Six components make up the practice of IA. Taxonomy classifies content into categories. Metadata describes attributes of individual content objects. Labeling gives categories, links, and pages clear names. Organizational structure determines the overall shape of the information space. Navigation systems let people move through that structure. Search systems let people find content directly, without browsing.
Each component is covered in depth later in this course. For now, the important idea is that they work together as a system, a weakness in one component often surfaces as a symptom somewhere else, which makes diagnosing real IA problems harder than it first appears.
A missing metadata field rarely looks like a metadata problem to the user. It looks like "I searched and nothing came up" or "I found the file but had no idea if it was current." Diagnosing IA problems means tracing the symptom back to its actual source.
1: Labeling. The page exists and the search system found it correctly by title, the problem is that the title, "Business Registration Continuation Form," uses none of the language a real user would type. This is a mismatch between organizational terminology and user vocabulary, which is a labeling failure, not a search engine failure.
2: Taxonomy / organizational structure. Forty links with no subgrouping means no one ever broke "Business Services" down into meaningful subcategories. The category exists, but it isn't actually organized: it's just a bucket.
3: Metadata. A file with no description and no visible date is missing descriptive and administrative metadata. The label "Form 27B" alone tells the user nothing about what it is or whether they can trust it.
There's no single right answer on priority, but the strongest case is fixing the labeling problem first. It's the very first point of contact, most users search before they browse, so a labeling fix that lets "renew license" find the right page removes the failure before the user ever reaches the taxonomy or metadata problems. Fixing taxonomy or metadata alone wouldn't help someone who never gets past the search box.That said, a reasonable case exists for taxonomy too, if a meaningful number of users start by browsing rather than searching, cleaning up the forty-link dump would have equally broad impact. The right answer depends on data you don't have yet in this scenario: how users are actually arriving at this problem. That's exactly the kind of question the research methods later in this course are designed to answer.