Hi Everyone,
We’ve been working on some potential changes to how we bill for Unity Multiplayer and would love your feedback. If you’re interested in doing a ~30 minute call to go over a some proposals and provide your opinions we’d love to hear from you. Just grab a time that works for you here: Calendly - Derek Bronson
If those times don’t work out for you message me privately and we’ll work something out.
Thanks,
Derek
Hey DBronson, i think it would be better if you just wrote down your potential changes 
Otherwise i have some suggestions,
500 CCU - unlimited or very high bandwidth limit like something along 15 kb/s - $50-100 (Plus or Pro also needed)
1000 CCU - Unlimited bandwidth with some kind of fair usage so people don’t go crazy maybe like 25 kb/s $100-150 (Plus or pro of course also needed)
Then if you want to go higher than 1000 CCU you can pay $100 monthly for every extra 1000
I feel this way would seem less “frightening” for users since you know how much you will pay and can budget accordingly, also at this moment i can’t even see how much i’ve used this month so in theory i don’t know if i will have a $1000 bill incoming. (Obviously i won’t but you get the point). Otherwise maybe i haven’t just been able to locate the tab showing currently monthly data usage in which case i would love if you could
Edit: I haven’t got my first bill yet, so i don’t know how prices compare to CCU really, but of course you should do the Multiplayer services as cheap as posible. Otherwise people will just get Dedicated servers instead
I think that it is essential to provide developers with a transparent option to clearly retrace the bandwidth consumptions which are actually charged by Unity so that we can understand the costs. As of today, however, it seems that such an option is not available yet. Good news is, on my question here, Richard has confirmed that it is planned to be made part of the analytics dashboard. This is good to know and I think it’s about time.
I would also like to suggest that you consider to separate matchmaking & server browser services from the relay server service.
For example, let’s say I would like to use a native unity solution to list available servers for my game, or, alternatively, to implement some other kind of matchmaking. Of course, this needs a server hosted by Unity. So, it is reasonable to pay a fee for this service. However, I could as well decide to use a third party solution for matchmaking.
Then, the next step is the actual connection. I think a quite good solution is like this:
Try direct connection first. If this works - fine. No need for everything else or to pay anything further.
If direct connection fails, try NAT punch through (a native solution for this seems to be currently in development by Unity, a solution in the asset store is available already).
If step 1 and 2 both fail, it’s nice to have the option to fallback to relay servers in order to guarantee a 100% connection in any case if wanted. Of course, this is a service that cannot be provided for free. Notably, however, this is an optional step. Developers could just decide to use steps 1 and 2, but not necessarily 3. Or, they could decide to use a third party solution for relay servers.
So, this is why I think it might be a good idea to separate these services to provide higher flexibility. Still, it might be an attractive solution to run a game with overall native Unity services if reasonable.
Please post your proposal, either straight up, or as poll(s). Who has time for phone calls, Derek?
I would suggest that you not forget the hobbyist, and realize that your competitors are offering 20 CCU for no cost. Want to take the business from them? Don’t jut match it, beat it, and offer something that sweetens the pot. Like SIMPLE networked object pooling, or some other means to optimize bandwidth/performance.
Just my 2¢
Mark.
@Rungsted93 - Thanks for the feedback. Your point regarding how “frightening” it is for users is spot on. We are currently working on reporting usage in the dashboard as well as adding better tools for measuring usage of your game before you release it.
@moco2k - There has been some talk on splitting out matchmaking/server browser into a standalone service. With regards to your other points we currently have a developer working on NAT punch through and ensuring that the pipeline with the service is smoothed out so developers can take advantage of the 1-2-3 flow you are describing. We don’t have an exact timeline but I can say that it is currently the #1 feature on that person’s plate.
@Mad_Mark - Point taken Mark. We’re finalizing an idea or two and will share them with several people for more offline based feedback.
@Mad_Mark - I just wanted to second this point. The ability to use the free tier as a stepping stone is an essential asset to some developers that have to face so much uncertainty. It nice to be able to hop back to shore if the stepping stone crumbles with as little cost as possible.
I haven’t actually got a full feature game built yet with Unet or even unity(used monogame previously), but If I hit success then naturally I have no problem paying a fair price for networking where more users connect to my apps. I also think it makes sense for the tiers to be part of the unity subscription but only if the Unet package scales with the unity price sensibly(the best price for 500 ccu will probably not match with Unitys plus price for a crazy example).
I also just want to post a link to this thread because I think it has a very good analysis on pricing and Is part of the reason I haven’t given up on Unet after looking at the competitors.
I just wanted to voice my opinion if it s not to late, though im sure that a competitive pricing plan will be implemented by Unity anyway.
I don’t know what the networking services market is like nowadays, mostly because I’m not doing anything big, but also because any larger CCUs are “contact us for prices”. This is code for “you can’t afford it”
But personally, I can make my own matchmaking service and run at least 300k CCUs at a very low fixed cost (one rented server dedicated to the purpose, no load-balancing). Yes, it’s a bit more work and fallback options add the cost of a server for each level of redundancy, plus monitoring. But it’s still cheaper than numbers I’ve calculated from AWS and similar elastic cloud services.
So just matchmaking should somehow match that, or add very useful benefits. Possible bonus features could be things like account systems (UT handle user accounts for your game too - huge benefit right there), payment gateways (they handle accountING as well), or at the very advanced level, host your game servers for games where that is a thing.
All the services I’ve seen use different schemes which make it hard to compare them, of course, so I think simply copying the competition, streamlining it and doing a bit more is how you can make your service the desirable one.
A friend of mine was saying how expensive networking is so I went to have a look and I must admit that I cannot find where he got his prices from. Getting a pro subscription includes more CCU and it cost less than what he described… but I digged deeper and went to use the calculator.
From my point of view, the pricing looked really cheap based on the data I was plugging in there. Of course, I don’t have a cooking clue how much data I will be transmitting yet so those numbers are bogus at best. Following a link to the forum I read that Unity is trying to charge you ONLY for what you use and thus keep the costs to a bare minimum and the poster there even said himself that the calculator is a best guess effort and should not be trusted yet… See their pricing as a general guideline and hope and pray…
Here are a couple of thoughts I have about networking prices, in no particular order:
I do NOT want to pay a once off free PER GAME for X users max. If I pay for 500 users and only get 3 ppl to ever play that game in total (much less at the same time) then I don’t want to pay another $2500 for another 1500 CCU for my next game.
I don’t want to be told “Pay us a fortune and you can have as many CCU as you want. You will have to provide and pay for your own hardware to and run your own server after paying us this small fortune, though”
I am not prepared to launch a game that I wish to become successful but with my lack of advertising experience might not even reach 10 people but if 21 people take an interest, then the game will simply deny everyone else access while I am trying to get the game discovered. 20 CCU to me is a joke and I wouldn’t even launch a demo on my website with that limitation.
In contrast, paying for 500 CCU every month, hoping that 10 people discover it is also daylight robbery.
The final part of this catch 22 is that Unity offers the option of one only paying for what one uses and no more. That sounds perfect… except for one small catch… I want to make free to play games so I might or might not get 500,000 downloads in my first month or three, with 0.001% paying me $1 per month. For all the $1000 I make per month I certainly can’t survive if 2 million people downloaded the game and just checked it out for a day or so then 99% said “nahhhh” and left me with a $20,000 bill to Unity because people just tested the game then left… ouch.
I am blowing the figures way out of proportion here but way back in the days of Unity 1 and poor discovery mechanisms and far far far less users than now, every single demo I released on the iStore to see if people like the game so I could do more with the game, nothing had less than 500,000 downloads by the time I stopped bothering to look. 500,000 downloads, 2 million plays, 1000 remaining players afterwards and a bill that cannot possibly be calculated in advance… This “we don’t know how much we are going to charge you” thing is the real reason why I don’t do multiplayer games…
It boils down to this:
Either you pay for a set number of players and if you are popular then you will just have to tell the bulk of your players “sorry, go away, I can only handle this many people” and never even know exactly how many people were interested…
another option is for my plan to automatically upgrade if and when the users exceed my cap, again limiting me to another cap I did not choose and just adding to my bill without my having a say in the matter … only to then stay that high for the following months even after interest has died. No automatic downgrading of your service is there? Just automatic taking of more money.
And the last option is to not have any cap on your users and just send you a bill for any amount under the sun. Maybe $3 maybe $25,000… just wait till the end of the month and be sure to pay up by the 5th, thank you. No way am I going that route…
So chase away your customers or bankrupt yourself. Pick one.
SOLUTION: I prefer option 4
What Unity has done all these many years so far… Give us a testing service that we can use to test our games on for free, no matter how many users end up playing it… but once we release the game, give us the software to run on our own servers so we can use the built in networking exactly as it is but from our own servers without any CCU cost as long as we run it on our own hardware. Of course, for those of us who don’t have the skill to set this up, let THEM choose to use your servers and pick from the above options…
I would rather figure out the installation of server software that you provide us (Master Server and facilitator) and not have to make the choice between showing customers the door or bankrupting myself. At the moment what I am doing is running a server instance of my game on my own server at home. I have a fast enough line to handle the game working as a master server (or an authoritative server if I so choose) and I have uncapped bandwidth from my ISP so that means all it is costing me is $2.50 for the static IP every month and the once off $150 for the second hand Mac Mini I bought as my server… IfI could have a master server like you used to provide so I don’t have to run a full game to act as my master server then I would prefer that option any day and just run that on a VPS that costs me $5 per month and doesn’t give a frag how many people are playing my game…
But as it stands I feel you give me only 2 options: Chace people away or bankrupt myself. Solve that issue for me and you will have the answer you are willing to make all those phone calls for. Find a means of offering a middle ground to those two extremes…
How about a free, limitless master server during testing and 3 months free usage after publishing. In 3 months we will have built up our userbase, seen if the game is going to be worth it or not to keep online and you can tell us exactly “This is what your bill would have been for these three months”. Based on our user base at that time and the figures you provide us we can then either say “Nah, I am going to remove this game from the iStore and cut my losses” or say “Okay, this could work” and know how much to factor into our budgets to pay for our networking. No more “No clue how much until the bill arrives” nonsense…
I think Unity should get rid of the free CCU for games in development so that the much smaller number of actually released games don’t subsidize the unreleased. That way Unity can drastically cut the cost per GB for released games, hopefully down to below $0.10 / GB. Replace free CCU with developer configurable daily usage caps to prevent runaway bandwidth usage during development.
Right now the high cost of data for a released game is basically a hard barrier for an indie developer putting together a free to play game, where there will be relatively high network usage compared to paying customers.
It doesn’t help the developer having free bandwidth during development if virtually all income to their project gets sucked up in networking costs after release.
EXCELLENT post and I fully agree. My biggest worry is the above. Based on demands from my user-base I was thinking of rewriting and upgrading my (successful) family quiz game on Apple app-store from XCODE OBC to Unity and UNET but as MrDude stated the feeling is that it may become very (too) expensive in the end. I know that a quiz game is not eating bandwidth but still it feels to uncertain. However, maybe a native XCODE Swift based Game center & Google integration is better.
I guess for us indie developer this somewhat sums up the discussion!!! …and why a lot of people looking for other solutions.
I wouldn’t worry about investigating new multiplayer pricing models when we got community members so fed up with unet they are fixing it themselves…
Your main competitor has a more complete HLAPI networking system that is optimized for smaller player counts, heck that whole engine is just a fps engine to me. You have an opportunity to provide something a little more unique, and since it is just an API the support would still be in place for those people who just need a few things and a single 1 scene server.
Basically why are you focusing on the hosting when you should be focusing on the network layer behind it? If you want to offer overpriced hosting just partner with one of the terrible game hosting companies like gpanel or something.
Hi,
Is it possible to make the matchmaker server and relay server running in every player’s client?Just like the idea of torrent.So Unity just stand there as a torrent seed,and fee will be low.That will be great!
@DBronson any updates or comments ?
Why actually not just get some real dedicated servers and let us rent them?
Hi!
I am very new to networking so please excuse me. Is it possible to fund these servers with Unity Ads at a $6 CPM? I am thinking of making a ticket system where users can choose to watch ads to get tickets(or buy them using IAPs). I was initially thinking of making 3 games per ticket. (each game runs for 10 minutes and uses data equivalent to that of Minecraft) Is this idea feasable and how much profit can be made out of it? (I suppose 50 Concurrent users)
Guys, so from what I read in other threads the choice is really simple here… Use Unity matchmaker and pay them what they think is fair… Or start a copy of your game as a dedicated master in headless mode on your VPS and have THAT be your master server, thus costing you nothing.
So those who want to use unity’s matchmaker service, they can… and those who want a free solution with no CCU caps or limitations of any kind other than what their hardware is capable of, just build an instance of your game, put it on your own VPS and there you go… all it will cost you is the price of your VPS… which start at about $5 p/month… including an SSL certificate.
Of course, if you are gonna do this then the cheapest package you can find is not the best idea, but it is a good start and all their packages come with free SSL. Again, just to reiterate, if you do this then you are not using Unity’s matchmaking and should not offer that option in your connection options.
Now that that option is on the table this discussion can get back on track again… but it means I have my answer and am out. Good luck to the rest of you 
You also forgot to mention this way you save the Plus or Pro Subscription which is needed to use the multiplayer services !
Ironically enough, I am considering getting Plus now that I have had some experience using it to deliver projects to clients
Note that you don’t need a subscription to use the 20 CCU… you just need to pay (subscription or otherwise) if you want a worth while number of players.
@pKallv I actually realised a stupid thing in that earlier options pro vs cons assessment of mine. That holds true for when you actually make multiplayer games where there are multiple players playing against each other directly… in most of the FTP games that I have played and the ones I intend to make will entice downloads by the masses BUT there is no real multiplayer involved. I.e. I will have 500,000 players connecting to my site and it’s servers to log into their account, to play and store their data and ultimately leave their mark on the leaderboards etc… but all of that will be handled by my own WordPress assets and with the exception of being able to see the scores of multiple players and to see them update with values of other players, these people are never actually playing a multiplayer game…
Even when I release my kit to have “friends” and to have their avatars fight alongside you in battles using their actual gear and stats etc… All of that will be driven by my own WordPress plugins for initial setup at the start of the game and for gear/xp/score updates etc but even then, there is no actual multiplayer going on.
I don’t make games where people play together because of the limitations of the CCU but that doesn’t mean that I can’t create f2p games with all it’s data stored in the cloud. This means that most of my “online f2p games” actually won’t even make use of networking at all.
Once I get my Plus subscription (will do so as soon as they launch a special that does not include random assets I either have no interest in or already own) then I will have the larger 50 CCU limit which is still rather pointless and there running my own server would definitely be a bonus… but there is one more thing that you guys seem to have forgotten… By buying Stan’s Assets iOS Native kit for $25 just to simplify your life; or else just figure this out yourself; you can actually change the “20 CCU” limitation into “20 CCAU”…
Think about it… Apple has GameCenter which comes with built in match making (I think the limit is 4 or 8 players in a match) but offers no CCU limit. This means that as long as you use GameCenter to do the matchmaking for your iOS builds then you only need to worry about 20 Concurrent Android Users via Unity’s matchmaker service. Just thin about that for a second… let’s say you have a 50/50 split between Android and iOS downloads… literally half of your game’s players will not have a CCU limit and the other half will be subject to it… That effectively doubles your chances of not having too many people playing at once…
Still, 20 CCU on Android is still very limiting… so concentrate on the unlimited iOS players instead! Worry about the Android version once you’ve tested the waters on iOS. Not ideal, but it gives you a free way of testing how many users you have and all that analytic stuff and then you will know how much you need to budget for Unity for the Android release… Again, not an exact science but it should give a far better deal that “No more than 20 players, people, thank you for your interest now f-off” or “Oh, let’s just gamble and see how much it costs me, shall we”…
@Rungsted93 and if you do it that way AND use the self hosted server then you also greatly reduce the load on your own server resources. Bit of extra coding in your game, loads of saved resources and far clearer view of what you need to budged for to use Unity multiplayer… (if anything at all)
EDIT: Oh snap! I was just looking at Stan’s product page and noticed his Ultimate Mobile bundle is on sale at the moment so if you buy THAT ( $25 on sale $50 normal price ) you get unlimited iOS and unlimited Android CCU!
Not everyone develops for mobile devices 
In my case it’s a PC game on Steam. However having a dedicated server or VPS is something i just really regret from the start…
So yeah i kinda regret going for UNET to be honest…