What can Steam do to improve game discovery?

[ZedClampet]

My current complaint involves the "More like this" section at the bottom of a game's store page. This section finds popular games with similar tags to the game you are looking at. That's fine, but why does it show me games I already own? Also, I would like to see some options with this, maybe a slider, to look for more rare games.

What kinds of things do you think would help discovery? Closing the new game floodgates isn't a realistic option.

 

[Brian Boru]

What kinds of things do you think would help discovery?

Some guy somewhere making a monthly post about new releases that month.

similar tags to the game you are looking at

That's the problem, the tags are pretty suspect if my 'More like this' is any guide.

I would like to see implementation of something like the Quantic Foundry Gamer Motivation Model to identify games close to the current choice.

It should go beyond the basic model by taking Hours Played relative to other players into account—the 'relative' so that a 100-hour slog doesn't overshadow a 15-hour gem. Incorporate a rating score too.

Beyond that, let Steam fund-support people or groups whose task is to evaluate & pub reviews of all small releases in the reviewer's fav genre(s)—those which are not covered by big outlets. It can run a week or month behind. I doubt Steam can employ people to do this, since dissing products you're selling won't go well with dev, pub, or Steam marketing.

 

[ZedClampet]

That's the problem, the tags are pretty suspect if my 'More like this' is any guide.

The thing is that you can make it work with the current tag system, apparently, because Steampeek does okay with it. It's not perfect, but it's better than Steam's system. I suspect this may be because Steam takes popularity into heavy account, so much so that you are constantly getting recommended the same games even if they aren't particularly similar. This really needs to stop because it's a system that perpetuates discovery problems.

Other than that, it would be nice if the tags were weighted somehow by major categories set by Valve. For instance, if two games are both "farming" games, that's much more important, probably, than other tags you could apply to them.

 

[Pifanjr]

What's in it for Steam? Statistically users are probably much more likely to buy a game that has been bought a lot already, so it makes sense showing popular games in the "more like this" section instead of obscure games that are actually more like the game you're looking at. Though I do agree that they shouldn't show games you already own.

 

[ZedClampet]

What's in it for Steam? Statistically users are probably much more likely to buy a game that has been bought a lot already, so it makes sense showing popular games in the "more like this" section instead of obscure games that are actually more like the game you're looking at. Though I do agree that they shouldn't show games you already own.

I don't think users in my category (and there are legions of us now on Steam even if I'm the only one on this board) are more likely to buy popular games. We're the explorers. We play odd games that only have a few reviews. Valve has mentioned this group before. I'm playing a game now that is basically impossible to get to show up on discovery, but it has still sold thousands of copies (per SteamDB estimates). But even if this group were likely to be swayed by popularity, then why call it "Games Like This One"? Just say, "Here are some more games you might like." Saying "games like this one" is misleading and frustrating.

And if you say, "Well, thousands of copies is nothing," then that fails to see that most of these are single developer games and you can make a decent living selling thousands of copies a year. And, of course, a lot of these games do better than that. Something as stupid as Pumping Simulator 2 has sold about 30k copies.

 

[Zloth]

Better searches would require a lot more effort, both from Steam and from us. We would have to get serious about nailing down exactly why we like/dislike particular games - none of this "it just didn't click with me" stuff. Meanwhile, Steam would have to collect data on all the craziness we find ourselves keying off of for every game they've got. Ghaaa! Too much work for too little gain!

I think iteration on what's there would be the better way to go. Letting players put in their own tags didn't work well. Is there a way players could at least verify that tags given by each game's marketing division be accurate?

 

[ZedClampet]

Better searches would require a lot more effort, both from Steam and from us. We would have to get serious about nailing down exactly why we like/dislike particular games - none of this "it just didn't click with me" stuff. Meanwhile, Steam would have to collect data on all the craziness we find ourselves keying off of for every game they've got. Ghaaa! Too much work for too little gain!

I think iteration on what's there would be the better way to go. Letting players put in their own tags didn't work well. Is there a way players could at least verify that tags given by each game's marketing division be accurate?

The closest we could come for a viable search would be if you used an AI that reads all the user reviews and forum posts and store page info and creates a detailed summary of the game that could be used by a search AI. Type in plain English what you want and it gives you the results. Obviously, some of the stuff it reads wouldn't be 100 percent accurate, but no system is going to be perfect.

As for tags, I think the current system works okay. If a game has 10 tags and one of them isn't right, that doesn't make the system broken. What is really broken is that there is no weight applied to any of the tags. I'd like to think that letting developers add 3 key tags would be helpful, but I know that in reality there would be a lot of developers who would just use whatever tags they think would get them the most views like "Palworld, roguelike, fun" and thus it would be useless and actually worse than what the players are doing.

 

[Brian Boru]

Wishful thinking, but would be great if university gaming depts organized students to play every new release on various platforms and post useful data incl categorizations and reviews to a central wiki—for course credit of course, so there'd be an incentive to do it properly.

AI will be the answer in the future, this is a task tailor-made for it.

 

[Colif]

Some guy somewhere making a monthly post about new releases that month.

Good idea, just need to find a forum where there are more than 20 odd (we know who you are) users and a bunch of spammers... then post it there... then more people might know

 

[ZedClampet]

Good idea, just need to find a forum where there are more than 20 odd (we know who you are) users and a bunch of spammers... then post it there... then more people might know

We do seem to have some readers who never post, just based on how many times each thread is looked at.

 

[Brian Boru]

We do seem to have some readers who never post

Sadly the opposite too—some posters who never read

how many times each thread is looked at

A lot of those views will be search spiders.

 

[Zloth]

The closest we could come...

I was thinking the same, but I don't believe user reviews would work very well. The AI would need data on what's in the game, like story elements, game mechanics, music styles, graphics, art direction, save mechanics, and so on. Reviews will sometimes list some of those, but that isn't the main point of reviews and few people make any attempt to list all of them. Pro reviews do a lot better, but smaller games get few, if any, of those.

Indentured serv... <ahem> students would work much better, especially if you can get a really wide variety of backgrounds and age groups in there.

We do seem to have some readers who never post, just based on how many times each thread is looked at.

LURKERS!! Who are they? Forever watching, never talking... What do they want? Are they judging us? Do they keep elaborate spreadsheets tracking our every grammatical mistake?? Maybe they look up who our old teachers were, then play something like Fantasy Football, drafting the teachers onto teams, then scoring them according to our posts! Or maybe it's time I took a nap?

P.S. Web crawlers should be identifying themselves to make them easy to take out of the counts.

 

[ZedClampet]

I was thinking the same, but I don't believe user reviews would work very well. The AI would need data on what's in the game, like story elements, game mechanics, music styles, graphics, art direction, save mechanics, and so on. Reviews will sometimes list some of those, but that isn't the main point of reviews and few people make any attempt to list all of them. Pro reviews do a lot better, but smaller games get few, if any, of those.

Indentured serv... <ahem> students would work much better, especially if you can get a really wide variety of backgrounds and age groups in there.


LURKERS!! Who are they? Forever watching, never talking... What do they want? Are they judging us? Do they keep elaborate spreadsheets tracking our every grammatical mistake?? Maybe they look up who our old teachers were, then play something like Fantasy Football, drafting the teachers onto teams, then scoring them according to our posts! Or maybe it's time I took a nap?

P.S. Web crawlers should be identifying themselves to make them easy to take out of the counts.

I think user reviews might still be helpful. I get enough concrete information from them to read them before I purchase a game. Certainly, they are imperfect and many of them don't list any mechanics or features, but I think the few that do might be helpful. What you don't want is for the AI to confuse suggestions with actual features, and that might be very difficult right now. The general purpose AI's like ChatGPT and Bard seem to struggle with that, but I'm not sure how much customization is possible with those. You might be able to modify them to know the difference.

 

[ZedClampet]

Sadly the opposite too—some posters who never read

TLDR

 

[Colif]

We do have some lurkers, a lot of whom might only show up when there is a competition running. That and a mention in discord normally brings them out of woodwork... if only to post in one thread and leave again.

The community being split up over 3 different chat interfaces doesn't help. Article comments not coming here means less reason to come here... discord server appears to get more users. I am not in there enough to know for sure.

One way that might get more eyes on that thread of Zed's about releases would be for someone to post a link to it into the announcements in discord server, but it would appear only CM can over there - which makes total sense. Would need to ask one of them.

 

[ZedClampet]

We do have some lurkers, a lot of whom might only show up when there is a competition running. That and a mention in discord normally brings them out of woodwork... if only to post in one thread and leave again.

The community being split up over 3 different chat interfaces doesn't help. Article comments not coming here means less reason to come here... discord server appears to get more users. I am not in there enough to know for sure.

One way that might get more eyes on that thread of Zed's about releases would be for someone to post a link to it into the announcements in discord server, but it would appear only CM can over there - which makes total sense. Would need to ask one of them.

Ew. Discord. No idea why people like posting on Discord.

I'm not really concerned about how many people see the games post. Currently I'm more concerned with it being unusable due to length and the inability to add navigation to it.

But a few people seem to like the post anyway, so I don't mind going back to making it. Plus, even if no one actually uses it, I used it myself and miss having it.

 

[Brian Boru]

inability to add navigation

Split it into a few posts, with the nav to other posts in #1.

When you post #1, make enough 'Reserved' replies immediately so no one else gets in the way. If I recall correctly, you have an hour to make your edits before that function disappears—but of course if necessary one of the serv mods can help with that.

 

[ZedClampet]

Split it into a few posts, with the nav to other posts in #1.

When you post #1, make enough 'Reserved' replies immediately so no one else gets in the way. If I recall correctly, you have an hour to make your edits before that function disappears—but of course if necessary one of the serv mods can help with that.

Did you read my post in the demo thread? Because this was where that was headed. I just figured out today how to link to a particular post within a thread.

 

[Brian Boru]

Did you read my post in the demo thread?

Yep, after this one

 

[ZedClampet]

Yep, after this one

Links pasted into Word don't explode when copied and pasted into this forum for some reason. I'm probably pasting too much background info from Word. Probably the easiest solution would be to paste it into Word in a way that doesn't bring all the baggage. I know how to do that in Excel, and I assume it's the same in Word, but I haven't tried it yet. I've never really used Word all that much except for simple document creation. I wrote a bad novel in it once, but that didn't teach me anything except to stop writing novels.

Public Service Announcement: Don't ever give my wife something to read if all you are looking for is positive reinforcement. She will smite you with the club of truth.

 

[Colif]

The unfurl function isn't fully understood. Two links that are exactly the same can respond in different ways when posted... we still not sure why.

So one way to work with it is to label the links with the name of the game
Starfailed
example above. but then none will explode as you say. At least all links would be the same.

 

[Brian Boru]

explode

Just curious—I assume you want them to explode?

If you want Word to not bring crap with it to forum when pasted, save finished DOCX as a TXT file, open that in text editor, and copy-paste from there.

Alternatively, use 'Paste as Text' shortcut Ctrl+Shift+V.
Or use 365 clipboard—the one which holds multiple entries.
There's also 'Paste Special' via the Ribbon's Paste dropdown.

paste it into Word in a way that doesn't bring all the baggage

If that's better for your workflow, above should also work for it. I do a lot of text cleaning, so I've reassigned my KB 'Break' key to 'Paste Plain Text'—nice time saver and less finger gymnastics

PS the Ctrl+K shortcut works in forum for making links like Colif's example, quite handy.

 

[Zloth]

Public Service Announcement: Don't ever give my wife something to read if all you are looking for is positive reinforcement. She will smite you with the club of truth.

Dragon Age told me it was a sword!

View: https://youtu.be/uwK5p6JrkK4?si=JNr6mRUmKIVY-KLn

 

[ZedClampet]

Just curious—I assume you want them to explode?

Why just unfurl when you can explode? This is a gaming forum. If we shoot a red barrel, do we want it to unfurl? Do we want the enemy to "pass on"? Do we want to open a chest and find "goodies"? No, dammit! We're all men here, except for the women, and we want stuff to explode!

But, yeah, I couldn't think of the actual word.

 

[Colif]

But, yeah, I couldn't think of the actual word.

I couldn't either, went and looked at one