A customer opens your help article, follows every step, and still gets stuck. The article exists. It just describes a screen your team redesigned two releases ago.
Knowledge gap analysis is how I catch that problem before the ticket arrives. It compares what people need to know to finish a task with the knowledge they can actually use today. Then it shows exactly where the two diverge.
TL;DR
- Knowledge gap analysis compares needed knowledge with usable knowledge, not just what has been written down.
- Gaps can be missing, outdated, hard to find, or held by one person.
- The best evidence comes from real tasks: searches, support questions, and watching someone try.
- Match each gap to a fitting fix. Another article is often the wrong answer.
- In fast-shipping teams, gaps reopen after releases, so analysis has to be continuous.
What is knowledge gap analysis?
Knowledge gap analysis is the process of comparing what people need to know to complete a task with what they can currently find, understand, or do.
The difference between needed and usable knowledge
The gap sits between those two states. A customer who needs to export invoices is still blocked if the export guide is buried under a vague title like “Data options.” Search never surfaces it.
You can run the analysis on customer-facing documentation or on knowledge held inside a team. A page can exist and still leave a gap. If your SSO setup article lists a field your product renamed, the knowledge is present but unusable.
How it differs from related assessments
These terms get used interchangeably. They answer different questions:
| Analysis | Core question | Typical finding |
|---|---|---|
| Knowledge gap analysis | Is the needed knowledge available and usable? | The billing guide shows an old path, so customers fail the task |
| Documentation audit | What is in our content, and what condition is it in? | 40 articles have outdated screenshots; this is evidence for a gap analysis |
| Skills gap analysis | Can people do what their roles require? | Support agents can’t yet debug API authentication errors |
| Needs assessment | What support or intervention is needed overall? | Onboarding needs an in-app checklist, not another document |
Why does knowledge gap analysis matter?
What a fast-shipping SaaS team gains
When you ship weekly, every release can quietly break an answer. Running this analysis regularly turns that hidden decay into a fixable list:
- Repeat tickets prevented: fixing the answer stops the same question from returning after every release.
- Support time saved: agents stop rewriting the same explanation by hand.
- Less dependence on single experts: procedures no longer stall when one person is away.
- Fewer customers acting on stale steps: you avoid failed setups and lost trust.
What counts as a knowledge gap?
A knowledge gap is any point where needed knowledge fails to reach the person who needs it, and missing content is only one version.
Missing, outdated, inaccessible, and unshared knowledge
- Missing: no instructions exist for a task, such as connecting a new integration.
- Outdated or inaccurate: an article describes an old interface or a wrong procedure.
- Inaccessible: the right answer exists, but its title, location, or wording hides it. Customers search “cancel plan” while the article is called “Subscription management.”
- Unshared: one colleague knows a procedure, and nobody else can perform it.
What these gaps look like in practice
Picture a release that moves notification settings from the profile menu into a new Workspace tab. The help article still shows the old steps and the old screenshot. Customers follow it exactly and land on a page that no longer has the toggle.
Now picture a team where only one engineer knows how to recover a failed data import. When that engineer is on vacation, support can’t answer the question reliably. Both are illustrative scenarios, yet they will feel familiar to most SaaS teams.
Why a missing article is not always the answer
Writing another page won’t fix an incorrect existing page. It won’t fix poor search wording either, and it won’t teach a teammate a procedure. Often a duplicate article makes things worse, because now two answers disagree.
What evidence reveals a knowledge gap?
The strongest signals come from what people tried to do and where they stopped.
Searches, article usage, and feedback
- Searches with no useful result, including failed AI answers, point to gaps. Check whether the information is truly missing or just named differently.
- Questions that continue after an article view suggest the article didn’t help. A click alone proves nothing.
- Article feedback pinpoints confusing steps or answers people no longer trust.
Support questions and shipped changes
- Recurring support questions, grouped by the task customers attempted, expose answers that fail in practice.
- GitHub pull requests and shipped Linear work reveal changed behavior that current articles don’t reflect.
- Release notes, checked against screenshots, steps, and terminology, catch pages you assumed were current.
What people can actually do with the knowledge
Run a simple user acceptance test. Ask someone to complete a task using only the help center, and watch where they hesitate or give up.
For internal knowledge, ask a teammate to explain or perform the procedure. This separates information that is unavailable from information they haven’t learned to apply.
How does a knowledge gap analysis work?
Herrmann outlines a four-part gap analysis model: define goals, assess current state, identify gaps, and plan how to close them. I apply that model here to knowledge and help centers.
Step 1: Define what people need to know
Start narrow. Pick one audience and one task, such as an account admin changing a billing setting after a release. Then write down what a correct answer must let them accomplish.
- Audience and specific task identified
- Definition of a complete, correct answer written
Step 2: Establish what is available now
Assemble the current answer for that one task. Use the signals from the evidence section above rather than gathering everything at once.
- Current answer located, including where it appears and whether it matches today’s product
- Related searches, support questions, and task-test observations pulled together
- For team knowledge, what people can demonstrate, not just what a document claims
Step 3: Identify and weigh the differences
Compare the required answer with the current one. Classify the gap as missing, inaccurate, hard to find, or not yet understood.
Then weigh it. Consider how badly it blocks the task, how often the question arises, and what a wrong answer costs. A failed search is a clue, not proof that you need a new page.
Step 4: Match the gap to a remedy and a check
Pair every finding with a response and a way to confirm people can now complete the task. A fix you haven’t verified is only a guess.
- Documentation example: correct an outdated procedure, then test each step against the current product.
- Team example: share and practice an undocumented procedure, then check whether another teammate can perform it.
A worked example: a help article after a product release
Here is a hypothetical: a merged pull request moves the “Change billing email” setting from Account into a new Billing page.
Compare the needed answer with the existing one
| Stage | Evidence in the example | Finding |
|---|---|---|
| Define the need | Customers must update their billing email after the release | They need the current path through the interface |
| Assess the current state | The existing article ranks first in search but shows the old path and screenshot; tickets say “I can’t find this setting” | The answer is findable but wrong |
| Identify the gap | No missing content; the steps no longer match the product | Primary gap is inaccurate knowledge, not an absent article |
| Choose the response | A new article would duplicate the old one | Update the existing instructions and screenshot, then verify them |
What the example teaches us
Counting articles would have shown full coverage. Search data alone would have shown success, because people found the page. Only comparing the customer’s task with the usable answer exposed the real problem.
How do you close gaps without creating new ones?
Closing a gap well means fixing its actual cause and keeping it fixed.
Choose the response that fits the gap
- Missing answer: create instructions for that specific task.
- Incorrect answer: revise the steps and screenshots, then verify them.
- Hard-to-find answer: improve the title, organization, or search terminology.
- Unshared team knowledge: document the procedure and confirm others can apply it.
Recheck fast-changing knowledge after releases
An audit captures one point in time. Teams that ship frequently need to check whether each relevant change makes an answer stale, because last month’s accurate article can break this week.
Ferndesk is one example of how this fits into the workflow. Its agent, Fern, monitors GitHub, Linear, and support conversations for potential gaps. It then drafts article or screenshot updates for a human to review. The point is accurate answers, with automation handling the watching.
Who owns the analysis and how often to run it
Name one owner, usually the support or docs lead, so the work doesn’t fall between teams. Support surfaces recurring questions, product flags shipped changes, and docs updates and verifies answers. Run release-triggered checks, plus a broader review each quarter.
How can you tell whether the gap is closed?
Publishing a fix is not the finish line. Here’s how to check it worked.
Look for successful tasks, not just more content
- Task completion: users can complete the task with the revised answer. Retest after the next release too.
- Search and feedback: people find the answer and rate it helpful. One good week is not a trend.
- Recurring questions: tickets about that task drop compared with before. Other factors also move ticket volume, so don’t credit one article for every change.
Know when the answer needs another look
Return to the analysis whenever the product, procedure, or audience changes. A gap is closed only while the knowledge stays correct, findable, and usable.
Conclusion
Knowledge gap analysis compares the knowledge people need with what they can actually use. The most useful gaps show up in evidence from real tasks, and the right fix depends on the cause. Verify the answer works, then keep checking as your product changes.
So, for your most-asked support question this month: can a customer complete that task today using only your help center?
FAQs: knowledge gap analysis
What is an example of a knowledge gap?
A help article that shows an old settings path after a redesign is a common example. The content exists, but customers can’t complete the task with it.
How is a knowledge gap analysis different from a documentation audit?
An audit inspects content condition. A knowledge gap analysis asks whether people can complete tasks, and it uses audit findings as one source of evidence.
How often should SaaS teams run a knowledge gap analysis?
Check affected articles after each relevant release. Then run a broader review roughly once a quarter.
Who should own knowledge gap analysis?
One accountable owner, typically the support or docs lead, with input from support and product teams.
Book a demo