Demo Theatre: the PIM demo scenario library
Demo Theatre is a library of 30 PIM demo scenarios across ten areas: onboarding, modelling, enrichment, workflow, syndication, compliance, AI, governance, integration and performance. Each scenario states the sample data the buyer provides, three to five actions the vendor must show live, measurable pass criteria, the typical failure pattern and the requirement IDs it tests.
What is Demo Theatre?
Demo Theatre is a library of 30 ready-to-adapt PIM demo scenarios, each built to run live on the buyer's own sample data. It supports the scenario-based demo step in weeks 8 and 9 of the 12-week PIM selection process.
Scenario demos built on the buyer's own data are the only demos that predict fit. A generic demo runs on showcase products that hide the modelling compromises your catalogue would force. Demo Theatre gives you the scripts to avoid that trap.
Each scenario has the same structure:
- Setup: the sample data and context the buyer provides.
- Show: three to five actions the vendor must perform live.
- Pass: two to four pass criteria, measurable where possible.
- Failure: the typical way vendors fall short.
- Industries: where the scenario matters most.
- Requirements: the requirement IDs it tests, from the four PIMvendors domains.
The requirement links connect each scenario to the RFP. You can compare a vendor's written answer with its live result, and weight the scenario by the priority of its requirements.
How do you choose scenarios for your demo script?
Choose 8 to 12 scenarios that cover your must-have requirements, with at least one from each area that matters to you. Half a day per vendor fits that number (PIMvendors estimate).
| Area | Scenarios | Include when |
|---|---|---|
| Onboarding | SC-01 to SC-03 | You receive supplier data or run frequent imports |
| Modelling | SC-04 to SC-06 | You have variants, kits, spare parts or new families |
| Enrichment | SC-07 to SC-09 | You publish to several channels, markets or languages |
| Workflow | SC-10 to SC-12 | Several teams approve product data before publication |
| Syndication | SC-13 to SC-15 | You feed GS1 GDSN, retailers, marketplaces or wholesalers |
| Compliance | SC-16 to SC-18 | Packaging, safety or passport rules apply to your products |
| AI | SC-19 to SC-21 | You plan AI descriptions, extraction or classification |
| Governance | SC-22 to SC-24 | Many roles, markets or auditors touch the data |
| Integration | SC-25 to SC-27 | ERP, webshop or other systems exchange data with the PIM |
| Performance | SC-28 to SC-30 | Your catalogue or user base is large |
Compliance scenarios deserve attention in 2026. The Packaging and Packaging Waste Regulation applies generally from 12 August 2026, and the General Product Safety Regulation has applied since 13 December 2024. Scenarios SC-16 and SC-17 test whether a platform holds that data as structured, governed attributes.
How do you adapt a scenario to your catalogue?
Adapt each scenario by replacing the sample volumes, standards and pass thresholds with your own, and keep the structure intact. The structure makes scenarios comparable across vendors. The content makes them predictive for you.
Work through four changes:
- Volumes: use your real catalogue size, supplier file size and variant counts.
- Standards: swap ETIM for ECLASS, BMEcat for another catalogue format, or GDSN for a retailer portal, as your customers require.
- Pass criteria: set thresholds from your requirements. A distributor publishing daily needs faster bulk edits than a brand with two seasons a year.
- Requirement IDs: link each scenario to the IDs in your own validated list, built from the requirements checklist.
Prepare the answer keys before the demo. For AI and classification scenarios, the buyer's experts classify or describe the sample products first, so the team can measure vendor accuracy against them.
Which scenarios separate PIM vendors most clearly?
The scenarios that separate vendors most clearly test structural capabilities that cannot be prepared on slides. In PIMvendors selections, four stand out.
Rolling back a bad import (SC-03) shows whether the platform tracks changes per job or only per product. Bulk-editing 10,000 products in under 60 seconds (SC-28) shows performance at volume. A variant family with 40 sizes and colours (SC-04) shows the depth of the variant model. A live change to configuration shows how much effort the platform needs after go-live.
AI scenarios separate vendors too, under one condition: they must run on your own products, with measured accuracy and a human approval step. AI in PIM is useful for attribute extraction, translation, auto-classification and channel-specific descriptions. It remains unreliable for regulatory data, so keep a human approval in every compliance scenario.
Score each scenario with the PIM demo scorecard, and watch for the demo warning signs in the Red Flag Register.
The scenario library
The 30 scenarios follow, grouped by area. Filter by area or industry to build your script.
Onboarding
Onboard 5,000 supplier SKUs and classify to ETIM
The buyer provides one supplier spreadsheet with 5,000 SKUs, mixed units and free-text descriptions, plus the target ETIM classes and feature list.
The vendor shows
- Map the spreadsheet columns to PIM attributes and save the mapping for the next delivery
- Normalise units and value lists during import, with rejected values listed per row
- Assign ETIM classes in bulk and show the features each class adds
- Report completeness against the ETIM features after import
Pass criteria
- Import of 5,000 rows completes in under 10 minutes with a downloadable error report
- At least 90 percent of rows land in the correct ETIM class after review by the buyer
- The saved mapping reruns on a second file without rework
Typical failure. The vendor imports a cleaned file prepared in advance and skips the unit and value-list normalisation.
Supplier self-service onboarding with validation
The buyer names two supplier personas and provides the mandatory attribute set for one product family, with 20 sample products per supplier.
The vendor shows
- Invite a supplier to a portal or template with only the attributes it must supply
- Validate supplier input against mandatory fields, formats and value lists before submission
- Route the submission to a category manager for approval or rejection with comments
- Show the supplier the status and the rejection reason
Pass criteria
- Supplier cannot submit records that fail mandatory validation
- Approved records appear in the main catalogue without manual re-keying
- Rejection comments reach the supplier within the same workflow
Typical failure. The supplier portal is a separate product with its own data model, and approved data needs a manual transfer.
Roll back a bad import
The buyer provides a supplier file with a deliberate error: 800 products receive wrong net weights and an overwritten marketing description.
The vendor shows
- Run the faulty import and show which products and attributes changed
- Identify the import job as the source of the changes in the history
- Restore the previous values for all affected products in one action
- Confirm that downstream channels receive the corrected values
Pass criteria
- All 800 products return to their previous values in a single job, without manual edits
- Rollback completes in under 5 minutes
- The audit trail records both the import and the rollback with user and timestamp
Typical failure. The vendor can show version history per product, but cannot reverse a whole import job in one action.
Modelling
Create a variant family with 40 sizes and colours
The buyer provides one product with 8 sizes and 5 colours, shared and variant-specific attributes, and images per colour.
The vendor shows
- Create the parent product and generate all 40 variants from size and colour axes
- Set attributes at parent, colour and size level with inheritance
- Override one inherited value on a single variant and show the override clearly
- Link colour images once and inherit them to all sizes of that colour
Pass criteria
- All 40 variants created in under 5 minutes by a business user
- A parent-level change propagates to all 40 variants in one save
- Overrides remain visible and reversible per variant
Typical failure. The platform supports only one variant axis, so the second axis is faked with separate products or text attributes.
Add a new product family with inherited attributes
The buyer describes a new product family that shares 60 percent of its attributes with an existing family and needs 15 new attributes.
The vendor shows
- Create the new family by reusing existing attribute groups
- Add the 15 new attributes with types, units and validation rules
- Define which attributes are mandatory for each channel
- Move three existing products into the new family without losing data
Pass criteria
- A business administrator completes the change without vendor or partner support
- Existing values survive the family change for all three products
- Channel completeness updates immediately after the change
Typical failure. Changing the family of existing products requires a data migration script from the vendor or partner.
Model kits, spare parts and accessories
The buyer provides one kit of 6 components, 10 spare parts for one machine and 5 accessories that fit several products.
The vendor shows
- Create a kit with quantities per component
- Relate spare parts to the machine and to each other with typed relationships
- Link one accessory to several products in one bulk action
- Export the relationships to a channel in a structured format
Pass criteria
- Relationship types are configurable by the buyer, without code
- Kit quantities and component references export correctly
- Relationships remain valid when a component is renamed or replaced
Typical failure. Relationships are stored as free text or comma-separated codes that break when a part number changes.
Enrichment
Measure completeness per channel and market
The buyer provides 200 products and the mandatory attribute lists for a webshop, a marketplace and a printed catalogue in two markets.
The vendor shows
- Configure completeness rules per channel and per market language
- Show which products are not ready for each channel and why
- Filter work queues to the missing attributes for one team
- Update the completeness score live after an edit
Pass criteria
- Completeness rules differ per channel without duplicate attribute definitions
- Missing attributes are listed per product, per channel
- Scores update within seconds of an edit
Typical failure. Completeness is one global score that ignores channel and market differences.
Translate and localise into four languages
The buyer provides 100 products in English with marketing and technical attributes, and target markets in Dutch, German, French and Polish.
The vendor shows
- Mark which attributes are translatable and which are shared across languages
- Send products for translation, by machine or agency, and track status per language
- Convert units and formats per market where required
- Flag translations as outdated when the source text changes
Pass criteria
- Technical values stay shared and are never translated by mistake
- Outdated translations are flagged automatically after a source edit
- Translation status is visible per product and per language
Typical failure. Every attribute is duplicated per language, so technical values drift apart across markets.
Link assets to products and variants
The buyer provides 300 images, 40 PDF datasheets and 10 videos with file names that encode SKU and view type.
The vendor shows
- Link assets to products automatically from file names or metadata
- Assign asset roles such as main image, gallery, datasheet and safety sheet
- Inherit colour images to all variants of that colour
- Show which channels require which asset roles
Pass criteria
- At least 95 percent of assets link automatically by naming rule
- Asset roles drive channel completeness
- Replacing an asset updates every linked product
Typical failure. Assets are uploaded one by one per product, and file-name rules need custom scripts.
Workflow
Run a new product introduction workflow
The buyer provides its current new product process with five roles: product manager, data steward, marketing, compliance and channel manager.
The vendor shows
- Configure the five-step workflow with tasks per role
- Show each role only its tasks and attribute groups
- Block publication until compliance approves
- Report lead time per step for products in progress
Pass criteria
- The workflow is configured by a business administrator in the session
- Publication is blocked until every mandatory approval is recorded
- Lead time per step is reportable without exporting data
Typical failure. Workflow exists only as status fields, with no tasks, notifications or blocking rules.
Change a live product with approval and audit
The buyer selects a published product whose material composition must change, which affects compliance and marketing attributes.
The vendor shows
- Edit the material attribute on a live product
- Trigger re-approval by compliance for the dependent attributes
- Keep the published version live until the change is approved
- Show the full audit trail of the change
Pass criteria
- Channels keep receiving the approved version until the new version is approved
- The audit trail shows old value, new value, user and timestamp
- Dependent attributes are flagged for review automatically
Typical failure. Edits go live immediately, and approval only exists for new products.
Schedule a seasonal launch across channels
The buyer provides 150 products for a seasonal collection with a fixed launch date and different go-live dates per market.
The vendor shows
- Group the products into a launch collection
- Set publication dates per channel and market
- Track readiness of the collection against the launch date
- Publish automatically on the scheduled date
Pass criteria
- Publication happens on the scheduled date without manual action
- Readiness is visible per channel and market for the whole collection
- Products that miss the deadline are listed with their missing attributes
Typical failure. Scheduling is handled in the webshop, so the PIM cannot show launch readiness.
Syndication
Publish products to GS1 GDSN
The buyer provides 50 products with GTINs, packaging hierarchy, net content and nutritional or technical data, plus a target data pool and recipient.
The vendor shows
- Map PIM attributes to the GDSN attributes for the target market
- Build the packaging hierarchy from consumer unit to pallet
- Validate against GDSN rules before publication
- Publish to the data pool and show the recipient response
Pass criteria
- Validation errors appear in the PIM before sending, with the attribute named
- Hierarchy and GTINs are published without manual data pool entry
- Recipient responses return to the PIM and are visible per product
Typical failure. The vendor exports a file for manual upload to the data pool and cannot show the response loop.
Export a retailer or marketplace feed
The buyer provides 500 products and the current attribute template of one retailer or marketplace it sells through.
The vendor shows
- Map PIM attributes and value lists to the channel template
- Generate channel-specific values such as titles and category codes
- Validate the feed against channel rules before export
- Export only changed products on the next run
Pass criteria
- Mapping is maintained by a business user, without code
- Feed passes the channel validation with fewer than 2 percent rejected products
- Delta export contains only changed products
Typical failure. Each channel needs a custom export built by the partner, and template changes need a new release.
Deliver a BMEcat catalogue with ETIM features
The buyer provides 1,000 products classified to ETIM and a wholesaler's BMEcat requirements, including version and mandatory fields.
The vendor shows
- Generate a BMEcat file in the requested version with ETIM features
- Include prices, packaging units and asset references as required
- Validate the file against the wholesaler's mandatory fields
- Schedule the export for regular delivery
Pass criteria
- The BMEcat file validates against the schema without manual edits
- ETIM features export with the correct units and value codes
- A scheduled export runs without user action
Typical failure. BMEcat export relies on a third-party converter outside the PIM that nobody on the buyer's team maintains.
Compliance
Capture PPWR packaging attributes
The buyer provides 30 products with packaging made of several components, with materials, weights and recycled content from suppliers.
The vendor shows
- Model packaging as components with material, weight and recycled content per component
- Mark reusable packaging and record the related information
- Validate that every component has a material and weight before approval
- Report packaging data by material across the catalogue
Pass criteria
- Packaging components are stored as a structured, repeatable model per product
- Products without complete packaging data cannot pass the compliance approval
- Catalogue-wide packaging reports run without export to spreadsheets
Typical failure. Packaging data is stored as one text field per product, so per-component reporting is impossible.
Publish GPSR safety information per market
The buyer provides 40 products sold online in three EU markets, with manufacturer details, the EU responsible person and safety warnings.
The vendor shows
- Store manufacturer and EU responsible person contact details once and link them to products
- Manage warnings and safety information per market language
- Block publication to online channels when mandatory safety data is missing
- Export the safety data to a webshop and a marketplace
Pass criteria
- Contact details change in one place and update every linked product
- Missing safety information blocks channel publication
- Warnings export in the correct language per market
Typical failure. Safety data sits in PDF attachments only, so online channels cannot display it as structured text.
Assemble a Digital Product Passport data set
The buyer provides one product and a list of passport data points it expects for its product group, including supplier-provided values.
The vendor shows
- Model the passport data points with source and owner per attribute
- Collect supplier values with evidence documents
- Generate a product identifier link to the passport data
- Export the data set in a structured, machine-readable format
Pass criteria
- Each data point shows its source, owner and last update
- Evidence documents link to the attribute they support
- Export is machine-readable and repeatable for the full product group
Typical failure. The vendor shows a static web page mock-up with no link to governed attribute data.
AI
Generate channel descriptions with AI and approval
The buyer provides 50 products with structured attributes, a brand tone guide and the length rules of two channels.
The vendor shows
- Generate a description per channel from attributes and the tone guide
- Show which attributes the AI used for each description
- Route generated text to an editor for approval before publication
- Regenerate one description after an attribute change
Pass criteria
- No AI text publishes without a recorded human approval
- At least 80 percent of descriptions are accepted with minor edits, judged by the buyer's editor
- Generated text stays within channel length limits
Typical failure. Generated text writes straight into the live description field with no approval step.
Extract attributes from supplier datasheets
The buyer provides 30 supplier PDF datasheets in two languages and the target attribute set for that product family.
The vendor shows
- Extract attribute values from the PDFs into the target attributes
- Show a confidence level and the source location for each value
- Send low-confidence values to a review queue
- Convert units to the buyer's standard during extraction
Pass criteria
- Each extracted value links to its source page or passage
- Values below the confidence threshold never write to approved attributes
- Extraction accuracy is measured against a buyer answer key for the 30 datasheets
Typical failure. Extraction runs on the vendor's prepared datasheets, and accuracy on the buyer's files is never measured.
Auto-classify products with confidence thresholds
The buyer provides 500 unclassified products with names and short descriptions, and its target classification tree.
The vendor shows
- Suggest a class per product with a confidence score
- Apply suggestions above a threshold the buyer sets, and queue the rest
- Let a data steward accept or correct suggestions in bulk
- Report classification accuracy after review
Pass criteria
- The buyer can set and change the confidence threshold
- Corrections are logged and visible in the audit trail
- Accuracy after review is reported per class
Typical failure. Classification applies directly to all products with no threshold, queue or accuracy report.
Governance
Restrict permissions by attribute group and market
The buyer defines four roles: a supplier, a German market editor, a compliance officer and an external agency.
The vendor shows
- Configure edit and view rights per attribute group and per locale
- Log in as each role and show what it can and cannot change
- Restrict the agency to marketing attributes for one brand
- Show the permission change in the audit trail
Pass criteria
- Each role sees and edits only its attribute groups and locales
- Permission changes require no code and take effect immediately
- Permission changes are logged with user and timestamp
Typical failure. Permissions apply only per product category, so locale and attribute-group rights need custom development.
Enforce data quality rules with a dashboard
The buyer provides 1,000 products with known quality issues: duplicates, invalid GTINs, missing units and inconsistent value lists.
The vendor shows
- Configure validation rules for GTIN check digits, units and value lists
- Detect likely duplicate products
- Show a quality dashboard by product family and owner
- Assign quality issues to owners as tasks
Pass criteria
- Invalid GTINs are detected on import and on edit
- The dashboard reports issues per owner and family without export
- Rules are configured by a data steward, without code
Typical failure. Quality checks run as one-off reports, with no rules that block bad data on entry.
Prove the history of a compliance attribute
The buyer selects a product whose hazard or allergen information changed twice in the past year, based on supplier updates.
The vendor shows
- Show every value the attribute held, with dates and users
- Show which value each channel received and when
- Link supplier evidence to each change
- Export the history for an auditor
Pass criteria
- Full attribute history is available for at least the contract retention period
- Channel publication history shows the value sent per date
- Export is readable by a non-technical auditor
Typical failure. History covers product saves only, so the value a channel received on a given date cannot be proven.
Integration
Create items from ERP with a clear master split
The buyer provides an ERP item extract of 2,000 items and states which attributes the ERP masters and which the PIM masters.
The vendor shows
- Import ERP items through a standard connector or API on a schedule
- Lock ERP-mastered attributes against editing in the PIM
- Show what happens when the ERP changes an item already enriched in the PIM
- Report and resolve a failed item in the integration log
Pass criteria
- ERP-mastered attributes are read-only in the PIM
- ERP updates never overwrite PIM-mastered enrichment
- Failed records appear in a log with the reason and can be retried
Typical failure. The connector overwrites enriched attributes on every sync because ownership is not configurable per attribute.
Publish deltas to a webshop through the API
The buyer names its commerce platform and provides 100 products that change in non-price attributes during the session.
The vendor shows
- Publish only changed products and attributes to the webshop
- Use events or webhooks to notify the webshop of approved changes
- Show API documentation and rate limits for the endpoints used
- Show the publication status per product
Pass criteria
- Changed products reach the webshop within the agreed time, measured in the session
- Only approved versions are published
- API limits are documented and fit the buyer's change volume
Typical failure. The integration exports the full catalogue every night, so urgent fixes wait a day.
Recover from a failed outbound sync
The buyer asks the vendor to simulate an outage of the target channel during a publication of 1,000 products.
The vendor shows
- Show how the failure is detected and reported
- Alert a named role by email or messaging
- Retry only the failed products once the channel is back
- Show the final publication status per product
Pass criteria
- Failure triggers an alert within minutes, without someone watching a screen
- Retry sends only failed products, without duplicates
- Final status is correct for all 1,000 products
Typical failure. The failed job must be restarted in full, and nobody is alerted until a channel manager notices missing products.
Performance
Bulk-edit 10,000 products in under 60 seconds
The buyer's sample environment holds at least 50,000 products, and the buyer selects 10,000 of them by a filter on family and brand.
The vendor shows
- Select 10,000 products by filter
- Change three attributes in one bulk action
- Show progress and the job log
- Show the changed values in the product view and the completeness score
Pass criteria
- Bulk edit of 10,000 products completes in under 60 seconds
- Users can keep working while the job runs
- The job log lists every changed product
Typical failure. The vendor runs the bulk edit on a few hundred products and extrapolates the timing.
Search and filter a catalogue of 500,000 SKUs
The buyer provides or the vendor generates a catalogue of 500,000 SKUs with the buyer's attribute structure and 4 locales.
The vendor shows
- Run a full-text search across names and descriptions in two locales
- Combine filters on family, attribute values and completeness
- Save the filter as a shared view for a team
- Export the filtered result
Pass criteria
- Search results return in under 2 seconds
- Combined filters return in under 5 seconds
- Saved views are shareable without admin rights
Typical failure. Performance is shown on a small demo catalogue, and large-catalogue figures come only from slides.
Keep editing fast during a large import
The buyer asks the vendor to run an import of 20,000 products while five users edit products in the same environment.
The vendor shows
- Start the import and show its progress
- Let five users edit and save products during the import
- Show save times for users during the import
- Resolve an edit conflict between the import and a user
Pass criteria
- User save times stay under 3 seconds during the import
- Edit conflicts are resolved by a documented rule, without lost changes
- The import completes without errors caused by concurrent edits
Typical failure. Imports lock the products or the environment, so editors wait until the job finishes.
Give every vendor the same script and the same data, and the differences show up on their own. Scenarios built on your products leave no room for showcase answers.
Want a second pair of eyes on your selection?
PIM Selection is published by PIMvendors.com, which runs scenario-based PIM selections for European manufacturers, distributors and brands. A 30-minute call is enough to see whether that helps.
See PIMvendors selection servicesBook a call with PIMvendors
Frequently asked questions
How many scenarios should a PIM demo include?
A PIM demo should include 8 to 12 scenarios, enough to cover the must-have requirements in half a day per vendor (PIMvendors estimate). Pick at least one scenario from each area that carries must-have requirements for you. Include at least one scenario that tests change after go-live, such as adding a product family, and one that tests performance at your real volume.
Can we use these demo scenarios without changes?
The Demo Theatre scenarios work as templates and need adapting to your catalogue. Replace the sample volumes with your real ones, swap standards such as ETIM for the ones your customers require and adjust pass criteria to your requirement priorities. Keep the structure: setup, live actions, pass criteria and requirement IDs. That structure makes the scenarios comparable across vendors.
How do we test AI features in a PIM demo?
Test AI features in a PIM demo by running them live on products from your own sample pack, never on curated examples. Ask for confidence scores, the source of each extracted value and a human approval step before publication. Measure results against an answer key you prepare in advance. Scenarios SC-19 to SC-21 in Demo Theatre cover descriptions, extraction and classification.
Which demo scenarios separate PIM vendors most clearly?
In PIMvendors selections, the scenarios that separate vendors most clearly are rolling back a bad import, bulk-editing 10,000 products, creating a variant family on two axes, and changing configuration live. They test structural capabilities that cannot be prepared on slides. AI scenarios also separate vendors, but only when they run on the buyer's own data with measured accuracy.
Sources and method
- EU packaging regulation starts applying from August 12
- General Product Safety Regulation (Wikipedia)
- Ecodesign for Sustainable Products (Department of Enterprise, Tourism and Employment, Ireland)
How this page was made: drafted by the PIM Selection editors from the sources listed, checked against the PIMvendors selection method, and signed off by a named expert. Editorial policy.
Cite this page
PIM Selection (2026). Demo Theatre: the PIM demo scenario library. PIMvendors.com. https://pimselection.com/process/demo-scenarios/ (version 2026-10-14).