Fitment search

Parts lookup by machine is the feature that decides everything.

Dealers usually think the risk in a parts store is the catalog size or the payment setup. It is neither. It is whether a customer standing next to a broken machine can prove to themselves that the part fits before they click buy.

Talk to us

Get your parts store scoped.

A real reply from an engineer. We won't put you on a mailing list.

Book a 30-min call

Two customers, two completely different searches

There are only two people buying parts from you online, and they behave nothing alike. Building for one and ignoring the other is the most common design mistake in dealer ecommerce.

The first knows the part number. They want the fastest possible path from a number to a checkout, and every extra step loses them. The second does not know the part number and never will. They know the machine, they know what broke, and they need to be walked to the right part.

Part number search must tolerate spaces, dashes, case and superseded numbers
Model search must accept the model plate wording, not your internal naming
Both paths must end on the same product page, with fitment confirmed on screen

What fitment search is made of

Fitment is a data problem dressed up as a search problem. The interface is straightforward once the relationships underneath it are correct, and hopeless when they are not.

A clean machine model hierarchy: brand, family, model, serial range
A parts-to-model mapping that survives brand-specific naming
Supersession chains resolved so old numbers land on current parts
Diagram groupings so customers can navigate visually
A search index tuned for exact-match part numbers and fuzzy model names
Fitment confirmation displayed on the product page, not buried in a spec table

Why multi-brand makes it harder

One brand is a mapping exercise. Five brands is a normalisation exercise, because every manufacturer describes machines and parts in its own vocabulary, at its own level of detail, with its own idea of what a serial break means.

For Scholten's Equipment, Kubota, CLAAS and JCB data all arrived differently. Each was connected on its own terms, then normalised into a single model hierarchy so a customer could search once and see everything the dealership could sell them. Over 3.5 million parts ended up behind one search box.

What good fitment search does to the business

The direct effect is fewer wrong-fit returns, which is where parts ecommerce margin usually goes to die. The indirect effect is bigger: customers stop calling the counter to ask whether a part fits.

Scholten's Equipment recovered roughly 200 hours of counter time that had been spent answering exactly that question. That time did not disappear, it went back into selling.

Search paths and what each one needs

Customer path	What they start with	What the store must provide
Part number	An exact or superseded number	Exact match tolerant of formatting, supersession resolution
Machine model	A model plate or serial number	Model hierarchy, serial ranges, compatible parts list
Exploded diagram	A visual idea of the part	Diagram groupings with clickable callouts to parts
Category browse	A maintenance job, such as a service interval	Filters, fluids and kits grouped by job, not by manufacturer taxonomy
Reorder	A previous purchase	Customer accounts with order history and one-click reorder

Questions dealers ask

Parts lookup by machine lets a customer choose their equipment, by brand, model and often serial range, and see only the parts that fit it. It is the opposite of a keyword search, and it is how most customers actually shop, because they know what broke but not what the part is called.

Because a customer who cannot confirm a part fits will not buy it. They either call your counter, which removes the point of the store, or they buy from a marketplace that does show fitment. Returns from wrong-fit orders then eat whatever margin the store made.

Both, and a parts store needs both paths. Fleet managers reordering filters have part numbers. An operator standing next to a broken machine has a model plate and a photo. Serving only one of those halves your addressable demand.

At minimum a machine model list, a mapping from models to parts, and resolved supersession chains. Diagram groupings and serial-range breaks make it considerably better, because they let a customer navigate visually rather than guess at part names.

Yes, but only after normalisation. Each brand describes machines and parts differently, so the model hierarchy and part relationships have to be mapped into one shape before a single search box can answer for all of them.

They are not required, but they convert. Diagram searches are high intent, and dealers who publish exploded views tend to capture customers who would otherwise call a competitor to describe the part over the phone.

The supersession chain is resolved before the catalog is published, so an old number and a current number both resolve to the current part, with a note explaining the replacement. This prevents the most common cause of returned parts orders.

It is usually the longest single piece of a parts store project, because the mapping work scales with brand count rather than with page count. Scholten's Equipment covered five brands and over 3.5 million parts in a roughly four-month build.