Refining a search that returns too much or too little

In this lesson

Diagnose whether a poor result set is a filter problem or a phrasing problem
Adjust one variable at a time to move toward a workable result set

Most search problems fall into one of two shapes: hundreds of results with no useful ordering, or almost none. Both have a small number of usual causes, and both are fixed faster by changing one thing at a time than by rebuilding the search from scratch.

Too many results

A flood of results almost always means a skill term is too broad, too common, or matched by the ontology into an unhelpfully wide neighbourhood. “Project management” alone will return most of the organisation; “Excel” returns nearly everyone. The fix is to add a second, more specific skill that narrows the field, rather than trying to make the broad term stricter.

Too few results

Almost nobody, or nobody, coming back is nearly always a filter being stricter than the work needs, not a genuine shortage of qualified people. Work through these in order:

Widen or remove the location filter first - it is the most common cause of an empty result set that should not be empty.
Check the availability window against the work’s actual start date. A window set to “immediately” when the work starts in six weeks excludes people who will be free by then.
Reconsider whether a skill phrase is too literal. Try a near-synonym, or remove one of several skill terms to see whether it was the one over-constraining the set.
If all of the above still returns too few, the organisation may genuinely be thin in that combination - worth confirming with your resourcing lead before concluding the search is broken.

Change one thing at a time

The temptation when a search is not working is to adjust several filters at once and re-run. Resist it. If you loosen location and broaden a skill term in the same step, and the result set improves, you will not know which change did the work, and you will repeat the same mistake on the next search. One change, one re-run, and you build a working sense of which filters are sensitive for which kinds of work.

Refining a search well is what turns “the system isn’t finding anyone” into a shortlist you can act on, usually within two or three iterations.

Key takeaways

Too many results usually means a skill term is too broad or too common
Too few usually means location or availability is filtering harder than intended
Change one filter at a time so you know what actually fixed it