liveBook · Manning

Criticism

As explained in AI-Powered Search

Negative feedback is information indicating that a system result or recommendation was unsatisfactory or irrelevant.

Criticism is negative feedback indicating that a search result or recommendation was unsatisfactory or irrelevant. Search systems can capture it through explicit actions such as thumbs-down votes, removals, returns, or negative reviews, and implicit behavior such as skipping a displayed document. section 8.4

A weighted ranking model incorporates criticism by assigning negative values to dissatisfaction signals. The resulting score combines positive and negative evidence, while domain-specific weights determine how strongly each action changes ranking. section 8.4

Typical criticism signals include:

positive_signals = (1 * sum(click_signals)) +
                   (25 * (purchase_signals)) +
                   (10 * sum(add_to_cart_signals)) +
                   (0.025 * sum(seen_doc_signals))

negative_signals = (-0.025 * sum(skipped_doc_signals)) +
                   (-20 * sum(remove_from_cart_signals)) +
                   (-100 * sum(returned_item_signals)) +
                   (-50 * sum(negative_post_about_item_in_review_signals))

type_based_signal_weight = positive_signals + negative_signals

Shows a configurable aggregation in which clicks, purchases, cart additions, skips, removals, returns, and negative reviews receive separate weights. section 8.4

How AI-Powered Search covers it

How deep it goes
Explains signal collection and weighted scoring, but defers click modeling and automated weight tuning.
Examples
Uses thumbs-down votes, skipped results, cart removals, product returns, and negative reviews as examples.
What it assumes
Assumes familiarity with clicks, add-to-cart events, purchases, and signal-based ranking.
What you can do after
Apply it via a weighted ranking formula that penalizes dissatisfaction alongside positive engagement signals.
How it connects
Connects negative feedback with user intent, signals-based ranking, click modeling, and learning to rank.

Also explained in

Compare how 2 books explain criticism →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of AI-Powered Search. The section links point to the parts of the book it draws on.