[object Object]

Most HubSpot knowledge bases are vanity. They look modern, have 80 articles, and deflect maybe 4% of tickets. The reason isn’t content quality, it’s structure and search relevance. Here’s how to fix both.

Map articles to ticket reasons, not product features

Open your top 50 ticket categories from the last 90 days. Every category needs exactly one article. If a category lacks coverage, that’s the next article you write. Skip the “Welcome to our product” filler.

Title every article as a question

“How do I reset my password?” outperforms “Password reset” by ~40% in HubSpot’s site search. The KB search engine is intent-matched; questions match the way customers actually type.

Front-load the answer in the first 50 words

The Breeze Customer Agent and the chat widget both pull from the article preview. If your answer starts in paragraph three, it never makes it into the bot reply. Lead with the resolution, then add context.

Tag for the chat widget, not for humans

Use the article’s kb_internal_tag property to mark articles as chat_eligible. Then configure your chat flow to query only those articles:

Filter: kb_internal_tag CONTAINS chat_eligible
AND last_updated > 180 days ago

Stale articles in chat replies destroy trust faster than no reply at all.

Wire ticket-to-article feedback

Every closed ticket should prompt the agent to tag which KB article would have prevented it (or “none”). Build a workflow that increments a custom property kb_gap_count per ticket reason. The top three gaps each month become next month’s articles.

Track deflection, not pageviews

Pageviews are noise. The metric is: (KB sessions that did NOT create a ticket within 24 hours) / (KB sessions). Pipe HubSpot’s KB analytics into a custom report. A healthy deflection rate sits at 25-40%.

Kill the long tail

Articles with under 10 views per quarter and zero “helpful” votes get unpublished. They drag down search relevance for everything else.

What to do this week

Pull your top 10 ticket reasons. Audit whether each has one focused, question-titled, answer-first article. Fix the gaps before adding new content.

[object Object]
Share