Bring your worst supplier file to the PIM demo

If you are choosing a PIM this quarter, stop scoring demos and start scoring scenarios. Pick five tasks that will decide whether the project works, write them down with your own data, and make every vendor perform them in the same order with the same file. Whatever a vendor refuses to run belongs in the score too.

Syndigo made this case in a long buyer's guide published on 13 October. Syndigo sells a PIM and would like to be on your shortlist, so read it with that in mind. The advice still holds, and it is more honest than most vendor content: three vendors, three demos and three strong scores tell you nothing, because each vendor picked the data, set the conditions and decided where the session stopped.

Why demos all look the same

A prepared demo is built to succeed. The record is complete, the category exists, the mapping was done last week and the exception path never shows up. Nothing in it is false. It answers a question you never asked.

Feature questionnaires fail the same way on paper. A "yes" to localisation or workflow only tells you the capability exists somewhere in the product. It does not tell you whether it covers your variant structure, whether it is in the price you were quoted, or whether someone on your team can configure it.

The result is a familiar Thursday afternoon after the last demo. The buying group argues about which presenter seemed more confident, because that is the only thing that actually differed.

My view is simple. Each vendor's demo should be treated as marketing, and the scenarios you script as the evaluation.

What to put in the script

Write your scenarios before the first vendor call and keep them fixed. Syndigo's guide lists seven elements per scenario. In practice four matter most: the real input, the workflow, the thing you will deliberately break, and what a correct result looks like, written down beforehand.

• Your worst supplier file. Bring the spreadsheet with merged header rows, two units in one column and a batch of likely duplicates. Watch the mapping happen live and ask who on your team would do that work every week. In Part 1 of our Trends in PIM series we argued that supplier onboarding is splitting off into its own layer in front of the PIM. That makes this test sharper: if the vendor says onboarding happens in a separate module or partner tool, price and test that tool too.
• A record that has to fail. Ask to see a validation rule reject something. Check where the failure shows up, what the message says to a non-technical user, whether the record can move on anyway, and who is allowed to change the rule.
• A task nobody picks up. Assign an enrichment task and let it sit. See who gets reminded, and whether your own admin can change the workflow steps without a services contract.
• An AI agent that breaks a rule. Have an agent or AI feature write a value it should not be allowed to write, such as a change to a regulated attribute or a locale outside its scope. In Part 2 of the series we argued that enrichment agents need AI governance, with the PIM enforcing who may change what. A demo where the AI only fills in empty descriptions shows none of that. Ask to see the blocked write, the audit entry and the rollback.
• The Monday remap. A retailer changes its required attributes. Ask the vendor to absorb that change during the session, then show you the export status and who owns the failed records.

If a vendor asks to swap in its own sample data or skip the exception, write that down. It is a finding.

Score what you would actually buy

Keep two lines separate on your scorecard. The first is core PIM: the data model, validation, workflow and change history. The second is everything sold near it, such as supplier portals, DAM, syndication and AI features. For each item, note whether it is included in the quote, needs another module or partner, needs custom work, or sits on a roadmap without a date.

Then follow one rule. A strong add-on never makes up for a weak essential. A beautiful analytics screen does nothing for a data model that cannot hold your variants.

Score your own capacity too. If the people meant to prepare data and own the rules are already fully booked, that will hurt the project more than any feature gap. Write it down as a risk before the contract does it for you.

Where analyst rankings fit

Syndigo was named a Leader in Gartner's 2026 Magic Quadrant for PIM the day before it published this guide. Its own guide says a quadrant can help build a shortlist and cannot tell you which vendor fits your catalogue. I agree. Use rankings to decide who gets invited, then let your file decide who wins.

The trade-off

Scripted scenarios take work. Expect a week or two of preparation to anonymise real files, agree priorities across teams and write expected outcomes. Some vendors will push back or ask for a paid proof of concept. You will also run fewer vendors through it, three at most.

That cost is small next to the alternative, which is finding out about your variant model in month four of implementation. Scripting also brings an internal argument forward. If merchandising and e-commerce cannot agree which scenarios matter most, you learn it now, when it is cheap.

Start this week. Open last month's worst supplier file and save a copy as demo scenario one.