In short
We have seen the same five failure modes across merchant and rural retail catalogues on both sides of the world.

A builders’ merchant catalogue looks simple from the outside. Timber, bricks, insulation, fixings. Hardly the stuff of complex data architecture. But then, you try to model it:
Most building supplies PIM projects underestimate all of this, and the catalogue workstream is where the time and budget go. We have seen the same five failure modes across merchant and rural retail catalogues on both sides of the world.
Three structural facts shape merchant catalogue product data, and none of them apply to a typical retailer.
1. Consolidation
The UK merchant sector has spent a decade buying itself. Huws Gray traded from 113 branches in early 2021, then bought Grafton Group’s UK merchanting businesses, including Buildbase and Civils and Lintels, in a £520 million deal covering around 200 branches. The group now trades from more than 330 branches across several brands. Every acquisition brings another item file, another ERP, and another set of supplier codes for the same products.
2. Range breadth
A merchant sells heavy side and light side together: aggregates by the tonne in the yard, door furniture by the each at the till.Very few retail businesses force one catalogue to span such structural variety.
3. The customer mix
Bowens, a fourth-generation family business with 20 stores across Melbourne and regional Victoria, serves professional builders first. Most UK merchants serve the trade and the DIY walk-in from the same catalogue. Two audiences, two price books, one product record.
These three facts produce the failure modes below. Each one is avoidable, and each one is expensive to fix after go-live
A merchant price is not a number. It is a matrix:
None of that logic belongs in a PIM, and teams who try to make the PIM the pricing engine build something fragile that fights the ERP for the rest of its life.
The opposite mistake is more common. Teams declare pricing out of scope and then ignore the product data that pricing depends on. Price unit and sell unit are product data.
If the PIM does not hold unit of measure, pack quantity and conversion factors cleanly, every downstream system invents its own version, and sooner or later the website shows a price the till cannot honour.
The fix is a clean boundary. Pricing logic lives in the ERP or pricing engine. The structural data that pricing depends on, meaning units, packs and conversions, lives in the PIM as mandatory attributes, agreed once and used everywhere.
In a fashion catalogue, a SKU is a size and colour of a garment. In a merchant catalogue, the same physical product can be five sellable things.
Merchants typically inherit three different ways of modelling this, often all at once after acquisitions: separate SKUs per sell unit, a single SKU with unit-of-measure conversions, or variant structures borrowed from retail platforms. Any of the three can work. Mixing them is what fails. Mixed models mean:
Duplicate detection breaks down
Site search returns three versions of the same product
Stock figures disagree depending on which record you ask
The fix: Pick one pattern, hold the conversion factors as attributes, and migrate the legacy records to it deliberately. This is unglamorous taxonomy work, and skipping it is one of the places where most product structures go wrong.
A national website backed by 300 branches is not one shop. It is 300 shops sharing a catalogue. Branches range differently. A branch in north Wales stocks different roofing lines from one in Kent. Agricultural products sell in rural branches and sit dead in urban ones. Acquired brands carry overlapping but not identical ranges.
A PIM does not hold stock and should not. It should hold ranging:
Projects that skip this publish a single national catalogue, then spend months fielding complaints from branches selling products the website does not show, and from customers ordering products their branch does not carry.
Click and collect makes the gap visible within weeks. The website promises a product at a branch that has never ranged it, or hides a product the branch sells every day because the national catalogue never captured it. Both errors cost trust with exactly the customers a merchant most wants online: regular trade accounts.
This matters double after consolidation. When one group trades through several fascias, the same product often exists in multiple item files with different codes and descriptions.
The fix: Deduplication plus explicit assortment flags is the only way one catalogue can serve many brands without flattening them into one.
Construction products carry more compliance data than almost anything else sold online:
Moreover, the regime is tightening rather than relaxing. The government’s Construction Products Reform White Paper[1], published in February 2026, proposes a General Safety Requirement covering all construction products, where today only around a third fall under designated standards.
Scrutiny since the Grenfell fire has pushed traceability questions down the supply chain to the merchants who sell products as well as the manufacturers who make them. The direction is clear even before the new rules land.
The failure pattern is always the same. Certificates live in email attachments and shared drives. Nobody links them to SKUs. Nobody tracks expiry dates. Then a customer, a building control officer or a trading standards query asks for a Declaration of Performance, and the answer takes three days of searching.
Compliance data is product data:
These are all attributes from the first data load and cause a missing certificate to block publication for regulated categories. Retrofitting this after go-live costs a multiple of doing it during implementation, and the regulatory direction of travel means you will have to do it eventually either way.
Most published PIM advice assumes you are a brand or retailer selling finished consumer goods. Invest in lifestyle imagery. Write persuasive descriptions. Syndicate to marketplaces. Optimise the conversion funnel.
A site manager buying carcassing timber at 7am does not read persuasive descriptions. They search by specification: dimensions, grade, treatment, length. They buy on account price and branch availability. The enrichment that matters in a merchant catalogue is:
Delivery pricing depends on all these, and a spec table always beats a story.
Structured classification makes the difference. Electrical wholesale has ETIM as a shared standard. Building supplies has no single equivalent, which makes a well-designed attribute schema more important for a merchant, not less. Nobody hands you the template.
Vendor demos compound the problem, because most run on fashion or FMCG datasets where the hard merchant cases never appear. We have written separately about why PIM demos do not reflect real life, and about how product data needs differ between B2B and B2C. Merchants are the sharpest example of the split. The same catalogue must serve a DIY customer who wants reassurance and a tradesperson who wants a fire rating before 8am.
Each failure mode is the same mistake in different clothes: treating a merchant catalogue as a retail catalogue with more rows. It is not. It is a B2B materials catalogue that happens to have a retail storefront. Units, pricing structure, branch ranging and compliance are not edge cases to handle in phase two. They are the data model.
Get them into the foundation and the consumer layer on top is straightforward. Leave them out and no amount of content enrichment rescues the project. This is the pattern behind many of the stalled implementations we get called into, and a large part of why PIM projects stall after implementation: the system went live on a model built for somebody else’s catalogue.
We have done this work with merchants and rural retailers, including Bowens in Australia and Mole Valley Farmers in the UK (see our case studies). The projects that work share four habits:
Supplier data is the other half of the job. A merchant catalogue is assembled from hundreds of manufacturer feeds in inconsistent formats, and a repeatable supplier data onboarding process does more for catalogue quality over time than any one-off enrichment push.
None of this requires exotic software. The mainstream PIM platforms can all model a merchant catalogue well. The difference is whether the implementation asks them to. That is a question of experience, which is why our PIM consulting engagements with merchants start with units, packs and ranging, not with feature lists.
If you run a merchant or building supplies catalogue and any of these five patterns feels familiar, the cheapest time to fix it is before the next phase of your PIM programme, not after. See how we work with building supplies merchants, or read how our PIM consulting team starts every merchant engagement with units, packs and ranging. Book a thirty-minute call and we will tell you which of these five you are most exposed to.
Fill in your details and we will send it straight over.
Sourcing, supplier onboarding and enrichment for product data, delivered as a service.