Article feedback

Every public article invites feedback. The helpful / not helpful widget and view counters create the feedback loop that tells you which articles work — and which need help.

What readers see

At the end of each article: "Was this article helpful?" with thumbs-up / thumbs-down actions. One vote per reader; the widget acknowledges the vote inline.

What's tracked

• Views — counted per article page load.

• Helpful count — positive votes.

• Not-helpful count — negative votes.

Counts are shown in Docs for authors, so writers see performance without exporting anything.

Where the signal flows

• Authoring — a high-traffic article with a poor helpful ratio is the top candidate for revision.

• Support operations — article feedback complements Coverage Gaps (see Support Knowledge Sources & Coverage): coverage gaps show where the AI couldn't answer; feedback shows where published content fails humans.

• Content strategy — sort top-viewed articles by helpful ratio to find both champions (link them from related articles) and fixes.

How votes and views are recorded

Readers don't need an account to vote, and feedback submissions are rate-limited to prevent abuse. Views are counted when the article page loads.

Tips

• Ask for the article link in follow-up surveys; the article view + vote history is the ground truth.

• After fixing an article, watch the helpful ratio trend over the next weeks — that's your A/B signal.