PIM KNOWLEDGE
PIM implementation
The sequence is deliberate: content and responsibilities first, technology second. Starting with the software often means simply recreating the problems of legacy systems.
The starting point is understanding how items, products and product groups relate to one another, along with their properties and relationships. This forms the basis of the data model with its categories, attributes and relationships. It also includes the processes for creating, translating, enriching and exchanging all content and media.
Who maintains which data, and who approves it? Responsibility should sit where the data delivers the greatest value. Define data owners and creators, roles, organisational structures and an operating model with the necessary resources. PIM projects fail more often because these questions remain unresolved than because of technical issues. See Data Governance.
The rules and responsibilities guide the selection and configuration of the PIM system. The IT landscape is designed around automated data flows: a flowchart showing integrations and formats, the division of responsibilities between PIM and DAM, supplier integration, a transition model, and print and mobile applications. Our PIM selection page explains how we find the right system.
Implementation requires clear objectives and a roadmap: an implementation plan, cost and resource planning, quick wins, change management, project organisation, risks and alternatives. For an overview of expected costs, see PIM implementation & costs.
Content rules: Which product groups do you have, and which attributes does each need? Which classifications do customers and retailers require (ETIM, ECLASS, GS1)? Which relationships are needed, such as accessories, spare parts and variants?
Content ownership: Who creates new items, and who adds technical data, marketing copy, images and translations? Who approves content for each channel? How many hours per week are actually available for this work?
Content technology: Which system is the system of record for which data (ERP, PIM, DAM, PLM)? Which channels and formats need to be supported? Cloud or self-hosted?
Implementation: Which product group and channel should we start with? How will we measure success? Who will lead the project on the business side?
In practice, the four steps span six project phases. The timings are indicative, based on projects in mid-sized businesses. Product range size, data quality and the number of interfaces can change them significantly.
Highlighted in dark: the core implementation up to the first live channel (usually 6–12 months).
A proven approach is to start with a clearly defined scope: one representative product group, one channel (usually your own online shop) and the most important languages. This tests the data model early with real data, gives business teams tangible results within a few months and reveals design flaws while they are still inexpensive to fix.
Further product groups, channels and automation then follow in defined stages. The key is to design the data model for the entire product range from the outset, even if only part of it is populated initially.
Where does your business stand?
Use the PIM checklist to find out in three minutes whether a PIM system is worthwhile for your business and which systems could be a good fit. We will email you an initial shortlist.
The goal of implementation is to flexibly serve every channel and publication with data and media for all products. This requires a consolidated master record, the Golden Record. It contains all information about a product in all its variations, and every channel draws on it.
We plan the Golden Record around the output, starting with the key channels and using representative examples of the most important products. Only once we have established what we want to show to whom, which shop features we need, and how creation, updates and approval will work can we reliably define hierarchies, relationships, master data and transactional data in PIM, ERP and CRM.
One of the most important implementation decisions is to define the system of record for each data type. A typical allocation:
Each data element has exactly one system of record. Maintaining the same data in two places is the most common cause of conflicting data across channels.
Implementation is guided by user stories for the tasks the PIM performs: collecting data (Collect), cleansing it (Consolidate), enriching it (Enrich), validating it (Qualify), distributing it (Distribute) and managing it with version control, permissions and workflows.
The most demanding part. The ERP interface must work reliably, as its master data usually forms the basis for all product data. Supplier data and classifications such as ECLASS or ETIM are also included. Your own data is then classified, mapped and enriched.
Once centralised data management is in place, the channels are connected: online shop, marketplaces, print, data pools and retailers. Each channel has its own formats, fields and approvals.
The project is a good opportunity to put your data in order. Waiting for perfect data, however, only delays implementation.
Product management, marketing, e-commerce, sales, technical documentation, translation and IT all work with the PIM, often alongside suppliers and agencies. The project itself needs clearly assigned roles:
| Role | Responsibilities | Typically drawn from |
|---|---|---|
| Sponsor | Sets objectives and budget, resolves conflicting priorities between departments | Senior management |
| PIM Product Owner | Defines the target vision and priorities, provides business sign-off | Product management or e-commerce |
| Data Owner | Accountable for a data domain, such as technical data or marketing copy | Business departments |
| Data Steward | Rules, quality checks, ongoing data maintenance | Product data management |
| Key User | Tests, trains colleagues and provides feedback from day-to-day use | All participating departments |
| IT / Integration | Interfaces, operations, permissions, security | Internal IT |
| Implementation Partner | Configuration, development, technical migration | Software vendor's service provider |
| PIM Consultant | Vendor-neutral selection, solution design and quality assurance | External consultancy |
The article Service partner or in-house team explores whether to implement with a partner or your own team.
Most PIM projects run into trouble not because of the software, but because of organisational issues, data and the approach taken. We see these patterns time and again in projects:
The software is purchased, but processes and responsibilities stay the same. The same gaps and inconsistencies simply end up in a new system.A better approach: Business teams lead, IT implements. The Product Owner comes from product management or e-commerce.
Legacy data from ERP, Excel and online stores is almost always less complete and less consistent than expected. Mapping, cleansing and enrichment take more time than configuration.A better approach: Assess data quality using representative samples before selection, and plan migration effort separately.
Some companies postpone implementation until all their data has been cleansed. This rarely works without the right tools.A better approach: Cleanse critical attributes upfront, then improve the rest in the PIM system using quality rules and workflows.
Launching all product groups, channels and languages on the same date significantly increases risk and extends timelines.A better approach: Start with one product group and one channel, then expand in clearly defined stages.
Simply copying fields and structures from ERP or Excel misses the opportunity to build a clean data model tailored to your channels.A better approach: Design the data model around output requirements: What does each channel and each customer need?
If the people maintaining the data only see the system at rollout, user buy-in and practical feedback will be missing.A better approach: Involve key users from the design stage, test early with real data, and plan for training.
Without figures from before the project, it is difficult to demonstrate the benefits later or secure funding for subsequent phases.A better approach: Measure completeness, data maintenance effort and time-to-market before the project starts.
Before the project starts, define how success will be measured and establish your baseline. A small set of clearly measurable KPIs has proven effective:
| KPI | What is measured |
|---|---|
| Completeness | Percentage of required attributes populated for each product group and channel |
| Time-to-market | Days from creating an item in the ERP to approval for all channels |
| Data maintenance effort | Minutes per new item or update, across everyone involved |
| Errors and queries | Complaints about product data, queries from retailers, marketplace rejections |
| Channel coverage | Number of items available across channels and languages |
| User adoption | Active users, data maintenance outside the PIM (separate Excel spreadsheets), satisfaction |
New EU requirements and AI applications are increasing the demands on structured product data. If you are planning your data model now, consider the following:
The Ecodesign for Sustainable Products Regulation (ESPR) introduces the Digital Product Passport in stages. The first is the battery passport under the Batteries Regulation, from 18 February 2027; other product groups such as textiles, furniture and tyres will follow through delegated acts. Required information includes materials, origin, repairability and recycling, data that is currently scattered across many businesses. A PIM is a natural source for this data.
The EU Packaging and Packaging Waste Regulation has applied since 12 August 2026. Packaging requires, among other things, a declaration of conformity and technical documentation, with stricter requirements for recyclability and recycled content taking effect from 2030. Packaging data for each item should therefore be included in the data model from the outset.
Since 13 December 2024, online listings have been required to include manufacturer details, product identification, warnings and safety information, among other things. This information belongs in the PIM as dedicated attributes so that it appears automatically in online shops and on marketplaces.
AI can draft and translate content, classify products and identify gaps. But this only works with clearly structured attributes and well-defined rules. When planning the implementation, include workflows for reviewing and approving AI suggestions, and document which content may be generated automatically. AI assistants and search engines are also increasingly accessing structured product data.
For mid-sized businesses, implementation through to the first live channel often takes six to twelve months. This is preceded by selection, which usually takes two to three months.
Almost always data migration: legacy data from ERP, Excel and older systems needs to be cleansed, mapped to the new data model and enriched. The effort involved is routinely underestimated.
The Golden Record is the complete master record for a product in the PIM. It contains all the data, text, media and relationships that every channel can draw on.
Usually yes, at least for configuration and interfaces. However, business tasks such as defining the data model, assigning responsibilities and maintaining data remain in-house.
In phases. Starting with one representative product group and one channel has proven effective. The data model should still be designed for the entire product range from the outset.
Yes. In many projects, Excel spreadsheets are the most important source alongside the ERP. Critical attributes should be cleansed before import; the rest can be improved in the PIM using data quality rules.
In addition to licence or subscription fees, there are costs for implementation, interfaces and data migration, plus internal effort from business departments, which is often underestimated. For an overview, see PIM implementation & costs.
Typical signs include product data spread across numerous Excel spreadsheets, long lead times before new items go online, inconsistent information across channels and growing demands from retailers and marketplaces. The PIM checklist provides an initial assessment.
Is your business ready for PIM implementation?
We provide a vendor-independent review of your data model, processes, organisation and systems, and identify where adjustments are needed before the project starts.
Just getting started? Begin with the free PIM checklist or get in touch directly: Contact.
PIM KNOWLEDGE
Find out in 3 minutes whether a PIM system could benefit your business and which PIM systems could be right for you!
Webdesign: Deutsche Stadtmarketing