Bring the search that failed before asking for more filters
A request for a feature is often a proposed explanation of a problem. “We need filters” may be correct, but it can also conceal a smaller difficulty: customers and the catalogue use different words for the same approved subject. The first conversation should expose that distinction before a new interface is priced.
The fictional example here concerns a supplier's search terminology. It reaches a bounded decision about one editorial synonym. It does not establish that superficially similar applications or products are technically equivalent.
Begin with the visitor's attempt
The supplier receives a request to add more catalogue filters because customers cannot find a particular family. Rather than listing possible filter fields, the project team asks for safe examples of the searches that led to follow-up calls.
Three anonymised enquiries contain the phrase “door stop.” The catalogue calls the relevant family “travel limiters.” The initial search checks product names and approved editorial synonyms, but no synonym connects that phrase to this family. Searching the literal reference TL-8 finds the intended product immediately.
This is useful evidence, but it is not yet enough to create a general relationship. “Door stop” can mean different products in different settings. The commercial search phrase needs to be examined with the person who understands the actual enquiries and the range.
Ask the product owner to resolve the meaning
The owner reviews the three examples and the pages the visitors were trying to find. For this narrow catalogue family, the enquiries concern the function described in the approved TL-series information. The owner agrees that the customer phrase can be an editorial search aid for that family.
The approval does not extend to every product called a stop or limiter. A separate application involving a floor-mounted stop belongs to another family and must not be silently merged. Nor does the synonym assert that TL-8 suits an installation without the normal technical assessment.
The decision record therefore names the phrase, the intended family, the supporting examples and the owner who approved the relationship. That is more precise than adding a broad list of related words generated from a dictionary or an AI suggestion.
Inspect what the existing search can do
In the imagined system, the current search supports an editor-maintained synonym field associated with a catalogue family. The project team verifies that the field can be changed through the ordinary editing process and that literal part-code matching remains a separate supported search route.
This finding changes the proposed scope. The first intervention can use an existing capability rather than introduce faceted navigation, new attribute data and another interface to maintain. Filters might still help a different comparison task, but they are not needed to answer the evidence currently in front of the team.
The distinction matters for review effort. A filter system would require reliable values across the range and decisions about combinations, empty results and mobile use. A single approved synonym requires a narrower editorial change and its own focused checks.
Complete the small decision
The fictional reviewer records these outcomes after the supported edit:
| Search task | Result |
|---|---|
| Enter the approved phrase “door stop” | The intended TL family is discoverable with its proper product names and references |
| Enter the literal code TL-8 | The same exact-reference result remains available |
| Search the separate floor-mounted application | Its distinct family remains discoverable; it is not relabelled as TL |
| Open the result | The approved product description, limitations and enquiry route are unchanged |
The owner accepts the narrow search improvement. The request for new filters is deferred until a separate task demonstrates their need. This is a completed scope decision, not a claim that the whole catalogue's search quality has been assessed.
Bring enough context to make that decision possible
For your first conversation, a useful search example includes the entered phrase, the intended result and why the person expected that result. Add the route they used and whether a product code was available. If staff supplied the answer later, explain the reasoning without forwarding private correspondence unnecessarily.
Include a contrasting example. A superficially similar phrase that should lead elsewhere helps prevent a narrow improvement from becoming an overbroad equivalence. The contrast also makes the product owner's decision easier to inspect.
Separate what you observed from what you propose. “These three visitors used this phrase and found no result” is evidence. “We need six filters” is an option. Both can belong in the conversation, but treating them as different things keeps the investigation open to the most useful intervention.
Maintain the editorial relationship
The handover identifies where the synonym is stored, who may approve changes and how to repeat the literal-code and contrasting-application checks. If the product family changes, the synonym needs review with the rest of its public information.
The catalogue service covers the wider relationships between products and documents. The project table helps collect the task and its exception. Send one example when you want to decide whether the first improvement belongs in terminology, navigation or a different part of the process.