PIM implementation: phases, timeline, cost drivers and life after launch

A PIM implementation is the project that sets up a product information management platform around your product data: it designs the data model, cleans and migrates existing data, connects the PIM to your ERP, DAM, commerce site and channels, and trains the team that will run it. Most implementations follow four phases: discover, design, build and migrate, then launch and adopt. How long one takes depends less on the platform than on three things: how many systems it must connect to, the condition of your source data and how many channels the first release has to serve.

Still choosing a platform? Start with How to choose a PIM. This guide picks up once the platform decision is made.

The four phases at a glance

1. Discover: agree the outcomes, map data flows and integrations, and set the scope of the first release.
2. Design: model taxonomy, attributes, workflows and governance, and get the model approved.
3. Build and migrate: configure the platform, build integrations and load cleansed data in increments.
4. Launch and adopt: train each role, go live, support the first weeks and hand over to the team that will run it.

The phases overlap in practice. Data clean-up often starts during discovery, and later releases go through design and build while the first release is already live.

Phase 1: Discover

Discovery turns the reasons you bought a PIM into a plan the whole team can work from. It usually runs as a series of workshops with the people who create, approve and publish product data, and with the IT owners of each connected system.

• Outcomes: what should be measurably different after launch, such as time to publish a new product or the share of products that meet each retailer’s requirements.
• Data flows: where each type of product data starts (ERP, PLM, supplier files, a legacy PIM or DAM) and where it has to go.
• Integrations: every system the PIM must read from or write to, with the owner of each and how data will move.
• Data condition: a sample-based review of completeness, duplicates and inconsistent values in the source data.
• First release scope: the product lines, channels and teams included first, and what is deliberately left for later.

The output is a scope and roadmap that business and IT both sign. If your team has not yet assessed its readiness, the PIM readiness guide covers the questions to answer before this phase starts.

Phase 2: Design

Design decides how product information will be structured and governed. Decisions made here are the most expensive to change later, so this phase deserves senior attention from your side.

• Taxonomy and data model: categories, product families, attributes, variants, relationships, units of measure and inheritance. Our article on building a good product data model covers the common missteps.
• Channel requirements: how each retailer, marketplace, data pool or site maps to your attributes, including industry standards such as GS1 attributes for GDSN.
• Workflows: who creates, enriches, reviews and approves each part of a product record, and in what order.
• Governance: validation rules, completeness rules by channel, roles and permissions, and who owns the data model after launch.
• Integration design: the direction, frequency and format of each data flow, and which system is the source for each attribute.

The output is an approved data model and integration design. Test it early with a handful of difficult real products, such as a kit, a variant-heavy item or a product with complex specifications, rather than only typical ones.

Phase 3: Build and migrate

In this phase the platform is configured, integrations are built and data is loaded. The work is usually delivered in short iterations so business users see real products in the PIM early and can correct the design before it is fixed.

• Configuration: attributes, families, validation rules, workflows, roles and user interfaces.
• Integrations: connections to ERP, DAM, commerce and channel systems, built and tested with real data. See our integrations practice for common patterns.
• Data migration: extract, cleanse, map and load data in increments, with reconciliation reports that show what loaded and what failed. Our data migrations team treats migration as its own workstream, because it usually is one.
• Channel testing: channel mapping and syndication tests against each destination’s requirements before launch.
• User acceptance testing: business users run agreed scenarios on real products and sign off.

Integrations and data migration are where most timelines slip. Start both as early as the design allows, and give each an owner on your side as well as on the partner’s.

Phase 4: Launch and adopt

A launch is successful when your team uses the PIM as the place where product data is managed, not when the platform is switched on.

• Training by role: administrators, data stewards, content editors and approvers each need different training, using your own products.
• Cutover plan: the order in which channels and teams move to the PIM, and when old spreadsheets and tools are retired.
• Launch support: a defined period of close support after go-live to fix issues quickly and answer questions.
• Handover: documentation of the data model, integrations and runbooks, and a named owner for each.

How long does a PIM implementation take?

There is no standard duration, because the scope varies so much between organizations. Two examples from Sitation projects show what a focused first release can achieve:

• Lordco Auto Parts completed its Akeneo implementation in 12 weeks, loading 862,000 product records and processing 55 million fitment records to support its eCommerce launch (case study).
• Warren Rupp launched Salsify in four months, connected to JD Edwards, Bynder and SAP Commerce, and its model became the preferred PIM approach for 54 IDEX Corporation subsidiaries (case study).

Both had a clear first-release scope. Enterprise programs that cover several data domains, regions or business units usually run as a series of releases rather than one long project.

These factors have the most influence on the timeline:

What drives the cost of a PIM implementation

Implementation cost is separate from the platform subscription, and it varies with the same factors as the timeline. When you compare proposals, check that each covers these items on the same assumptions:

• Discovery and design: workshops, data model and integration design.
• Configuration: setting up the platform, workflows, rules and roles.
• Integrations: each connection to ERP, DAM, commerce and channel systems, including testing.
• Data migration and clean-up: often underestimated, and often the largest variable.
• Channel set-up: connectors, syndication mapping and testing for each destination.
• Training and change management: for each role that will use the PIM.
• Your internal time: product, IT and content people who take part in workshops, testing and decisions.
• After launch: administration, support and new releases, handled in-house, by a partner or both.

Platform pricing models differ by vendor. For one example, our guide to how Salsify pricing works explains what drives that platform’s subscription.

How to keep an implementation on schedule

• Scope the first release tightly. One product line and one or two channels prove value sooner and teach you what to change before you scale.
• Start data work in discovery. Profiling and cleaning source data early prevents the migration from becoming the critical path.
• Name the decision owners. A product owner who can make data model and workflow decisions saves weeks of waiting.
• Test with real products. Include your most complex products in design reviews and acceptance tests.
• Build integrations in parallel. Once the integration design is agreed, integration work does not need to wait for every configuration detail.
• Plan the day after launch. Decide who will administer the platform before go-live, not after.

Our article on the value of project management in PIM implementations describes how we run this cadence in practice.

What happens after launch

Go-live is the start of the PIM’s working life. In the first weeks, expect a period of close support while teams adjust and small issues are fixed. After that, the platform needs regular attention: new attributes and channels, retailer requirement changes, platform upgrades, user support and data quality monitoring.

Organizations usually choose one of three ways to run the platform:

• In-house: a dedicated PIM administrator and data stewards, trained during the implementation.
• Managed services: a partner runs and improves the platform under an agreed scope. Sitation’s PIM managed services cover platform operations, taxonomy, syndication and reporting.
• A mix: your team owns the content and decisions, and a partner handles administration, integrations and new releases.

Plan the next releases as part of the roadmap from discovery, so the PIM keeps expanding to new product lines, channels and uses such as AI for product content once the data foundation is in place.

Common implementation mistakes

• Migrating data as it is. Moving poor data into a new platform moves the problem too.
• Treating integrations as a detail. They often decide the timeline and the budget.
• Designing only for typical products. Kits, variants and complex specifications break models designed for the easy cases.
• Launching every channel at once. A smaller first release goes live sooner and informs the rest.
• Leaving users out until training. Content and category teams should shape workflows during design.
• No owner after launch. Without a named owner, data quality declines and the PIM is bypassed.

PIM implementation checklist

• Outcomes agreed by business and IT, with measures you can track after launch.
• Data flows, source systems and channels mapped, with an owner for each integration.
• Source data profiled, with a clean-up plan and owner.
• First release scoped to specific product lines, channels and teams.
• Data model and workflows approved, and tested with difficult real products.
• Integration design agreed, with the source system named for each attribute.
• Migration in increments, with reconciliation reports.
• Channel mapping tested against each destination’s requirements.
• User acceptance scenarios run on real products and signed off.
• Training by role, a cutover plan and a defined launch support period.
• A named owner for the platform after launch, in-house, managed or both.

How Sitation helps

Sitation helps retailers, distributors, brands and manufacturers put their product data to work, with PIM, MDM and DAM implementation, managed services, AI and Plezio software. More than 400 of them have chosen Sitation as their partner. We implement Salsify, Akeneo, Stibo Systems and Syndigo, as a Salsify Platinum Partner with more than 200 Salsify implementations, an Akeneo Gold Partner, a Stibo Systems Gold Partner and a Syndigo Silver Partner. Our PIM implementation service covers discovery, data model design, integrations, data migration, training and launch support, and our managed services team can run the platform after launch.

Frequently asked questions

How long does a PIM implementation take?

It depends on the number of integrations, the condition of the source data and the number of channels in the first release. As examples, Lordco Auto Parts completed its Akeneo implementation in 12 weeks, and Warren Rupp launched Salsify, connected to JD Edwards, Bynder and SAP Commerce, in four months. Larger programs are usually delivered as a series of releases.

How much does a PIM implementation cost?

Implementation cost is separate from the platform subscription and depends on scope: discovery and design, configuration, each integration, data migration and clean-up, channel set-up, training and support after launch. Ask each partner to price the same scope and assumptions so proposals can be compared line by line.

What are the phases of a PIM implementation?

Most implementations run in four phases: discover (outcomes, data flows and scope), design (data model, workflows and governance), build and migrate (configuration, integrations and data loads), and launch and adopt (training, go-live, launch support and handover).

Should we clean our product data before migrating it to a PIM?

Yes, as far as possible. Profile source data during discovery, fix the issues that would break the new data model, and migrate in increments with reconciliation reports. Some clean-up is easier inside the PIM once validation rules are in place, so agree which issues are fixed before and after migration.

Can a PIM be implemented in phases?

Yes, and it usually should be. A first release that covers one product line and one or two channels proves value sooner. Later releases add product lines, channels, regions and new uses, following the roadmap agreed in discovery.

Who should run the PIM after launch?

Either your own team, a managed services partner or a mix of both. Decide before go-live, name an owner for the data model and data quality, and make sure the chosen team is trained during the implementation.