A few early ratings can lock an item's long-term visibility

What it is

Early-rating lock-in is the long-term visibility consequence of a few ratings on an item when ranking, recommendation, and social proof amplify them. Random error in a tiny sample becomes path dependence that later users struggle to correct.

Why it happens

Initial rating affects display; display creates more similar rating, while low display prevents later disagreement. Visible scores induce conformity, so data both measure and change the object. “Seen first” becomes mistaken for “better.”

Studying it

• Paradigms: randomly vary or hide initial rating and trace later rank, rating, and independent quality.
• Variables: initial sample, score visibility, exposure, later rating, rank, and quality.
• Methodological caution: separate rating effect from posting time and author-network causes.

Where it stops holding

Early quality signals can sometimes predict well; the issue is small, opaque samples with hidden uncertainty. Urgent material may need immediate rank but should be reviewed.

Applying it

• Show uncertainty and reduce rank weight under small samples while retaining exploration.
• Delay totals or collect independent ratings first.
• Re-sample and recalibrate historically low-exposure items.
• Verification: audit sensitivity of long-term rank to first ratings and discovery of high-quality low-sample items.

Related

• Same group: V9.07.1 Group estimates outperform individuals only when judgments are independent · V9.07.2 Visible choices break independence and induce conformity · V9.07.3 Collaborative filtering reinforces popular items and buries the long tail · V9.07.4 Groups do not outperform experts on specialised questions
• Nearby: V8.03 Content quality · V7.03 Reputation mechanisms
• Search terms: path dependence · early rating · popularity lock-in