Finding a game should not take longer than deciding what you want to play. Yet a player can open a store with a clear intention and leave with a longer wishlist rather than a game for tonight.
\nA useful recommendation begins with constraints. Which device is available? Is the player alone or with friends? Do they have twenty minutes or an entire evening? Are they looking for a familiar experience or something different?
\nThose questions describe a decision. A catalogue describes supply. Showing more supply does not necessarily help someone make that decision.
\nConversational discovery offers a different starting point. A player can describe what they want in ordinary language and receive a small selection with a reason for each suggestion.
\nThe explanation matters. “You asked for a short co-op session on PC” is more useful than a generic claim that a title is popular. It also gives the player something to correct: less competitive, easier to learn, or suitable for a different device.
\nThat feedback can improve the next recommendation. It should not be presented as proof that a system understands someone perfectly or will never make a mistake.
\nA friendly interface cannot compensate for incorrect platform support, an unavailable store link or misleading gameplay information. Recommendations need a reliable catalogue and clear distinctions between a game’s actual features and an inferred match.
\nSponsored placement should also be distinguishable from a recommendation based on fit. Otherwise a new interface risks reproducing the same trust problem as a crowded feed.
\nFor a developer, a recommendation is useful when it brings an appropriate player closer to a meaningful experience. Clicks alone cannot show that. A demo start, a completed first session and a later return answer different questions about the quality of the match.
\nMy view is that discovery should be judged by how well it helps people choose and start playing. Reducing scrolling is a useful design goal; sustained player interest is the harder test.
\nWould you rather start with a conversation or a catalogue?
","content_text":"Finding a game should not take longer than deciding what you want to play. Yet a player can open a store with a clear intention and leave with a longer wishlist rather than a game for tonight.\n\nStart with the player’s situation\n\nA useful recommendation begins with constraints. Which device is available? Is the player alone or with friends? Do they have twenty minutes or an entire evening? Are they looking for a familiar experience or something different?\n\nThose questions describe a decision. A catalogue describes supply. Showing more supply does not necessarily help someone make that decision.\n\nRecommend, explain, refine\n\nConversational discovery offers a different starting point. A player can describe what they want in ordinary language and receive a small selection with a reason for each suggestion.\n\nThe explanation matters. “You asked for a short co-op session on PC” is more useful than a generic claim that a title is popular. It also gives the player something to correct: less competitive, easier to learn, or suitable for a different device.\n\nThat feedback can improve the next recommendation. It should not be presented as proof that a system understands someone perfectly or will never make a mistake.\n\nGood discovery still needs reliable information\n\nA friendly interface cannot compensate for incorrect platform support, an unavailable store link or misleading gameplay information. Recommendations need a reliable catalogue and clear distinctions between a game’s actual features and an inferred match.\n\nSponsored placement should also be distinguishable from a recommendation based on fit. Otherwise a new interface risks reproducing the same trust problem as a crowded feed.\n\nMeasure what happens after the suggestion\n\nFor a developer, a recommendation is useful when it brings an appropriate player closer to a meaningful experience. Clicks alone cannot show that. A demo start, a completed first session and a later return answer different questions about the quality of the match.\n\nMy view is that discovery should be judged by how well it helps people choose and start playing. Reducing scrolling is a useful design goal; sustained player interest is the harder test.\n\nWould you rather start with a conversation or a catalogue?","referenced_links":[],"social_image":"https://reinouttebrake.com/assets/editorial/discovery-photo-7862378-1280.webp","citation":"Reinout te Brake. Game Discovery Without Endless Scrolling. 2025-10-02. https://reinouttebrake.com/stories/game-discovery-without-endless-scrolling"}