The concept rank in keyword expansion isn't configurable, and here's why.

If you’ve ever poked around Relativity’s keyword expansion, you’ve probably spotted a lot of moving parts. One question that tends to pop up is about concept rank: can you tweak it, or is it fixed? Here’s the straightforward take, with a practical angle you can use in real work.

What is concept rank, anyway?

Think of keyword expansion as a way to connect the words you search with the ideas that sit behind them. When you type a term, the system maps it to related concepts—concepts could be topics, entities, or ideas that the taxonomy recognizes. Each concept gets a sense of importance or relevance tied to that query. That sense of importance is what we call the concept rank.

In most setups, that rank isn’t something you flip with a switch during a search. It’s defined by the system’s underlying structure—the taxonomy, the way concepts are linked to keywords, and how relevance scoring is designed. The goal is to surface the results that matter most, not to have every user juggling a dozen knobs mid-flight. So, in the common scenario, the concept rank is predefined.

Why the rank tends to stay fixed

Here’s the thing: if you allow users to change how concepts are ranked for every keyword expansion, you end up with a lot of variability. Results can swing wildly from one user to the next, or from one project to another, making it harder to trust the overall relevance signal. A stable, pre-established rank helps keep results consistent, which is valuable when teams are trying to compare findings across a large set of documents.

Another reason is governance. Taxonomies and concept mappings are often built with input from subject-matter experts over time. Those experts want to make sure that, across the board, the same concepts get the same weight when the same keywords show up. If the rank could be changed casually, governance becomes a moving target. The outcome? A lot of noise and a potential drift away from the terms that actually matter for the matter at hand.

Edge cases worth noting (without sidestepping the main point)

That said, there are a few scenarios where you’ll see knobs or settings that touch related areas, even if the core rank itself isn’t meant to be user-adjustable during a single keyword expansion:

• Taxonomy-driven adjustments: Some environments allow admins to refine how concepts in a taxonomy map to a given set of keywords. This doesn’t re-rank concepts on the fly, but it can influence which concepts show up in the first place and how strongly they’re connected to the query.

• Expansion scope controls: You can often define how broad or narrow an expansion should be. For example, you might limit expansions to a particular domain or language, or tie expansions to a controlled vocabulary. This shapes what concepts appear, not how the rank between those concepts is ordered.

• Document-type considerations: In some systems, the relevance of concepts can be affected by document type or metadata. The top-level ranking logic remains fixed, but the context in which it’s applied can shift how results feel to the user.

• Post-search ranking tweaks: After you get results, you might have options to sort, filter, or bias results based on metadata, document lineage, or other signals. These aren’t changes to the concept rank itself, but they influence what you see at the top of the list.

Putting it plainly: the default concept order stays steady, but you can influence what you see by shaping what gets expanded and how you slice the results.

What you can configure instead

If the goal is to steer outcomes without altering the foundational rank, here are practical levers you can pull:

• Refine synonyms and related terms: Feed the system a cleaner set of synonyms so expansions stay aligned with how your team actually talks about topics. This makes the expansion more precise without touching the rank.

• Tighten taxonomy and concept definitions: Regularly review and update how concepts are defined and linked to keywords. A well-tuned taxonomy keeps results relevant and reduces noise.

• Gate expansions by domain or language: Limit expansions to the domains you care about, or to the languages you operate in. This focuses the signal and avoids irrelevant noise from outside the scope.

• Use controlled vocabularies: When possible, anchor expansions to a curated list of terms that your team recognizes as significant. Consistency beats chaos here.

• Leverage post-processing filters: After results come back, apply filters by date ranges, custodians, project areas, or other meaningful dimensions. It’s a way to highlight what’s important without fiddling with the core ranking.

A mental model you can carry into daily work

Picture the concept rank as the spine of a well-curated library. The spine is reliable; it tells you where to look first. You don’t loosen the spine for every reader who visits; instead, you improve the section-by-section organization, the cataloging, and the cross-references. When you do those improvements, your searches become faster, your findings more consistent, and your team spends less time chasing the wrong threads.

That’s not to say you should ignore nuance. Sometimes, a topic has sub-areas that deserve different emphasis depending on your matter. In those moments, a combination of taxonomy tweaks and careful query design can help steer results toward the right concepts, even when the rank itself remains fixed.

Relativity-specific flavor — what this means in practice

For teams using Relativity or similar platforms, the practical takeaway is clear: you shouldn’t expect to flip the concept rank switch during a normal keyword expansion. The design is to provide a robust, stable relevance signal. When you’re building searches, you’ll want to invest in a solid taxonomy and a thoughtful expansion strategy. That means collaborating with subject-matter experts to ensure terms map cleanly to concepts, and that the expansion covers the cases your team actually encounters.

If you’re curious about the experience from a user perspective, you might notice how the system surfaces related concepts alongside the primary term. The goal is to help you discover connections you might not have anticipated—without letting the results drift away from what truly matters for your matter.

A quick checklist you can apply

• Do I have a clean, well-documented taxonomy for core concepts? If not, that’s a good place to start.

• Are synonyms aligned with how the team actually talks about topics in daily work? If gaps exist, consider a targeted update.

• Is expansion scoped to the relevant domain and language? If expansion bleeds into areas you don’t care about, trim it back.

• Can I apply filters after results return to spotlight the pieces that matter most? If not, look for a way to layer metadata-driven sorting or filtering.

• Do I have governance around how concepts are defined and linked to keywords? If governance is loose, establish a lightweight process to keep things stable.

Common questions (and straightforward answers)

• Is the concept rank configurable for individual users? No, in the typical setup the rank is predefined to keep results consistent across users and projects.

• Can document type affect how results are ranked? It can influence visibility and relevance signals, but the core rank isn’t something users adjust on the fly.

• What should I adjust if results feel off? Start with taxonomy and expansion term curation—clean up synonyms, tighten the scope, and use filters to focus the output.

Real-world touchpoints and analogies

• Imagine you’re searching a digital archive for old photos. The system knows which topics tend to cluster with your search terms (people, places, events). The weight assigned to those clusters helps decide what shows up first. You don’t reweight the clusters with every search; you improve the labels and connections so the cluster map better matches what you’re really after.

• Or think about a librarian trained to classify new books. The librarian’s system uses a fixed set of categories. When a new book lands, the librarian updates the catalog so future searches point to the same familiar shelves. The shelf labels don’t jump around every time a person looks for something; the labels stay put, guiding you to what you need.

Final takeaway

The statement True: “The concept rank is configurable in a keyword expansion” isn’t generally how these systems work. In most Relativity-like environments, the concept rank is built into the framework and kept stable to preserve consistent relevance signals. What you can influence—much more often than not—instead is how you map concepts, which terms you expand, and how you filter and sort results after they appear. A thoughtful approach to taxonomy, term management, and post-search refinement will deliver clearer, more actionable findings without wrestling with the ranking dial mid-quest.

If you’re charting a path through Relativity’s search landscape, this perspective can keep you grounded. Focus on building a rock-solid taxonomy, curate your term expansions with care, and use post-search controls to spotlight the nuggets that matter. It’s a steady, practical way to harness the power of keyword expansion without chasing shifting sands.

Bottom line: the rank stays where it is, but your understanding—and your workflow—can keep getting sharper. And in a field where precision matters, that clarity can make all the difference.