Forgiven 
Steam uses several discovery systems rather than one public master algorithm. The practical signals are the information and player responses attached to a particular store surface.
Relevance signals
Tags, categories, features, languages and store text describe the game. Accurate metadata helps players and store systems understand the product; keyword stuffing creates poor matches.
Presentation signals
Capsules, trailer and screenshots determine whether a relevant impression becomes a useful store visit. They are not ranking “weights”, but they affect player understanding.
Commercial and player response
Purchases, play, refunds and reviews describe parts of player response. Valve does not publish a formula assigning each one a universal ranking value.
Wishlists and notifications
Wishlists let users save a game and can trigger release or discount notifications. Valve does not publish a guaranteed threshold for placement.
Events, updates and discounts
Eligible events, meaningful updates and compliant discounts can create additional discovery opportunities. Each has its own rules and timing.
External activity
Creators, communities, email and advertising can bring qualified visitors. UTM Analytics reports some attributable actions but does not prove an automatic ranking bonus.
Use the official map
Valve’s visibility documentation is the starting point. Where Valve does not describe a signal, label the claim as unknown.
