Reader feedback

Readers of a published article can rate it on a 1–5 star scale from the article page. Feedback is aggregated per article and surfaced in the admin grid as an average score, letting editors identify articles readers loved and articles that fell flat — signals for what topics resonate, which authors reliably deliver, and which posts might benefit from a refresh.

This article covers how readers submit feedback, what the aggregate score means, how it is presented to editors, and how to interpret the numbers in the article grid.

Where readers submit feedback

Every published article on the public site has a feedback widget at the bottom of the body — five clickable stars, a submit button, and a short thank-you confirmation after submission. The widget appears after the article content and before any comments section.

The widget is:

• Anonymous by default. No login required. The platform records the rating but not who submitted it (an anonymous browser fingerprint prevents the same visitor from stuffing the ballot, but the identity behind the rating is never exposed).
• One rating per visitor per article. A visitor who tries to rate a second time sees their previous rating and can update it, but cannot inflate the count.
• Optional per site theme. Some organisations disable the widget for editorial reasons (news articles where reader approval is not the goal); check with your theme configuration.

What the numbers mean

Each rating is a number from 1 to 5. The article's aggregate score is the arithmetic mean of every rating received, rounded to one decimal place for display.

A companion count — how many ratings the score is based on — is tracked alongside. A 5.0 average from three ratings is different from a 4.7 average from 200; the admin surfaces both so operators can judge confidence.

Where the aggregate appears

Article grid column

The Rating column on /post shows a coloured pill for every article that has received any feedback:

• Green pill (≥ 4.0) — well-liked. The article resonated with its readers.
• Amber pill (3.0–3.9) — mixed. Some readers found value, some did not.
• Rose pill (< 3.0) — poorly received. Consider reviewing.

Articles with zero ratings show a grey dash (—) rather than a number — a zero average would mislead, since zero from no data is different from zero from lots of one-star ratings.

Hovering the pill shows the tooltip with the exact numbers: "4.7 average from 23 ratings", plus the aggregate age (when the first rating landed).

Sortable

The column is sortable. Sort ascending (worst first) to identify articles that might need improvement; sort descending (best first) to identify articles worth featuring more prominently or using as templates for what worked.

Note: sorting by rating is asymmetric because unrated articles are excluded from the sort direction — they show at the top or bottom depending on ordering. Use the rating column in combination with the view-count column to focus on articles that have both traffic AND enough feedback to interpret.

What the aggregate does NOT tell you

Interpret with care.

• Selection bias. Only readers who chose to rate are counted. Casual readers who liked the article but did not click stars are invisible. Enthusiasts and complainers are over-represented.
• Small samples. A 5.0 average from four ratings is not statistically meaningful. Wait for at least 20 ratings before drawing conclusions.
• Cultural variance. Different reader populations rate differently. A 3.5 average from technical readers may indicate the same article that would receive 4.5 from casual readers.
• Zero-star as a protest. Some readers use a one-star rating as a way to disagree with the article's thesis rather than to judge its quality. Read the accompanying comments (see Comments and moderation) for context.

Tips

• Combine with view count. High views + low rating = worth investigating (perhaps traffic finds the article via search but leaves disappointed). Low views + high rating = candidate for promotion (the few who found it loved it; get it in front of more people).
• Combine with comment sentiment. The rating aggregate is a single number; comments are prose. Reading a handful of comments alongside the rating tells you WHY readers scored the way they did.
• Do not chase the aggregate as a KPI. Editorial teams tempted to make every article "safe" to protect the rating tend to produce dull content that never delights. The bold article that alienates some but delights others is often more valuable than the middle-of-the-road piece nobody hates.
• Re-rate after a rewrite. After significantly updating an article, the historical rating stays — new readers rate the new version but their votes are mixed with old votes on the old version. To reset the counter for a major overhaul, contact support (rare; typically the mix is fine).

Public API exposure

The public API exposes:

• ratingAverage — the aggregate score, or undefined if the article has no ratings.
• ratingCount — the number of ratings the average is based on.

An integrator can present the aggregate however they like — as stars, as a numeric badge, or as an editorial signal — on their own product.

See API integration.

Moderation

The rating itself has no free-text component — a visitor cannot attach a written explanation to their rating. For that, use the comments section (see Comments and moderation).

Ratings cannot be individually deleted or moderated by editors. The aggregate reflects real reader input; interventions would undermine the signal's honesty. If a specific rating is clearly spam or abuse (very rare for a five-star system), contact support for platform-level removal.

Related

• Post management overview
• Comments and moderation — for the free-text companion.
• Views and analytics — the other reader-signal to combine with ratings.
• API integration