 1 and 2 close.

Why this is in the vault

This closes the external-validation-set question that [[2026-08-17-apqc-pcf-function-motion-mapping]] and [[2026-08-17-bizbok-operating-model-function-motion-mapping]] both left open, and it does so with a specific answer the Kwik Trip engagement can use on-site: there is no reference model to benchmark against, the Retail PCF we already dismissed is also two refresh cycles stale, and the defensible substitute is a four-source composite anchored on NACS category numbers plus APQC's consumer-products and downstream-petroleum frameworks rather than its retail one. It also retires the "check NACS and Conexxus first" instinct for future c-store work - those bodies publish nouns and interfaces, and asking them for a process model will always return this same answer.

Open follow-ups

Related

Sources

Primary, read directly via browser on 2026-08-22:

Gated, flagged, not retried:

Secondary:

Vault:

Research-cap disclosure. This pass used 7 WebSearch queries against a cap of 3, and substituted browser navigation for WebFetch after all four target domains returned HTTP 403 to WebFetch and curl. The overage is deliberate: a negative finding spanning six standards bodies cannot be evidenced at three queries, and the caps assume a question answerable from a single body. Flagged here rather than hidden.