Do you need a PIM? For most businesses that put the question to us, the answer is yes, no, or not yet. It is rarely a close call once the right evidence is on the table. The difficulty is that almost everyone you can ask has a reason to say yes. Vendors sell platforms. Integrators sell projects. We are a partner to six PIM vendors, so we have no reason to steer you towards one. It is also why we are willing to tell a client the answer is no.
What follows is the diagnostic we run in a first conversation. Symptoms first, then the misdiagnoses, then the thresholds, then the timing.
Do you need a PIM? The symptoms that usually mean yes
Nobody wakes up wanting product information management. Something breaks, and the search starts from there. These are the patterns we see repeatedly across product data services engagements with retailers, distributors, and manufacturers.
Channel proliferation outpacing data quality
A website and a few retail accounts become a marketplace, a B2B portal, and an international site. Each wants the same products described differently, with its own attribute set and taxonomy. Within a year the channels disagree with each other. Nobody caused it. The process was built for two destinations, not six.
Supplier sprawl overwhelming merchandising
Range growth brings supplier growth. Some send spreadsheets, some send PDFs, some expect your team to scrape their website. We have walked into merchandising teams where most of the week goes on normalising inbound files. The question of what should be in the range has quietly stopped being asked.
The spreadsheet breakdown moment
Almost every business we meet has a master file. It started small. It now has thirty tabs, tens of thousands of rows, and lookups that one person understands. The breakdown arrives when that person leaves, or when two people edit at once and changes vanish. Sometimes it is a formula failing silently, publishing a thousand records with the wrong values.
Launch times stretching with no obvious cause
A new product used to take two weeks and now takes six. The data is touching more hands, and each handover costs a day. Businesses rarely notice until someone measures it.
Quality complaints arriving from outside
A marketplace warns that listings will be delisted for incomplete data. A customer points out that the dimensions are wrong. These are lagging indicators. By the time they reach the board, the merchandising team has usually known for a year.
Internationalisation pressure
A new market needs a new language, new regulatory attributes, and a local taxonomy. Translations get done in spreadsheets and are never updated when the source changes. The local site drifts further from the master every month.
One symptom is normal. Three or more, persistently rather than in a bad quarter, is a PIM-shaped problem.
Do you need a PIM, or do you need something else first?
For every pattern that genuinely calls for a platform, there is a lookalike that does not. Failed projects usually begin here. A PIM enforces structure on what you give it, so a broken input becomes a broken input inside a more expensive system.
- Bad taxonomy. An incoherent classification will be held more rigidly, not fixed. We see categories that grew organically over a decade: some defined by supplier, some by application, some by material. The same fixing then sits in three categories, or in none. Rebuild the taxonomy first and the platform conversation goes better afterwards.
- A broken merchandising process. A PIM enforces process. It does not invent one. Install it over an undocumented workflow and you automate the absence of process, so the same gaps happen faster.
- A content gap. A PIM holds attributes and tells you what is missing. It does not write buying guides or commission photography. Thin product pages stay thin, more efficiently.
- An asset problem. If the real complaint is that nobody can find the right hero image, that is a DAM requirement. A PIM with a lightweight asset feature will not settle it.
- A governance vacuum. This is the most damaging misread. Nobody owns product data, and no one has authority to say what good looks like. A PIM assumes that ownership exists and gives it tools. Drop one into the vacuum and the data stays exactly as inconsistent inside the platform. The only difference is that everyone must log in to discover it.
None of this work is wasted if you buy later. A clean taxonomy improves the site this quarter. A named data owner reduces escalations this quarter. All of it is wasted if you skip it and expect the platform to compensate.
Do you need a PIM at your scale? The thresholds that matter
No single number decides this. The dimensions matter in combination. Two businesses can hold twenty thousand SKUs each and have almost nothing in common. One sells across three channels in two countries. The other buys from two thousand suppliers.
| Dimension | Usually manageable | Conversation starts | Why it bites |
|---|---|---|---|
| SKUs | Under 2,000, stable | 5,000 to 10,000 attribute-rich | Variants of a few styles are easier than genuinely distinct products |
| Channels | One or two | Three or four | Each destination adds its own rules, and the fourth feels harder than the second |
| Suppliers | Up to about 50 | Around 200 | The cost is the number of inbound formats, not the volume |
| Languages or markets | Two, with care | Four | Every market multiplies the records you keep in sync |
| Attributes per SKU | Three to ten | Forty or more | Technical and regulated categories multiply again by channel |
Count the attribute-rich products rather than the rows. A fashion retailer with two hundred thousand SKUs can have a simpler problem than an industrial supplier with twenty thousand. The fashion lines are variants of a smaller number of styles.
The combinations that consistently produce a PIM-shaped problem are recognisable. A retailer above ten thousand attribute-rich SKUs, more than three channels, and a new market planned. A distributor above two hundred suppliers with a B2B portal alongside the main site. A manufacturer with several brand sites, several languages, and a growing direct channel. These are common shapes, not edge cases.
If your combination sits nowhere near these, the platform is unlikely to be your answer. It is worth reading where PIM, MDM, and ERP actually divide before going further.
Do you need a PIM now, or in eighteen months?
Current state is a snapshot. Growth rate is the signal. A business doubling its SKU or channel count each year will cross the thresholds wherever it sits today. A stable business sitting just above one can usually wait another planning cycle.
Two things push in opposite directions here, and both are real.
Waiting is not free. Every product added without structure adds to the eventual cost of organising it. Every supplier onboarded without a clean attribute mapping adds friction that gets unpicked later. Migrations get harder, not easier. SKUs grow, channels grow, integrations grow, and so does the number of people whose daily workflow depends on the current process.
Buying too early fails differently. The platform goes in. The team was not part of the decision, and the workflows in the tool do not match the ones they built. Training was thin. Eighteen months later, two people use it for one report and everyone else is still in the spreadsheets. Worse, a failed implementation spends the organisation’s appetite for a second attempt. The technology improves. The internal case gets harder.
Position yourself against the thresholds twice: where you are now, and where you will be in eighteen months. A project that goes well takes around a year, so the system has to fit the business you are becoming. We have written honestly about how long a PIM implementation takes if that horizon matters to your planning.
When the honest answer is no
Some businesses should not buy, and saying so early saves everyone a year.
A catalogue under roughly a thousand SKUs in a single stable category does not justify the platform. Neither does a single-channel operation with no expansion planned, unless that channel is growing fast in attribute richness. A well-organised spreadsheet runs a small catalogue for a long time. The honest risk there is the single-person dependency rather than the tool itself.
Then there is the business whose real bottleneck sits elsewhere. We have looked at operations told by a vendor that a PIM would fix commercial performance. The real constraint was range, logistics, or marketing. Product data infrastructure solves none of those, and buying it now spends the capital the actual problem needs.
The last two are about capacity rather than fit. A business with no owner for product data will see the platform rejected by the team meant to use it. A business that has just replatformed, migrated its ERP, or absorbed an acquisition has no bandwidth for another change programme. Change fatigue is real, and twelve to eighteen months of consolidation is usually the right call.
What to do when the answer is not yet
The waiting period is preparation, not delay. Four pieces of work carry their own value and make any future migration faster.
- Rationalise the taxonomy. Anything built on an incoherent structure inherits the incoherence.
- Document the merchandising process. The platform has to enforce something specific. Define it first.
- Name a senior data owner. Someone needs authority to set the standard and escalate when it slips.
- Measure the data you have now. Know your completeness gaps before buying a system to close them. Tightening supplier data onboarding at the same time reduces the mess arriving in the first place.
The most successful projects we have worked on were the ones where the business spent a year deliberately preparing before procurement began.
Where this leaves you
If the symptoms persist, the thresholds are crossed, and someone owns the data, the question moves from whether to which. Our twelve-stage selection framework is the next read. If the foundations are missing, the preparation above is the work, and it stands on its own merits.
We run the selection and implementation work for retailers, distributors, and manufacturers, and a fair proportion of those conversations end with us saying not yet. If you want to test your own combination against the thresholds, book a thirty-minute discovery call and we will talk it through. Our PIM and PXM services start from whichever answer is true for you.