Most articles about PIM implementation describe the project stages from the agency's perspective: analysis, configuration, testing, and launch. This is true, but incomplete. A sales director or e-commerce manager is interested in something different: what should I do before engaging an implementation partner? And what are my responsibilities during the project?
This roadmap answers exactly this question – separately for the manufacturer and distributor, because the starting points are different.
Before You Begin: 3 Questions You Need to Ask Yourself
Before you send your first request for proposal (RFP) to a PIM agency, answer three questions. Not just to "pass" it, but to prevent the project from starting with a mid-scope rewrite.
1. Where does your product data live today? Count how many places your company stores product information: ERP, Excel spreadsheets, hard drive folders, supplier emails, and the store's CMS. Each of these places is a potential source of data migration and a separate item on the roadmap.
2. Who in the company makes decisions about data structure? PIM isn't an IT project. It's a project encompassing sales, marketing, logistics, and IT departments simultaneously. You need one person with the mandate to make decisions about product attributes, categories, and distribution channels.
3. Do you want to implement PIM alone or as part of a larger project? Implementing PIM "in a vacuum" makes sense, but we more often see it as part of a broader project: PIM + B2B e-commerce , PIM + ERP integration, PIM + marketplace. The decision on scope will influence the choice of system and partner.
Don't miss out on more valuable content about the digitalization of B2B trade!
Roadmap for the manufacturer
A manufacturer starts with a unique challenge: product data often doesn't exist in any structured system. It's in ERP (incomplete), in Excel (inconsistent), and in the heads of product specialists.
Stage 0 – Internal Data Audit (2–4 weeks, before system selection)
Before choosing a PIM system, do a simple inventory:
- How many SKUs do you have? How many product families?
- What attributes describe the products in each family?
- Who in the company creates and updates product data?
- How many sales channels will you publish data to?
You can complete this step internally or with the help of a consultant. The result is an "implementation brief"—a document that each potential partner will receive as the basis for their valuation.
Stage 1 – Data modeling (as part of pre-implementation analysis)
For a manufacturer, this is a crucial step. A good data model—that is, how you define product families, attributes, and their values—will determine whether employees will want to use the system or avoid it.
The principle we follow: start with an MVP. Establish 10–15 attributes per product family. You're not aiming for perfection on Day 1—you're aiming for a system that works and can be expanded.
Stage 2 – Data Migration and Enrichment
Data from ERP enters PIM as a skeleton—names, codes, and basic parameters. Your team (or automated AI tools available, for example, from Akeneo) then enriches it with marketing descriptions, photos, and technical documentation.
Roadmap for the distributor
The distributor has a different challenge: the data exists, but it comes from dozens or hundreds of suppliers, in different formats and of varying quality.
Stage 0 – Map of suppliers and data formats
Before you start a PIM project, inventory your suppliers for data:
- How many suppliers provide product data?
- In what formats (CSV, XML, API, PDF, email)?
- What is the quality of this data – is it complete and consistent?
This is an introduction to the decision whether to implement Supplier Portal – a module in Akeneo that transfers responsibility for data quality to the suppliers themselves.
Stage 1 – Data normalization and master data management
Data from supplier A names the product "Cable 3×1,5mm²," while supplier B names it "Wire 3×1,5." It's the same product. Normalization is a tedious but necessary step—if well planned in the PIM data model, it later becomes automatic.
Stage 2 – Integration with sales platforms
A distributor typically publishes data to more channels than a manufacturer: their own B2B and B2C store, marketplaces (Allegro, Amazon), PDF catalogs, and integrations with customer systems. PIM becomes a central hub from which data is automatically distributed everywhere.
Common stages: from kick-off to launch
Regardless of whether you are a manufacturer or distributor, the steps involved in working with an implementation partner are similar:
Kick-off – confirmation of scope, schedule, and roles in the project. Each party knows who makes which decisions.
System configuration – PIM installation, data model mapping, channel and permission configuration.
Building integration – connecting with an ERP, e-commerce platform , or marketplace. This is usually the most time-consuming step.
Data migration – transferring data from existing sources to PIM and verifying their completeness.
Testing and training – testing with target users, user training. This stage is often shortened under schedule pressure – a mistake that backfires after launch.
Launch (go-live) – transfer of integration to the production environment, verification of data exchange in real time.
The most common mistakes we see before implementation
- Too small scope of analysis – companies want to sign a contract and move on to implementation as quickly as possible. A shortened analysis provides a sense of progress, but leads to rewriting the scope mid-project.
- Lack of a decision-making Product Owner – PIM requires dozens of decisions about data structure. If each one requires consultation with five people, the project won't get off the ground.
- Underestimating the work involved in data processing – Data migration and enrichment typically accounts for 30–40% of the entire project time. Companies plan for 10%.
- Implementation of the "ideal system" immediately – instead of an MVP. The result: the project takes a year, and after launch, it turns out that half of the attributes are unnecessary.
- Lack of developed processes and responsibility for data after implementation – this is a mistake that implementation agencies rarely address openly. Most companies implement PIM from a situation where no one specific person is responsible for product data – it's scattered, incomplete, and owned by no one. A PIM system centralizes this responsibility in one place. This is a change for the better, but only if the company consciously implements it. Software implementation alone won't create new processes: who enters data, who verifies its quality, who approves its publication to sales channels.
How to Get Started? First Step with Cognize
At Cognize, we begin every PIM implementation with a pre-implementation analysis – a conversation in which we map your processes, data, and goals before proposing a specific solution. We don't assume you need Akeneo Enterprise. We consider Ergonode, PIMCore, and most often, we choose the free Akeneo Community. Sometimes, we choose something else entirely. If you'd like to see where to start in your company, schedule a free consultation.
