Skip to content
Home » Insight » PIM Selection Process: a 12-Stage Buyer’s Framework

PIM Selection Process: a 12-Stage Buyer’s Framework

Most failed PIM projects can be traced back to a flawed selection. The vendor was picked before the requirements were clear, demos showed scripted scenarios rather than real catalogue conditions, and the scoring framework was assembled after the favoured vendor had already emerged. A disciplined PIM selection process avoids that outcome. The 12 stages below set out how to choose a PIM in the order that actually works, drawn from the selection projects Start with Data has run for distributors, retailers and manufacturers across the UK, EU and Australia.

Before the process starts

Selecting a PIM is not a software decision. It is a data, process and organisational decision that happens to end with a software contract. Treating it the other way round is the single most common reason projects stall after go-live.

The 12 stages split into three blocks:

  • Readiness and requirements (stages 1 to 4)
  • Market engagement and evaluation (stages 5 to 10)
  • Commercial close (stages 11 to 12)

Each block has its own working group and its own artefacts. Skip a stage and the next one inherits the gap.

If the question of whether the organisation is even ready to buy is still open, start with how to tell if you’re ready for a PIM before opening this checklist.

Stage 1: Pre-selection readiness check

Before approaching any vendor, run a structured readiness check. The point is to surface the things that will derail the project later regardless of which platform is chosen.

A readiness check covers:

  • Catalogue volume today and a credible three-year forecast
  • Source systems holding product data (ERP, spreadsheets, supplier portals, DAM)
  • Channels the data has to reach (own ecommerce, marketplaces, B2B portals, print)
  • Data quality baseline: completeness rate, attribute coverage, duplication
  • Internal capacity for change management and ongoing data work

This is also where data debt gets surfaced. If the existing catalogue is 40 percent complete and full of duplicates, the issue is not the absence of a PIM. It is the data. The trade-off is covered in detail in when to fix data before choosing a PIM.

Aim for a one-page readiness brief signed off by the project sponsor.

Stage 2: Stakeholder alignment

PIM touches more functions than any other piece of mid-market software. A typical implementation pulls in ecommerce, marketing, merchandising, IT, supply chain, finance and customer service. If those functions disagree on what the PIM is for, no vendor will fix it.

The deliverable at this stage is a stakeholder map and a written charter. The charter answers four questions:

  • Who owns the PIM after go-live? Almost always merchandising or ecommerce, rarely IT.
  • What problems is it solving, in priority order? Time to launch, data quality, channel reach, cost, compliance.
  • What is explicitly out of scope? Order management, pricing, promotions, customer data.
  • Who has decision rights and who has veto rights?

A stakeholder map without a written charter is decoration. Get the charter signed before stage 3.

Stage 3: Functional requirements

Functional requirements describe what the PIM has to do for the people using it day to day. This is where most RFPs go wrong: they list 400 generic features copied from a vendor brochure and lose the specific things that actually matter to the business.

Keep functional requirements to 60 to 100 items grouped by user role. Suggested groupings:

  • Catalogue authoring (product creation, variants, bundles, kits)
  • Attribute management (mandatory attributes, conditional logic, lookup lists)
  • Taxonomy and classification (multiple taxonomies, channel-specific categories)
  • Workflow and approvals (review states, multi-step approval, role permissions)
  • Localisation (languages, regional pricing, market-specific content)
  • Asset management (DAM integration or native, image variants, video)
  • Channel publishing (which channels, which fields, frequency)
  • Reporting and dashboards (completeness, time to enrich, throughput)

Each requirement must be testable. “Strong workflow capability” is not a requirement. “Support a four-step approval workflow with role-based reassignment and audit log retention of 24 months” is.

Stage 4: Technical requirements

Technical requirements describe how the PIM has to fit into the wider stack. These come from IT and integration architects, not business users.

Areas to cover:

  • Hosting model: SaaS, single-tenant, on-premise, hybrid
  • API surface: REST, GraphQL, event streams, batch import
  • Authentication and SSO requirements
  • Existing integrations to inherit: ERP, ecommerce platform, DAM, ESB, marketplace feeds
  • Data residency and compliance: GDPR, sectoral regulations, hosting region
  • Performance: catalogue size, expected concurrent users, peak import volumes
  • Disaster recovery and uptime SLAs

Pay disproportionate attention to integrations. Industry research[1] has found that organisations typically underestimate integration costs by 30 to 40 percent at the planning stage, focusing on licensing while overlooking architecture complexity. On most mid-market PIM projects, the integration surface is the single biggest line item. A vendor with a perfect functional fit and a weak integration story will end up more expensive than the reverse.

Stage 5: RFP build

The RFP is the document the longlist vendors respond to. A good RFP is short, specific and structured for comparable answers. A bad one is 80 pages of vague feature questions that produce 80 pages of vague feature answers.

Sections to include:

  • Business context (one page on the company, channels and catalogue)
  • Project objectives and success metrics
  • Functional requirements with response format (must-have, should-have, nice-to-have)
  • Technical requirements with the same prioritisation
  • Implementation expectations (timeline, methodology, change management)
  • Commercial section (licence model, professional services, ongoing costs)
  • Reference requirements (three named clients of comparable size and complexity)
  • Response format and submission deadline
  • Length target: 25 to 40 pages. Anything longer favours large vendors with bid teams and disadvantages the leaner platforms that are often the better fit.

Stage 6: Vendor longlist

The longlist is the universe of plausible vendors before any filtering. For a UK or EU mid-market PIM selection, the longlist typically contains 8 to 14 platforms across enterprise (Akeneo, Stibo, inriver, Salsify), mid-market (Bluestone PIM, Pimcore, Sales Layer) and lighter or sector-specific options (Plytix, Bluemeteor). The 2025 Gartner Market Guide for Product Information Management Solutions catalogues the representative vendors and is a reasonable starting reference for the market shape.

Build the longlist using:

  • Independent analyst lists (Gartner Market Guide, Forrester) as a starting point only
  • Peer recommendations from comparable companies in the sector
  • Existing partner relationships if a strong technology partner is in place
  • Platform-specific fit: a Shopify retailer with 5,000 SKUs has a different shortlist than an electrical wholesaler with 300,000

Avoid letting incumbent vendor relationships shrink the longlist prematurely. The point of the longlist is breadth.

Stage 7: Shortlist

The shortlist narrows the longlist to three or, at most, four vendors that go through full evaluation. More than four creates evaluation fatigue and dilutes the depth of demos and reference calls. Fewer than three removes the negotiating leverage that comparable shortlist members create.

Shortlisting criteria:

  • Functional fit against the must-have requirements (no gaps on any must)
  • Technical fit on the integration surface
  • Sector experience with comparable buyers (size, complexity, channel mix)
  • Commercial viability (licence cost in the right ballpark for the budget envelope)
  • Vendor stability (revenue, growth, ownership structure)

The shortlist is also where bias starts to creep in. If one vendor was already known to the project sponsor, it will get unconscious preferential treatment unless the criteria are written down before the shortlist is locked.

Stage 8: Demo facilitation

Vendor demos are the most over-weighted and least informative stage of a typical PIM selection. Sales engineers demo a clean configured environment with curated data and scripted workflows. This pattern is well documented in why PIM demos don’t reflect real life.

The fix is to take control of the demo agenda:

  • Send the same scripted scenario to every shortlisted vendor in advance
  • Use a real subset of the actual catalogue, not the vendor’s demo data
  • Ask each vendor to perform the same five tasks in the same order
  • Score each task against the same criteria immediately after the demo
  • Block all sales pitches and capability tours; reject vendors who insist on running them

The five tasks should cover

  • Catalogue authoring
  • Attribute management
  • Workflow
  • Channel publishing
  • A realistic data import

Two-hour demos per vendor across two sessions are enough. Anything longer is theatre.

Stage 9: Reference calls

Reference calls are where vendor performance gets pressure-tested against real implementation experience. The trap is to accept vendor-supplied references at face value. Vendors offer their happiest clients. The signal is in the questions you ask, not the references you accept.

Useful questions for a reference call:

  • How long did the implementation actually take versus the original plan?
  • What was the all-in cost including services and internal time?
  • What broke after go-live that nobody flagged during selection?
  • How responsive is the vendor’s support team in week three of an outage?
  • Would they pick the same vendor again with the benefit of hindsight?

Aim for two reference calls per shortlisted vendor, ideally one matched on size and one matched on sector. Where possible, find an off-list reference (a peer in the same industry not supplied by the vendor) to triangulate.

Stage 10: Scoring framework

The scoring framework converts demo and RFP responses into a comparable number. Without one, the decision defaults to whoever ran the most charismatic demo.

A workable framework uses weighted scoring across four dimensions:

  • Functional fit: 40 percent typically
  • Technical fit: 25 percent
  • Implementation and partner quality: 20 percent
  • Commercials: 15 percent

The discipline is to set the weights and the scoring rubric before the demos, not after. Adjusting weights after the demos is how favoured vendors win selections they should have lost. Document the weights, lock them, then score.

Stage 11: Commercial negotiation

Commercial negotiation starts only after the preferred vendor is chosen. Running parallel commercial conversations with multiple vendors creates leverage but also signals indecision and slows the process.

Negotiation areas that move on most PIM contracts:

Annual licence fee: industry-wide SaaS benchmarks put typical discounts at 10 to 30 percent in 2025, with the upper end tied to multi-year commitment and early-renewal timing

  • Implementation services scope and rate card
  • Payment terms: annual versus quarterly, milestone-based versus upfront
  • Contract length: multi-year discounts in exchange for commitment
  • Exit clauses, data export rights and termination rights
  • SLA terms and credits for breach
  • Future price increase caps: CPI plus a fixed percentage cap. SaaS price increases ran at 9 to 25 percent in 2025 on many platforms, so an uncapped contract has real downside

Pay particular attention to data export rights. Getting product data out of a PIM should be a contractual right with a defined format and timeframe, not a vendor-discretionary courtesy.

Stage 12: Decision sign-off

The final stage is the formal sign-off. The decision paper goes to the steering committee or board with the recommendation, the scoring summary, the financial case, the risk register and the implementation plan.

The decision paper should include:

  • Recommended vendor and the runner-up
  • Total cost of ownership over five years
  • Expected business benefits and how they will be measured
  • Implementation timeline and resource plan
  • Top five risks and mitigations
  • What changes about the operating model after go-live

Sign-off is the moment the project transitions from selection to implementation. The selection team should be retained into early implementation to preserve context. Handing a fresh implementation team a contract and a vendor name with no transfer of selection rationale is how requirements risk drifting in the first 60 days.

The full timeline

A typical mid-market PIM selection runs 12 to 16 weeks from stage 1 to sign-off:

  • Stages 1 to 4 (readiness and requirements): 4 to 6 weeks
  • Stages 5 to 7 (RFP through shortlist): 3 to 4 weeks
  • Stages 8 to 10 (demos through scoring): 3 to 4 weeks
  • Stages 11 to 12 (commercials through sign-off): 2 to 3 weeks

Compressing the process below 10 weeks usually means cutting stage 1 or stage 9, both of which are the stages whose absence shows up later as project failure.

Common mistakes that derail PIM selection

Three patterns recur across selection projects that go wrong.

1. The shortcut from longlist straight to demo, skipping written requirements. The demo becomes the requirements process, and the vendor with the best demonstration of features the buyer did not know they needed wins.

2. The sponsor with a preferred vendor before stage 1. Every subsequent stage gets retrofitted to the predetermined answer. The scoring framework loses meaning.

3. Selecting a PIM before fixing the data problem. The new platform inherits the old catalogue. Six months later, the project is judged a failure because the data is still wrong, even though the platform is doing exactly what it was built to do.

For organisations starting their first PIM selection, an independent advisor adds discipline at the points where shortcuts are most tempting. Start with Data‘s PIM consulting team runs selection projects on a vendor-agnostic basis, including the readiness check, requirements build, RFP, demo facilitation and scoring through to commercial negotiation.

The PIM selection checklist

Use this as a working tracker through the process.

1.  Pre-selection readiness check: readiness brief signed by sponsor

2.  Stakeholder alignment: written charter with named decision rights

3.  Functional requirements: 60 to 100 testable items, role-grouped

4.  Technical requirements: integration map and architecture summary

5. RFP build: 25 to 40 page document with prioritised requirements

6.  Vendor longlist: 8 to 14 platforms, sector-appropriate

7. Shortlist: 3 to 4 vendors with no functional must-have gaps

8. Demo facilitation: identical scripted scenarios on real catalogue data

9.  Reference calls: two per vendor, one off-list

10. Scoring framework: weighted, locked before demos

11. Commercial negotiation: with the preferred vendor only

12. Decision sign-off: board paper with TCO, risks and runner-up

This is the spine. The detail in each stage above is what makes it work.

Key takeaways

The PIM selection process is twelve stages, not three (requirements, demos, decision).

  • Stages 1 and 2, readiness and stakeholder alignment, are the highest-leverage stages and the most often skipped.
  • Functional and technical requirements should be testable, not aspirational.
  • Demos are theatre unless the buyer controls the agenda and uses real catalogue data.
  • The scoring framework must be set before the demos, not after.
  • Commercial negotiation comes after the preferred vendor is chosen, not in parallel with the shortlist.

To run a structured selection process with an independent partner, see Start with Data’s PIM consulting practice and book a PIM Selection consultation.

External sources cited