Bring the citation before asking for better search
“We need better search” can describe several problems. A reader may know a subject but not a title. Another may have an exact citation yet be unable to distinguish an individual issue from a compilation. Those tasks need different evidence before a useful first scope can be chosen.
This fictional first conversation concerns a small publication series. It shows how two supplied references change the proposed work from a broad search feature to a specific reading route.
The request as it arrives
An organisation publishes a series called Community Learning Notes. Its website lists several cards with that same title. The team asks for an intelligent search box because readers keep contacting the office for help.
The initial discussion asks for the references that two readers actually supplied. One reader cites “Community Learning Notes, issue 4, 2025.” The other asks for the “2025 collected edition.” Neither is making a broad subject enquiry. Each is identifying a particular publication object.
The editor brings approved cover pages and the current records:
| Record | Material | Distinguishing evidence |
|---|---|---|
| P-27-4 | Issue 4, published in 2025 | Cover identifies issue 4; includes its own article sequence |
| P-27-C | 2025 collected edition | Cover identifies collected edition; combines issues 1–4 with a new introduction |
| P-27-5 | Planned issue 5 | No approved public file yet |
The website currently reduces the first two records to the repeated series title and a generic download label. The files are different and correctly stored. The cards hide the distinction the readers already know.
Ask a narrower question
The team changes the first question from “How can search understand our publications?” to “Can a reader with either of these references identify the intended item?”
That question can be answered using the existing collection and approved source material. It does not require a new retrieval service or an assumption about how every future reader will search. It also prevents the planned fifth issue from being invented as a public resource simply to complete a sequence.
The editor confirms that the collected edition is a separately issued work. It is related to the series and contains earlier material, but it is not the same file as issue 4. Its new introduction belongs to that compilation. This is the editorial fact the interface must preserve.
Choose the first delivery
The team selects a series page with explicitly named issue records and a separate collected-edition entry. The card titles become “Community Learning Notes — issue 4 (2025)” and “Community Learning Notes — collected edition 2025.” Each card describes its own scope and leads to its own record.
The series page explains that the collected edition contains issues 1–4 and an additional introduction. It does not redirect the issue 4 citation to the compilation. A reader may prefer the compilation, but that is a choice to make with the relationship visible.
Issue 5 stays outside the public download list because no approved public file exists. The internal preparation record remains available to its authorised editor. The first scope does not promise to complete missing publication work.
Check the two tasks separately
In the fictional review, a reader receives the issue 4 citation and starts at the series page. They select the issue 4 record and open the file whose cover identifies that issue. A second reader receives the collected-edition reference and finds the compilation with its new introduction.
The reviewer then asks what is available for issue 5. The public page does not offer a nonexistent file or label another item as the fifth issue. The editor can explain its preparation status through the organisation's chosen enquiry route.
These results support the selected first delivery. They do not establish that subject search is unnecessary for the whole collection. A later reader study may reveal a separate need to search themes, contributors or passages.
Turn the evidence into a useful enquiry
For your own project, bring the reader's words, the reference they used and the material you expect them to find. A safe reconstruction is acceptable if the original request contains personal information. Keep the distinguishing publication facts intact.
Record the proposed task, the current obstacle and the smallest material set that demonstrates it. Identify who can confirm the publication relationships and what is unavailable. That makes a first conversation productive without requiring you to choose a technology in advance.
The project-planning page helps assemble the wider context. If the same difficulty affects a larger collection, discuss the publication-library service.