How to choose a PIM solution for your product data

You’ve updated a product, but customer service still finds the old information in the customer portal. The reseller has received a different version, and marketing is waiting for the right image. Someone has to compare the files before the next publication.

A PIM solution can give you control over that flow. Its value depends on which errors you need to prevent, who is responsible for the information and how changes reach your channels. That’s why the technology choice should start with your actual data errors.

What a PIM solution is and where it sits

The same product can have different names in the business system, on the website and in the reseller’s item file. The customer portal may show technical dimensions, while the website needs images and descriptions of use. The differences can be intentional, but the product’s properties still have to match.

PIM is a system for structuring, enriching and distributing product information. That is where you manage things like copy, attributes, dimensions, images and documents. The solution acts as a hub between the business system and the channels where customers or resellers use the information.

The business system should remain responsible for internal item numbers and transactional data. The PIM needs to be able to link that information to the content each channel requires. A technical name can then be kept internally while the customer gets an understandable product description.

Product information is part of the customer’s ability to find, compare and buy. Wrong dimensions or missing documents therefore need to be treated as business problems.

Start by separating a missing piece of information from a failure in distribution. If the right information already exists with the product manager, you need to look at how it reaches the channels. If no one knows the correct value, a decision on who should determine it comes first.

Three business problems a PIM solution solves

The first problem arises when staff have to search for product information in spreadsheets, emails and the business system. Imagine a manufacturer where the product manager emails new dimensions to marketing. The customer portal is updated, but the reseller file still has the old information. Record how many manual corrections each publication requires and which task causes them.

The second problem is that channels use different versions. An item can be complete on the website while the reseller is missing packaging dimensions. Measure the share of complete items per channel, based on each channel’s mandatory fields. The reasoning about PIM for e-commerce needs to cover the whole route to the recipient’s accepted item.

The third problem shows when a change in the business system has the wrong consequences further down the line. A changed unit can make a pack appear as a single product. Prices and stock levels may also follow their own flows that need to be troubleshot separately. Track the time from an approved change to publication and log rejected reseller files with clear reasons.

Decide what counts as a correction and when publication time starts. Use the same definition before and after an implementation. A PIM can help you detect and stop errors, but incorrect source values still need to be corrected by the person responsible.

A data model and enrichment that hold up across channels

Start the data model with the product’s identity. Where GTIN is used, it should be a stable identifier kept separate from the business system’s internal item number. The link between identifiers should be explicit, so channel presentations can change without items getting mixed up.

GS1 Sweden describes GDSN as a network for sharing digital product information through certified data pools. In Sweden, Validoo is the GS1-certified data pool. For that kind of exchange, internal attributes need to be mapped to standardised information on identity, logistics, marketing information, images and sustainability.

Then build checks based on concrete errors. An item with an empty material field is stopped if the recipient requires material. A length value without a unit needs to be completed before it can be used safely. Language and dimension formats should also be checked, so that Swedish descriptions or dimensions in the wrong format are caught before delivery.

Relationships need their own rules. A product image on the base item may suit several variants, while dimensions and packaging details need to be linked to the right item. If a variant lacks documentation, the system should show that clearly. Automatically inheriting a document could otherwise give the customer instructions for the wrong version.

Then enrich with approved copy, images and documents per channel. Keep that work separate from pricing rules and stock information, which may need different update flows. Version the mappings and put failed transfers in an error queue with re-runs. The person responsible can then fix the cause and resend the affected items without redoing the whole publication.

ERP integrations and implementation options

An integration needs to describe which information is moved, who owns it and how errors are handled. Product content covers copy, attributes, images and documents. The business system’s flows may cover item numbers, base data, pricing logic and transaction-related information. These flows need to be changeable and testable with their different uses in mind.

Introducing a new PIM and a new website at the same time gives you the chance to shape the data model around the same channel needs. At the same time, the customer has to make more interdependent decisions during implementation. You need to set aside time for product structure, content and testing in the new web environment. This approach suits cases where both solutions need to change and those responsible can work in a coordinated way.

Introducing PIM alongside an existing website limits the scope of change. The customer then needs to document the website’s current fields and find out what limitations exist on the receiving side. The deeper article on integration with the business system is relevant when you need to work out dependencies between systems. Also check whether the existing website can receive the item relationships you need.

A ready-made integration with a fixed scope suits cases where your needs fit within the agreed flows. With us, that approach is built on the Function Library’s tested, compatible functions and integrations at fixed scope and fixed price. The customer still needs to appoint a data owner, describe channels and attributes and take part in mapping and testing.

A small range with few changes can work well enough in a controlled spreadsheet. If no one can decide which source field applies, you first need to sort out the base data. A new connection would otherwise spread the same uncertainty to more recipients. The technology choice can wait until those responsible can describe what should be transferred.

Governance and AI in product data after go-live

After go-live, someone needs to be able to decide when information is contradictory. Decide ownership per field, who may make changes and who approves publication. The change log should show what was changed, by whom and from which source. That helps those responsible investigate a discrepancy without comparing old emails.

An update to the business system can affect a field the PIM uses. Ongoing management therefore needs to include checks of changed formats, mappings and how channels receive the data. We take long-term product ownership after launch. The boundaries of responsibility need to be agreed so the customer knows who investigates errors between systems.

Statistics Sweden reports that 35.0 percent of Swedish companies with at least ten employees used AI in 2025, compared with 25.2 percent in 2024. The figures show use of the technology and give no measure of the quality of product copy. You need to assess AI-generated content against the same source requirements as other content.

A practical rule of thumb is to start with fields where errors have limited consequences. AI can, for example, suggest a shorter headline based on approved properties. Information on safety, compatibility and sustainability should require human approval. This is a recommended working rule, not a measured effect.

The PIM can then act as a control layer between source information, AI and channels. Save the material each suggestion is based on and keep suggestions separate from approved content. Track error rates, approval times and returns or support cases related to product data. Only when you measure these values in your own flow can you judge what automation actually delivers.

Scenarios where product data has to serve several channels

A manufacturer may have few purchases on its own website and still send product information to many recipients. The customer portal needs technical documents, the reseller needs packaging information and salespeople need material for quotes. Statistics Sweden reports that 21 percent of manufacturing companies sold via their own website, app or e-commerce platform in 2024, compared with 30 percent overall. Start by deciding which recipient first needs reliable information.

A web shop that also publishes on a marketplace may need several languages and different categorisations. The same colour can be described freely in your own shop but have to be chosen from a fixed list of values by the recipient. In that setup, a correct internal value can still lead to a rejected item. The choice of first channel should be based on where you spend the most effort correcting deliveries.

In a group, several brands may share components but use different names and images. The shared component’s properties need to be kept together, while each brand is responsible for its own presentation. If each company copies the whole record, it becomes hard to see which information is genuinely shared. Prioritise the channel where a shared product structure can be tested with clear responsibility.

Work on the digital product passport also needs to be tied to the product’s identity and available source material. Keep documented properties separate from marketing copy when you map the material. Then decide which recipient needs the material first and which information you can actually substantiate.

Three questions to ask the supplier

Ask the supplier to describe how internal item numbers are linked to GTIN and what happens when a mapping fails. The answer shows whether there is a working error queue, a test environment and a practical way to resume the transfer. A statement that the systems can be connected isn’t enough of an answer about error handling.

Base the walkthrough on a real item from your range. Ask to see a failed flow and who gets notified. That makes it clear whether your staff can handle the deviation or need technical help every time.

Then ask for a clear answer on which fields may be changed in the PIM and which the business system always owns. The answer reveals whether the supplier has defined responsibility when several systems hold the same information. A useful answer should be something you can hand to the product manager and the finance manager without them having to interpret technical assumptions.

Finally, ask the supplier to explain what happens to mappings and channels when the business system or platform is updated. The answer shows whether follow-up, compatibility and bug fixing are covered by a continued commitment after go-live. It should also be clear which customer changes require a new decision on scope.

We set out our long-term product ownership in the agreement, including what the customer needs to contribute. The same clarity is needed whoever you choose. If the answers stop at launch, you still have no answer on who handles the next version change.

Start with your product information as the next step

A useful starting point can be built around a representative product family. Choose items with variants, documents and channel-dependent information. Include a known error that staff regularly have to correct. You can then test how a proposed way of working handles a problem you already understand.

Gather the source information, the published version and the recipient’s requirements. Describe who does the work and where someone needs to make a manual decision. The material should show what those responsible need to be able to check when an item leaves the PIM. Also keep examples of what works well and should be kept.

The deeper look at product information and enrichment in PIM gives you a picture of how content and syndication can be connected to business systems and channels. Use it as support when you assess your own flow. The key is being able to follow a piece of information from its source to the place where the customer uses it.

Also set aside time with the person who knows the products and the person responsible for the systems. Together they need to be able to judge whether a proposed field can be filled with reliable information. A clear scope makes it easier to assess both the way of working and what the organisation needs to contribute.

Our PIM guide walks through the data model, responsibilities and flows step by step. Download the guide to PIM in practice and bring a product family and a concrete publishing error to a free conversation with one of our PIM experts.