Average Visit Length doesn't account for teleports

Reproduction Steps

  1. Create a place in Roblox Studio that teleports to itself.
    You can use the “Teleport to a Reserved Server via Chat” example code here:
    TeleportService:ReserveServer

  2. Join the place and teleport to the reserved server.

  3. Check the place’s stats and note that the Average Visit Length tracker considers each teleport as its own unique session.

Expected Behavior
I expect Average Visit Length to reflect the total amount of time a player spends in my game, regardless of whether they teleport to private instances or other places within the game universe.

Actual Behavior
Average Visit Length treats each teleport as its own session, thus heavily cutting down on the average time a player spends in the game.
You can see the point at which I started teleporting new players to a tutorial place.

Workaround
External services such as GameAnalytics can keep track of total playtime regardless of teleports.

Issue Area: Roblox Website
Page URL: https://www.roblox.com/places//stats
Impact: Moderate
Frequency: Constantly
Date First Experienced: 2021-08-14 05:08:00 (-04:00)

34 Likes

Technically they are different sessions but the most recent data on the graph is usually far lower than the rest

1 Like

This is not being caused by the most recent data being lower. This is being caused by the play sessions where a user teleports to the tutorial being counted in the average.

2 Likes

Thanks, we are looking into this.

5 Likes

Bumping this, just because it looks like whoever responded doesn’t work at Roblox anymore lol.

I’d like to add, my team and I are worried that this affects our experience’s engagement, and thus has prevented it from showing up on any special sorts. What we suspect is that data used by those sorts use the average visit length from the main place rather than the total playtime across the entire experience (so including teleports across the same experience).

Yesterday, we showed up on the Survival sort and gained 200k+ new players, which accounted for over 70% of the players playing. However in our game, when new players join, they are immediately teleported to the tutorial, which leads to a lot of ~1 second sessions which we suspect destroyed our average visit length that day - and was inconsistent when comparing the total sessions against the total playtime. Given how many experiences use the teleportation system, this needs to be fixed so that there is no confusion, and so that our data can be consistent and accurately reflect the true engagement.

I’d also like some insight into how the special sorts (i.e. roleplay, fighting, survival) take engagement into account, and if this might have been affecting it.

8 Likes

Yeah I’ve experienced this bug as well.

Teleporations absolutely destroyed the entire playtime viewer and even if a player has a combined total of 20 minutes in a teleport based session, it’s considered X amount of time before the first teleport.

1 Like

Games like Fish Game and many of the other round-based ones that have teleports for each game within 1-2 minutes don’t have that problem, though if you’re truly teleporting them after 1 second that might fall into a different category.

2 Likes

We have pushed a server browser and server creation update yesterday on SCP: Roleplay and despite having higher retention, similar unique visits and higher concurrent player counts, our Average Visit Length has gone down.

This issue needs to be addressed, it hinders the expansion of the platform with user-generated content based games and affects games that require teleports disproportionately.

5 Likes

Just want to say that this is something that I have seen for years and years. When i first saw that my average visit time was only 10 minutes i thought oh well apparently they just play for a short time. But when i started playtesting in populated servers i noticed that people on average stay longer than 10 minutes in game. Its just that about every 10 minutes or so theres a disaster in my game that players have to escape from which happens via teleporting. I wish the visit time was for overall time spent in an experience as well. It would make this data way better, both in accuracy and usability.

5 Likes

I believe the Average Visit Length only counts the starter place.

From tests I’ve done no matter how long I spend in other places in my game, only spending time in the initial starter place affects it.

This is an extremely big flaw for games with multiple places as there is no way to tell (on Roblox) how long players are actually spending in the game hurting testing and making that analytic stat worthless. Not sure why this hasn’t been fixed yet.

3 Likes

Hi. Can you please share some updates about how this is going? This is an extremely annoying bug.

I’m not sure if this is a solution, but the average visit length is still accurate considering, at least in my eyes, visit length is for the current place, where session length is for as long as the player stays in your experience. The average session length for your game is now shown under the Create Dashboard. I am not sure if this is a proper solution if you wish for all games to have access to it immediately.

For me, this seems quite accurate, where visit length isn’t what I exactly need. That’s half a minute compared to 20 minutes.

image


On top of this, I would like assurance that Roblox is using session length for game sorts and such, as my game currently doesn’t appear on any sorts (source: RoMonitor), which is surprising considering how many there are, even if my game isn’t in its prime. It did seem like my game declined after I switched to a hub model, which basically screwed over my start place’s visit length. I’m not sure if this is because of this, or if there’s some other source which doesn’t relate. This could be evident with this feature request, but I couldn’t know which Roblox uses for discovery.

3 Likes

This is still an issue. My game recently received a decent spike in players, averaging 100+ concurrent users. I have a main “lobby” place, where players enter different gamemodes, and are teleported to the relative places. When this system was implemented, playtime dropped from 4-6 minutes, to 1-2 minutes.

Keep in mind, the game wasn’t the most engaging to begin with - it was originally a 1 player game, so player retention was low. Since opening up multiplayer, average players spiked from 20-100(ish). Playtime should increase. However, since players are being teleported away from the starter place, and into private servers, playtime isn’t being accurately tracked.

I don’t know if this effects premium payouts, or the game sorts on the front page, but it is an issue that’s been around for a while. If for no reason other than accurate data, this should be fixed.

4 Likes

This is, unfortunately, still an issue. Data has smoothed out over time, but still the same trends. I have averaged 100-300(ish) players.

Once teleports were implemented in september (sep '22), it dropped to the average time people are in the main lobby. It doesn’t seem to factor the solo/multiplayer lobbies, where people spend most of their time.

All of the places under the universe have visits. As expected, the lobby gets the most, followed by multiplayer. I would really like to know average visit length for multiplayer.

image

Does teleporting players out of a main lobby affect premium payouts?

FEATURE SUGGESTION

Under the Roblox Creator, there is a nifty “Places” section. Why wouldn’t there be a display for average visit length under these places? It seems to just let you edit the description/title. I would love to see place-specific data under this section.

3 Likes

Average playtime cut in half on Pet Simulator X after I added teleports 3 months ago. That’s absurd. Since sorts rely on average playtime, why is this fix not a priority? My game isn’t appearing on any sorts anymore.

4 Likes

I recently added mobile-only servers to my game and it is impossible to tell if they are having a positive or negative impact because the teleporting from the main place completely ruins the play time analytics

Would also be great to have confirmation on whether this affects premium payouts and visibility on sorts

2 Likes

Hi Preston:
Could you let me know where you get this idea on the average playtime cut in half?
If you check the latest metrics in Creator Dashboard, the average play time does not change significantly for the game you mentioned. Could you provide more details?

Hi Rezrift:
We recently open our new Creator Analytics to all experiences with 50 DAU+. Your experience should also have this metrics. You could now refer to the Engagement metrics to get the most accurate average play time and total visits. We plan to remove developer stats page you are using since that is not accurate. Please let us know if you still have any concern.
image

2 Likes

I still have a concern. Even if Roblox plans to hide the developer stats page in the future, they are still collecting this completely inaccurate data. A lot of developers using TeleportService would rest easier having an official confirmation from Roblox that this inaccurate statistic is not being used for critical things like game discovery (making games with higher average playtime more prominent, for example), Premium Payouts (monetization partially based on how long players spend in my game), and verification status (which is supposedly determined partially by average playtime in my games). If these features are using Average Session Time from Creator Analytics, all is fair. If these features are using Average Visit Length from the Developer Stats page, we have a problem.

4 Likes

I would 100% second this. If all of the algorithms are updated and accurate, an official announcement would ease my (and others’) concerns. TeleportService, as a heavily-utilized Roblox service, shouldn’t disadvantage games using it. It is a really nifty service - very useful - but if it only decreases game accessibility, I am reluctant to utilize it.

An official statement would ease my concerns.

1 Like