Term base best practices: How to build a living terminology system

What is a term base?

A term base is a glossary of approved terms and their translations that teams use to keep terminology consistent across languages.

A good term base captures what a term means in your product or domain, plus the rules for how it should (and shouldn’t) be translated. It often includes context, part of speech, definitions, usage notes, and do-not-use variants.

You can think of a term base as a living data model. Each entry represents a concept (like a feature name, product capability, plan tier, or compliance phrase), plus the details people need to use it correctly. That includes:

• Definitions
• Context
• Approved and forbidden variants

This is how translators, writers, and reviewers don’t have to play guesswork.
 

What a term base typically contains

A strong term base entry usually includes:

• Concept / term (the canonical term you want used)
• Definition (what it means in your product or brand)
• Context (where it appears, who it’s for, and any constraints)
• Approved variants (if more than one is acceptable, and when)
• Forbidden variants (what not to use, for different brand, legal, or SEO reasons)
• Notes for translators/reviewers (edge cases, tone, UI limits, examples)

To make this concrete, here’s a sample term base entry for a pricing tier, where consistency matters across marketing, product, and legal docs. We’ll use a learning platform as an example software product, and we’ll imagine this SaaS company offers multiple plan tiers across regions and languages.

As you can see, it’s important to define the concept, lock down where the wording must be exact (pricing, checkout, renewals), and clarify when shorthand is acceptable. This makes “wrong-but-plausible” alternatives hard to miss, which is exactly what you want.

Glossary vs term base: what’s the difference?

A glossary is a reference list. It helps people look up what a term means and what it’s usually called. A term base is a system. It defines terminology as structured, reusable knowledge, and makes it harder for teams to accidentally drift away from it as they write, translate, and ship content.

In other words:

• A glossary helps you remember what the term means
• A term base helps you stay consistent when using the term across different languages

Top 5 term base best practices that hold up at scale

A term base only becomes a strategic asset when it’s designed for real-world usage. Realistically, you’re going to have a lot of contributors, different content types, and many edge cases.

These best practices are what separate a term base people trust from one people just ignore.

1. Design it like a data model, not a word list

Start with the idea you’re trying to protect. It’s the concept you’re conveying. Then define how that concept should show up in language.

At minimum, each entry should answer:

• What is this concept? (definition)
• Where does it appear? (context)
• What’s allowed? (approved variants)
• What’s not allowed? (forbidden variants)
• What should translators/reviewers watch for? (notes + examples)

This structure matters because it prevents “correct but still wrong” usage. Two terms can be linguistically valid and still be brand-incorrect if they refer to different concepts in your product.

2. Make ownership and review unavoidable

Term bases decay when everyone can add terms but no one is responsible for quality. The fix lies in clearer governance.

You need to define:

• Who can create entries (often product marketing, product, or localization owners)
• Who approves (one accountable owner per domain: product tiers, features, legal, SEO)
• How changes get reviewed (lightweight workflow + change log)

Even simple rules help. For example, you can say “New terms must include definition + context,” and “Plan tier terminology can’t be changed without approval.” Write things down and communicate across the team.

3. Add context that prevents mistakes

Context is what makes a term base usable when the same word could mean different things depending on where it appears. For each entry, capture the details that stop confusion:

• Where it shows up (UI label vs help center vs sales deck)
• Audience and intent (admin vs learner, marketing vs legal)
• Constraints (character limits, capitalization, punctuation, tone)
• Examples (one correct sentence, one incorrect-but-plausible sentence)

This is also what makes term bases work better with AI-assisted drafting, because it narrows the “reasonable” options.

4. Treat forbidden terms as first-class citizens

Forbidden terms aren’t just “things we don’t like.” They’re often the most valuable part of the term base, because they prevent costly mistakes. Common reasons a term is forbidden:

• Brand clarity: avoids confusing near-synonyms (track vs learning path)
• Sales accuracy: prevents tier mislabeling (Enterprise vs Business)
• Legal/compliance: avoids claims you can’t make (guaranteed, perfect)
• SEO strategy: reduces legacy naming and keyword noise

Write forbidden-term rules clearly, with the “why” included. People follow rules faster when they understand the cost of breaking them.

5. Keep it small, then expand based on impact

A term base doesn’t need to cover every single noun in your product. But it does need to cover the terms that create the most risk and rework.

A practical starting set is:

• Plan names and tier descriptors
• Top feature names
• UI navigation labels
• Compliance and legal phrases
• High-traffic SEO terms (and their forbidden variants)

Once those are stable, you can expand into deeper areas (industry terms, campaign language, partner terminology) without overwhelming contributors.

Terminology management automation with Lokalise

A term base becomes genuinely useful when it keeps up with reality. If you ship new features, or your pricing plan names evolve, or marketing campaigns introduce new language, it needs to keep up.
If terminology updates depend on someone remembering to edit a spreadsheet, the system will fall behind.

In Lokalise, the goal is to make terminology management feel less like manual upkeep and more like an operational workflow. Terms get added easily, and they are reviewed, approved, and enforced with minimal friction.

Automate term intake

Most teams already have a steady stream of “new terms,” they just don’t capture them systematically:

• a new feature name from a release
• a renamed setting in the UI
• a campaign tagline that becomes semi-permanent
• a compliance phrase legal wants locked down
• a keyword decision that impacts SEO across locales

With a modern term base workflow, those terms get added as candidates with enough context to be reviewed once, then reused everywhere.
 

Automate review and approval

Terminology is rarely “one person’s call.” Product, marketing, support, and localization all have a stake. This is especially the case with plan tiers, feature names, and SEO-sensitive phrases.

Instead of resolving terminology through scattered comments and one-off decisions, treat approval like a lightweight gate:

• candidate (new term proposed)
• reviewed (definition and context validated)
• approved (canonical + variants locked in)
• deprecated (legacy terms kept for reference and flagged as forbidden)

When the decision is recorded once, you don’t have to re-litigate it in every translation job.

Automate distribution so everyone uses the same source of truth

A term base fails when different teams maintain their own versions: product has one list, marketing has another, and translators work from a third. Even if each list is “mostly right,” the inconsistencies accumulate.

With Lokalise, the term base becomes shared infrastructure for terminology:

• Translators and reviewers can rely on the same definitions and rules
• Terminology decisions stay consistent across projects and languages
• Updates don’t require re-sending files or chasing “latest versions”

This is what makes the term base a strategic asset. It stays current, it stays enforceable, and it scales with the amount of content your organization produces (especially when AI increases the volume of drafts).
 

Your term base needs to be dynamic

A static glossary within a spreadsheet can tell you what the preferred term is. But translation management systems like Lokalise can help you apply that term consistently.

Terminology becomes part of the workflow, so people don’t have to invest an effort to check an external doc. It’s right there.

Translators and reviewers can see the approved term while they’re working, rather than guessing or switching tabs, or just hoping they used the right variant.

Want to see what this looks like in Lokalise? Give it a try for 14-days for free, no credit card required.