When choosing the product is the hard part

In a great many technical businesses the obstacle is not persuading someone to buy — it is that they cannot work out which of your four hundred products they need. Presenting all four hundred, however well designed the page, does not solve that. It transfers the problem to a customer who is less equipped to solve it than you are.

The businesses that get this right stop building catalogues and start building selectors.

What makes a product hard to choose?

One of three things, and often all three.

The choice depends on the application, not the product. The right adhesive depends on substrate, tile size, location and conditions. The right pump depends on flow, head and fluid. The right cable depends on load, distance and installation method. None of that is visible in a product list — the customer has to already know the answer to find it.

Compatibility is a rule. A fitting must match the pipe’s diameter, class and material. A breaker must suit the board. A seat must fit the pan. Present these as independent products and customers will assemble combinations that cannot be installed.

The range multiplies. Material by size by rating by variant produces thousands of items that look nearly identical in a list and differ in one number that matters.

Why does filtering not fix it?

Because filtering assumes the customer knows what to filter by.

A filter panel offering material, diameter, pressure class and connection type is useful to a specifier and useless to everyone else. The person who knows they need a 40mm class 6 CPVC elbow does not need help. The person who knows they are connecting two pipes under a sink does.

	Filtering	A selector
Asks about	Product attributes	The customer’s situation
Requires	Technical knowledge	Knowledge of their own job
Returns	A narrowed list	A recommendation
Suits	Specifiers and repeat buyers	Everyone else
When it fails	Customer filters wrongly and leaves	Rarely — the questions are answerable

Build both. Filtering for people who know, a selector for people who do not. Most technical catalogues offer only the first and lose the second group entirely — and the second group is usually larger.

What does a good selector ask?

Questions the customer can answer without looking anything up.

Not “what pressure class do you need” — that is the answer, not the question. Instead: what are you connecting, where is it, is it under pressure, is it exposed. Four questions about their situation, from which the specification is derived.

Three things separate one that works from one that irritates.

• Few questions. Four or five. Beyond that people abandon, and each additional question usually narrows less than the one before.
• An answer, not a list. “This product, and here is why” beats twelve results ranked by relevance.
• A route to a person. Some cases are genuinely unusual. The selector should recognise when it is out of its depth and hand over rather than guessing.

That last point is the same principle that governs a good chatbot: the valuable behaviour is knowing when not to answer. We have written about it in building a chatbot that answers from your own documents.

Why does this matter commercially?

Because the customer who cannot decide does not buy a different product from you. They ask a dealer, who recommends whatever they know best — which may not be yours.

Every unresolved choice is a decision handed to somebody else. In channel businesses that somebody has their own preferences, their own stock position and their own margins.

There is a second effect that shows up in returns. A customer who chose from a list without understanding the application chose wrongly some of the time, and the wrong product comes back — or worse, gets installed and fails months later, which becomes a warranty claim and a complaint about your product rather than about their choice.

What does this require underneath?

Product data held as structured attributes, which is where most of these projects actually stall.

A selector cannot reason about products whose specifications live inside a description field or a PDF. Material, rating, size, application and compatibility have to be fields — queryable, consistent, and maintained by someone.

Most businesses discover this at the point of building the selector and find their catalogue is a price list with photographs. That is not a reason to abandon the project; it is the first phase of it, and it makes everything afterwards possible — see what to digitise first.

Common questions

Is this the same as a product finder or configurator?

Related but not identical. A selector narrows to an existing product from your range. A configurator builds a specification for something made to order. The underlying data requirement is similar; the output is not.

Should dealers have access to it too?

It is often more valuable to them than to end customers. A counter salesperson answering a technical question in seconds, correctly, sells more and returns less — and it means your product gets recommended rather than whichever one they remember.

What if our products genuinely need an engineer to specify?

Then the selector’s job is to narrow and route, not to decide. Getting the enquiry to the right technical person with the application details already captured is still a large improvement on a contact form.

How do we handle products where the answer depends on a calculation?

Do the calculation. Coverage, sizing, load and quantity are deterministic — they should be computed rather than approximated, and the result is usually the most useful thing on the page.

More on this: selling online when nobody pays online and catalogues with thousands of SKUs. See also product catalogues & dealer locators.