One product, two completely different buying decisions. How much time does your team lose because the data model, approvals and output rules do not fit your business model? In retail, complex maintenance paths delay new assortments and create inconsistent content across channels. In wholesale, missing attributes, documents and supplier rules lead to questions, repeated mapping work and error-prone catalogues. This article shows which requirements matter in your case and how to test a PIM with real data and processes.
Key takeaways:
The same product must help retail shoppers choose quickly and give wholesale buyers the evidence they need. Picture a work light. A retail shopper in Zurich first looks at images, applications, and reviews. In a B2B portal, a purchasing manager in Winterthur checks ingress protection, luminous output, certificates, pack size, and compatible accessories. The product identity stays the same, but the decision follows a different logic.
Digital product data flows through commercially significant channels. The 2025 B2B Market Monitor was published by ECC KÖLN in collaboration with FIS and Shopware. In 2024, German manufacturers and wholesalers generated EUR 509 billion through online shops and marketplaces, 7 percent above 2023 and 12.1 percent of revenue.
| The retail shopper asks | The wholesale buyer asks |
|---|---|
| What does the product look like, and how will it help me? | Does the product meet our technical and regulatory requirements? |
| Which variant suits my application? | Which pack size, spare parts, and compatible components apply? |
| Can I trust the image, description, and reviews? | Are the datasheet, certificate, classification, and approval current? |
Retail information often needs to build understanding and confidence quickly. Wholesale information must help a professional team reduce procurement risk. The comparison helps you identify the requirements of your own business model. It keeps your focus on the functions and priorities that fit your model.
Every organisation needs a controlled product core whose fields remain stable within its own business model. A retailer defines that core around ranges, variants, brands, and rapid channel output. A wholesaler gives more weight to technical attributes, supplier sources, classifications, and customer views. The architecture principle is comparable, while its implementation depends on the model.
Brand, market, channel, and buyer contexts sit around that core. The contexts you need depend on your case. A retailer may manage image selection, tone of voice, and campaign fields. A wholesaler is more likely to prioritise languages, technical documents, customer catalogues, and access rules.
| Layer | Retail | Wholesale | Question for your case |
|---|---|---|---|
| Product core | ID, variant, base specification, brand | ID, technical features, family, classification | Which fields stay stable across all outputs in your model? |
| Market | Language, range, legal notices, campaigns | Language, units, standards, available catalogues | Which values may vary locally, and who approves them? |
| Channel | Shop, app, marketplace, store | Portal, e-procurement, EDI, print | Which required fields and formats do your target channels expect? |
| Buyer | Benefits, images, reviews, variant guidance | Specifications, documents, compatibility, pack size | Which information reduces risk for your actual buyer? |
Every variation within your model needs a rule. A value may override the core only in an approved field, with a named business owner responsible for the change. Provenance matters just as much. Teams need to see whether a value came from ERP, a supplier, a translation, or manual enrichment. A robust record stores value, context, status, source, and source priority.
Product Information Management provides that controlled framework. Our PIM fundamentals explain it in detail. Here, the focus is how to define the core and contexts for your own business model.
Central management can produce different outputs for each context. For every dimension, the comparison shows its priority in your business model. It also identifies which fields remain contextual and which system owns them. This helps you test which configuration fits your model.
Akeneo commissioned 3Gem to survey 650 B2B professionals across Germany, France, the UK, and the US in 2024. The survey shows which information B2B buyers value most highly. Clear descriptions were critical or very important to 79 percent, technical datasheets to 78 percent, and warranty and compatibility information to 75 percent each.
| Dimension | Retail | Wholesale | Decision | Leading system | Error prevented |
|---|---|---|---|---|---|
| Buyer | Orientation and confidence | Suitability and evidence | Separate view | PIM | Wrong information |
| Content | Benefits, images, reviews | Specifications and instructions | Context-specific | PIM and DAM | Wrong content |
| Product model | Variants, bundles, accessories | Families, spares, successors | Central to each model, relationships by use case | PIM | Duplicate products |
| Documents | Selected when useful | Datasheet, certificate, CAD | Required by profile | DAM or MAM | Expired evidence |
| Channels | Shop, app, marketplace, store | Portal, e-procurement, EDI, print | Separate profiles | PIM and channel | Rejected output |
| Updates | Launches and campaigns | Feeds, standards, customer catalogues | Priority by model | PIM | Overwritten enrichment |
| Commercial logic | List price and promotion | Contract price, tiers, MOQ | Outside PIM | ERP or commerce | Wrong commercial terms |
| Data intake | Brand and merchandising | Suppliers, ERP, classifications | Mapping by source | Source system and PIM | Wrong unit |
| Governance | Marketing and buying | Master data, engineering, compliance | Owner by field group | Defined by the business | Unreviewed approval |
Use the matrix beyond the workshop. Turn every relevant row into a technical rule for your own case. Define what inherits, what may vary, and what happens after an update. Retail may use a shorter title while wholesale preserves the formal technical name. A certificate may be mandatory in a B2B catalogue and optional in a shop. Record these decisions in the data model so they remain consistent and clear to every team.
Technical trade adds standards to the picture. BMEcat transports catalogue data. ETIM and ECLASS structure classes, features, and values. You still need to account for versions, exceptions, and internal taxonomies when mapping the data.
Retail and wholesale each need an end-to-end workflow, but their sources, enrichment, reviewers, and outputs differ. The 2025 SME Procurement Barometer from BME, Onventis, and ESB Business School examines supplier management. Only one in four surveyed companies, or 25 percent, used a digital platform to collaborate with suppliers. Of the more than 350 European procurement professionals surveyed, 80 percent saw supplier management as their greatest digitalisation need. This gives your PIM evaluation a practical question: How will supplier data enter your own workflow?
A retail flow often starts with brand, supplier, and merchandising data. Variants, content, media, campaign approval, and rapid channel output follow. Wholesale sources are more likely to enter staging from feeds and ERP. Mapping, units, classifications, and technical documents are checked before portal, customer catalogue, or EDI output is approved.
Both workflows need useful error feedback. A generic rejected-record message is too vague for a retail team or supplier. Name the product ID, source field, failed rule, expected format, and responsible recipient. Configure stored mappings and failure paths for the process your organisation actually repeats.
Our guide to building an omnichannel experience with PIM provides the broader channel context.
A PIM fits your business model when every field has one leading system and one permitted write direction. Unclear field ownership can let a shop overwrite a PIM description. ERP may return an older value while a media change reaches just one channel. Synchronisation then becomes permanent conflict management.
| Field group | Leading system | Write permission | Frequency | Destination | Failure path |
|---|---|---|---|---|---|
| Price, stock, order | ERP, inventory system, or commerce | Operations team | Event or short interval | Shop, portal, OMS | Stop transaction and alert owner |
| Descriptions, attributes, classification | PIM | Business owner | Change or batch | Profiles in your model | Block approval and show source |
| Media, documents, rights | DAM or MAM | Asset owner | Change or expiry date | PIM and channels | Block asset and request replacement |
| Channel rules and presentation | PIM and commerce | Clearly separated by rule | Release | Channel profile | Stop output and return error |
| Configuration and quote | CPQ, commerce, or ERP | Sales or product team | Request or release | Quote and order | Reject invalid combination |
Write direction and frequency belong together. A product description changes on a different cycle from stock data, and media rights need different controls from tiered pricing rules. For every field group, document the leading system, trigger, transfer frequency, and failure path. These details keep incidents traceable and make a technically working interface safe for business use.
Choose the transfer pattern by frequency and cost of failure. Daily catalogue updates may suit a batch; frequent individual changes may need APIs or events; EDI suits standardised business messages. Monitoring, retry, and rollback keep this data orchestration manageable.
Picture two companies that sell work lights. A retailer in Zurich sells to consumers. A technical wholesaler in Basel supplies trade customers. Both need reliable product data, but each builds an independent model for its buyers, sources, and channels.
The German Economic Institute examined how companies provide digital product data. A 2024 survey of 1,078 companies, published in 2025, found around half provided digital product data; 18 percent of providers used a standardised format. In practice, you should therefore expect source data to arrive in different structures.
The retailer lists a light as RL-480. The wholesaler lists a comparable model as WL-480. Values such as 48 W, 6,200 lm, and IP65 are fictional assumptions. They illustrate the model differences and were not taken from a real catalogue.
| Aspect | Retailer RL-480 | Wholesaler WL-480 |
|---|---|---|
| Content | Images, benefit points, variant guidance | Technical features, datasheets, pack size |
| Relationships | Accessory recommendation | Compatibility, spare, successor |
| Output | Shop and marketplace | B2B portal, customer catalogue, BMEcat |
Product relationships also follow the use case. The retailer models accessories for guidance and cross-selling. The wholesaler needs directed compatibility, replacement, and succession with technical approval. Type, direction, condition, source, owner, and review date are useful in both cases, but they support different decisions.
A hybrid company may later derive several contexts from one core. Most readers can start by configuring PIM cleanly for their own retail or wholesale process.
Retail and wholesale need similar capability categories, but their priorities differ. Akeneo commissioned 3Gem to survey 650 B2B professionals in Germany, France, the UK, and the US in 2024. Forty percent said their organisation still managed product information manually. For selection, identify the manual steps a system would actually handle in your process.
Derive selection criteria from your sources, data objects, roles, and target channels. This connection shows whether a capability is useful in the actual process.
| Capability | Retail priority | Wholesale priority | Question for your model |
|---|---|---|---|
| Data model and inheritance | Variants, bundles, brand, campaigns | Families, technical attributes, classifications | Can the system model your key product objects without copies? |
| Roles and approvals | Merchandising, marketing, buying | Master data, engineering, compliance, sales | Can you model your actual reviewers and risks? |
| Interfaces and transformation | Shop, app, marketplace, DAM | ERP, portal, EDI, BMEcat, partner formats | Are mapping, errors, retries, and monitoring documented? |
| Validation and controlled AI | Content, language, media, channel rules | Units, technical facts, documents, standards | Does the system show source, uncertainty, rule violation, and review? |
| Deployment and data location | Operations, updates, rights, market data | Supplier knowledge, technical data, customer views | Where do data and models run, who operates them, and what changes? |
| Channel-level readiness | Required content by shop, market, or campaign | Required attributes, documents, customer catalogues | Does the system measure readiness for your relevant outputs separately? |
Review AI support by its possible error. A marketing description carries a different risk from an added voltage or selected certificate. The system should show sources, flag uncertainty, and hold critical changes for approval.
A general PIM evaluation checklist can help build the longlist. Next, test one real flow that matches your business model.
A feature list becomes meaningful when you test it against a real data flow. A retailer takes that flow through to shop or marketplace output. A wholesaler tests a supplier feed, technical review, and B2B output. Only a hybrid business needs both workflows in the same pilot.
Apply one change and one failure within your chosen flow. This tests the data model, roles, approvals, rollback, and monitoring under realistic conditions.
| Test point | Yes | No | Unclear | Evidence | Next action |
|---|---|---|---|---|---|
| Import preserves ID, unit, version, and provenance | □ | □ | □ | Preview, log, record | Owner and due date |
| Rules and required fields match your business model | □ | □ | □ | Real target output | Correct the rule |
| Failure is blocked and returned clearly | □ | □ | □ | Error message and recipient | Close failure path |
| Repeat update protects reviewed enrichment | □ | □ | □ | Before-and-after comparison | Adjust mapping |
| Rollback and repeatable runs are evidenced | □ | □ | □ | History and monitoring | Test operating path |
Measure channel-level completeness, error rate, lead time, and automatically published records before the pilot. Your own baseline shows whether the pilot improves the process.
The demo should include your most difficult sources and target channels. Pristine sample data alone cannot provide a robust test. For vendors, costs, and general criteria, use our PIM systems comparison.
elytra can be configured around the process of your trading model. The retail path focuses on product content, translation, validation, and consumer-channel output. The wholesale path focuses on technical product data, ETIM and ECLASS, and connections to SAP, shops, and industry networks.
The requirements from the first eight sections provide the selection framework. No-code configuration, role-based workflows, and several deployment models are available to adapt the data model and approvals. OpenAPI 3.1.1, the Elytra Query Language, and the SAP ERP Connector link source and target systems. Rule-based validation and local containerised models provide additional controls for sensitive product data.
In a joint pilot, we select one real data flow from your operation. For retail, we may test variants, content, media, and shop output. For wholesale, we test feed, mapping, technical rules, and a B2B channel. If your company genuinely operates both models, the second path becomes a separate test.
The Pistor reference covers a clearly defined use case. It describes around 21,000 products, with elytra as the primary PIM, InDesign IDML templates, and Docscape output based on XML rules. Semi-automated publishing has been live since mid-2022. The reference documents publishing and system roles. Supplier onboarding, ETIM/ECLASS, and combined retail and wholesale operation fall outside its scope.
| Your case | What elytra provides | What we align in the pilot |
|---|---|---|
| Retail | Product enrichment, translation, validation, and role-based workflows | Variants, content rules, languages, approvals, target channels |
| Wholesale | ETIM/ECLASS mapping, OpenAPI, EQL, and named SAP ERP Connector | Feed formats, versions, mappings, BMEcat profile, customer catalogue |
| Sensitive product data | Rule-based validation, local containerised models, deployment choice | Sources, review, history, data location, responsibilities |
| Hybrid as an exception | Configurable data models, roles, and workflows | Test the second process only after the first path is robust |
Retail and wholesale set different priorities for PIM. Both need reliable product data, but buyer questions, sources, required fields, approvals, and target channels determine the configuration. The comparison helps you define the requirements of your own model.
elytra supports either case through separate paths. Choose the path that matches your business and test it against real data. A hybrid model remains possible, but it is the exception and should be added as a deliberate second process.
Do you operate in retail or wholesale? Bring one real feed and the relevant target channels to a working session. We will align the data model, rules, interfaces, and approvals with your actual process.
Retail places more weight on content, variants, media, and rapid channel output. Wholesale often needs deeper technical attributes, documents, supplier processes, classifications, and customer-specific catalogue views.
Retail needs a controlled foundation for identity, variants, base specifications, and approved media. Brand, market, and channel contexts then add campaign content, local rules, and approvals.
Wholesale often needs pack sizes, minimum quantities, certificates, datasheets, technical features, compatibility, and customer catalogues. The exact scope follows product group, customer, standard, and destination.
BMEcat is an XML-based format for exchanging electronic product catalogues. ETIM and ECLASS structure classes, features, and values and can be used within that exchange process.
That depends on the architecture, but price, stock, and orders often remain governed by ERP, inventory, or commerce systems. PIM holds the product context and identifiers needed for consistent output.
No. Most organisations configure PIM for one of the two models. Only a genuine hybrid business needs several contexts and separate workflows on one platform.
AI can extract data, propose mappings, enrich content, and flag anomalies. Sources, rules, confidence, human review, history, and approval should remain traceable.
Take one real feed through your most important process to the relevant target channels. Test import, rules, approval, integration, output, failure handling, rollback, and recurring updates.