Genuinely shocked how long this has been swept under the rug. I dont have anything else to say since my last few replies have been saying virtually the same thing - constant loss of stats with no sign of recovery. Search sucks. Home sucks. The algorithm pushes new games and tosses out old ones. Yada yada yada
2 Likes
Glad to see I’m no the only developer having this issue. My game wen’t from 1.5k ccu to 100 in the span of 1 week after the release of Roblox on Playstation on the 10th of October. All this during the time where my stats where looking best and I was Updating atleast once per week.
3 Likes
This happens to games that hasn’t been “sorted” yet. When I search up the name of my game, It never shows up for me, but I still get 4-6 new players daily, who somehow find them. I asked them if they searched for it directly and they tell me no, that they were looking around and the game just popped up.
1 Like
What does it mean that a game is “sorted”? How do I get our game sorted, to appear on searches?
Go to the Recommended section of the game page. “Unsorted” will look something like this, where the recommended games are completely irrelevant. This is the recommended for my TYCOON game.
Here is a Tycoon game that is “Sorted”
But I’m pretty sure i’m partially wrong about the whole thing. Here’s Child Labour Tycoon 2 which shows up when you search it but isn’t sorted.
I think stuff showing up when you search the name is more visit dependent.
2 Likes
That Recommended Experiences thing is the EXACT same as on my game!
And on any other game with low visits
exact same thing…
It seems like the Recommended Experiences just use some default experiences if they aren’t popular enough…
And I’m pretty sure those recommended games have been the same for at least a couple years
5 Likes
We need an algorithm update asap.
2 Likes
It’s still private though, but here’s a screenshot
That is with around 150 games in the system, so this is not representative of how my discovery game will be, but each sort will contain around 1k games, and some sorts have some randomness, which should create a good variety
6 Likes
Interesting. So you’re making your own discovery page?
3 Likes
This looks good though?? Please don’t let this idea die.
4 Likes
My game has been hit again for the third time this year, truly unsustainable.
4 Likes
To all Roblox game developers who are struggling to promote their games, I want to give you all some encouragement, as a fellow developer.
As a developer myself, I know we put so much effort into our games, and how frustrating it is when that effort goes unrecognized. Especially after trying over and over to market games without success, or with meager results.
But don’t give up on your games you’ve worked so hard to design. For all the effort you’ve put into your games, they’re more than capable of helping people, by bringing happiness, even if to only a few people for now.
The marketing and visibility struggle is real, I’ve experienced much of it myself. But your projects are valuable and filled with inspiration whether they’re seen by players or not.
4 Likes
I’ve a similar project with +1k games I made ~1 year ago. I made a tag system similar to Steam so you can filter games by (multiple) genres and also find stuff you didn’t play yet. One of the major goals was to have granular search to find new games on genres that you like.
Unfortunately, it never took off as Roblox algo doesn’t recommend this type of experience (minimal playtime due to the nature of the experience and no monetization). There are other experiences with similar intents (record all story games, find random games), but all are empty. The only exception with a few hundred of players is Walmart Discovered, but I suspect most of the player base comes from immersive ads, as I’ve seen the ad quite a lot.
In my own experience this approach to try to give players and developers something beyond recommendation doesn’t work at all. 
The Recommended Experiences are the same for every new game/game that never got any traction until the algo establishes their genre internally and associate with similar experiences. But this is not a major problem, since once the game gets some traction, that tab usually gets filled with relevant suggestions. Meaning this default experiences must not get almost any traction from this default state.
I also can confirm these games are the default for at least 1 year because I first noticed this when I developed my gamepage around 1 year ago.
1 Like
I remember before around 2021, it used to just be so easy to gain a playerbase for your game from just search discovery alone. My new games from the last 3 years don’t even show up even when you search their exact title. It’s honestly just so demoralizing spending years creating these new games and getting absolutely zero traction.
Sure, I can try to sponsor/advertise, but the popularity only lasts for roughly a week because there are literally no new players coming in from search discovery.
It’s a dilemma as well. They say the algorithm promotes games with high engagement, retention and monetization, but how the hell are you supposed to achieve any of that if you are not even given a chance in the first place.
In addition, many of the games promoted by the algorithm are straight up copies, scams, or have very little effort put into them. I am just so disappointed that I’ve invested all of this time working on my games and nobody can even find them.
7 Likes
We released another game just about a month ago. Here’s how it went.
This game, Tropical Resort Tycoon 2, is the sequel to our most popular game, mentioned above in the post I’m replying to. We improved our statistics across all categories, getting them higher than any of our other games, and they have stayed well above top tiers of available analytics page benchmarks. Our rating sits stable at 98% after millions of visits and over 40 thousand votes.
Let’s take a look at our analytics and try to break them down. Here’s our first month of acquisition:
The algorithm clearly picked up on the game on the 10th of December. We’re very grateful to have been shown to so many players at that time, but what happened after is extremely concerning to me - because we have seen only one correlating statistic:
Home Recommendations Conversion Rate, AKA how many players click your icon, and then click play.
Your first instinct when seeing this may be something along the lines of “If your conversion rate is lower, it makes sense that new users per impressions is lower that day”.
However, taking a closer look, that’s not the case at all. It actually takes one or two days for the conversion rate changes to show up in the new users graph. The response is proportional, ping-ponging back and forth until reaching some equilibrium, presumably due to there being more competition with click rates when your game gets shown more among other hyper-optimized game icons, titles, and descriptions.
So, I have some questions:
- Why is this single statistic so much more important than the others in terms of having your game recommended to players?
- Why are there no benchmarks for conversion rate statistics?
- What are we, as developers, supposed to do?
Are we forced to make our icons and descriptions as clickbait-y as possible?
Are we supposed to keep changing them to keep click rates up, deceiving players and making them think it’s a new game / update, when it isn’t?
How is this at all sustainable for us, or Roblox as a platform?
10 Likes
I believe there is an alternative explanation you can infer from both graphs:
The lower the amount of impressions the easier it is to fall on the extremes. This seems to be the case in the graphs that you showed since in dec 8-9 you had very few new users and a really high conversion rate. Then, when the impressions went up (algo pushing your game) and new users went up, the conversion fell to a more realistic value and started looking for the equilibrium since impressions are a finite (and uncertain) resource.
So, to me, this shows the algorithm adjusting how many impressions you are getting based on the perceived conversion rate and performance of your game and thus how many players are reached by your game. The algo decided that this is the optimum for your game given it’s current resources.
To me it seems that once the algo decides your game belongs on a certain spot, due to the lack of discoverability options, you are forced to stay in that spot and there is a hard ceiling. This has pros and cons: one single super-optimized game wont gobble all the impressions but everyone’s impact on the algorithm is also capped.
3 Likes
Yeah, this is pretty much all the same as what I meant with the “ping-ponging back and forth until reaching equilibrium” part. I don’t think there’s any disagreement about what the graphs mean. It wasn’t really the main point though and the questions remain.
Conversion / click rate having such an extreme amount of influence on the discoverability of a game, seemingly overpowering any other metric, is concerning. This absolutely can, and probably does, kill would-be high-quality and long-lasting games, in favor of games with a fresh coat of paint. Roblox has always been this way to some extent (I’m a 15-year veteran on the platform) but now seems much more skewed.
3 Likes