A help center with excellent, accurate articles can still fail its customers if they can't find the right one. Structure, categories, and naming decide whether a correct answer is two clicks away or buried under a wrong guess.
The core principle: structure around customer goals, not your internal product architecture.
Direct answer
Structure a SaaS help center around what customers are trying to accomplish ("Getting started," "Managing your team," "Billing") rather than your internal feature or system names. Keep categories to 5-8 top-level groups, name articles as the question a customer would type, and put a search bar above the fold since most returning visitors search rather than browse. The structure should mirror the customer's mental model of your product, not your engineering team's.
Category design that works
- 5-8 top-level categories is the sweet spot; more than that and customers can't scan them at a glance.
- Name categories after goals ("Getting started," "Managing billing," "Integrations"), not internal system names ("Account provisioning," "Subscription service").
- A universal starting point: Getting Started, Account & Billing, Core Features (broken into subcategories per major feature), Integrations, Troubleshooting.
- Put the highest-traffic category first, not the one your team considers most important internally.
Article naming and search
Title each article as close to the customer's actual search phrase as possible: "How to invite a teammate" beats "Workspace membership management" even though they describe the same feature. This isn't just about clarity, it directly affects which articles rank in Google and get retrieved by an AI answer widget.
Frequently asked questions
How many categories should a help center have?
5 to 8 top-level categories is a practical range. Fewer than that and categories get too broad to be useful; more than that and customers can't scan the list to find the right one.
Should help center articles be organized by feature or by customer goal?
By customer goal. Customers search for what they're trying to accomplish, not your internal feature or system names, and articles titled that way also perform better in search and AI-answer retrieval.
Is search or navigation more important in a help center?
Search, for returning or specific-question visitors. Navigation matters more for new customers browsing to understand what's available. A help center needs both, but weak search is the more common and more costly gap.
Conclusion
- Structure a help center around customer goals, not internal product architecture.
- Keep top-level categories to 5-8, and name articles the way customers would search for them.
- Search quality matters as much as category design for returning visitors.
- DocsKoala keeps that structure accurate automatically as your product's features change.