How did Supercell pay for its servers at the start of Clash of Clans

Hi everybody

I am passionate gamer of Clash Royale and researched the history of Supercell. I am interested in creating a game like Clash Royale and would be working together with a group of people to reach that goal.

I have read various posts on the topic of financing such a game and conclude that creating a similar game like that would not be a significant problem.

The elephant in the room for me is the server cost. On the internet I found a thumbrule that the server cost for one player/user amounts to 1 cent per day. Now if the game has initial success and you get 200'000 Users a month that amounts to 60'000.-
That is a huge amount to have in cash especially if you follow the approach of supercell to offer the game for free and wait for ingame purchases. Ingame purchases happen after one has paid for the server cost.

I understand that supercell could afford the clash royale server cost because they were already swimming in the money thanks to Clash of clans. But how did they come up with the money for running the servers for clash of clans. I have no idea how much and if their predessor games hayday and boom beach made.

In conclusion I want to ask you guys if it is even possible to create such a game with a dedicated group who does not have the 60'000.- in cash to pay every month or if I would need to create a game that is mainly played offline or without real time multiplayer modus (then goodluck for making such a game still a hit).

What are you thoughts on issue?

Many thanks for your answers. By the way I am new to the forum, so hello unity forum

 

For what type of game?

200'000 a month could be anywhere from 6500 per day to same 200'000 users every day for a full month.
If (best case for server) there are only 6500 users online per day - that's $65 per day of operation. - Only $2'015 a month

Worst case, which you accounted for, would require all 200'000 users to be online whole day every day of the month which would be 2'000$ per day. - $62'000 a month.

The truth is somewhere in between, as it usually is, but I guarantee that, if they allowed such a server population jump then they had capital from previous projects or they were already profiting from the game well in excess of server costs.

 

Also after a quick look at the wiki pages:

"Accel Partners invested $12 million in Supercell in 2011"

"Supercell started developing games for mobile devices. Since then, the company has fully released four mobile games: Hay Day, Clash of Clans, Boom Beach, and Clash Royale, which are freemium games and have been very successful for the company, the first two generating revenue of $2.4 million a day in 2013.[3]"

"Clash of Clans is a freemium mobile strategy video game developed and published by Finnishgame developer Supercell. The game was released for iOS platforms on August 2, 2012, and on Google Play for Android on October 7, 2013."

Of course they had the money from the previous title already.

 

Hi seyn73,

I would suggest that, unless your team already has a huge following or user base from previous releases, obtaining 200k users a month is a little down the road for you yet!

If you begin by using Photon networking, for example, you will get 20 concurrent users for free. Whilst that does not sound like much, it will get you a long way down the road towards an initial working product. If I recall correctly, I think they offer a sliding scale for increasing that user limit. That should allow you to grow organically and within limited budget constraints.

Honestly, I think that creating a quality product is your immediate concern - along with the huge marketing efforts to encourage large scale participation in your game. Worry about the 200k users as and when they materialise.

Consider running some sort of dev blog if you are serious about seeing this large project through to maturity. That may help with raising users' awareness and helping you guage the potential interest that you are generating.

But I admire your enthusiasm and wish you best of luck.

 

Many thanks for your answers. I like where the discussion is going but I wish to dig deeper

Concerning the marketing and creating the quality product, my background allows me to view these milestones as possible to achieve (of course with hard work).

@Doug_B
I thank you for your post and yes 200'000 users is a high number. But what if your game is good and word of mouth publicity with marketing efforts can generate quickly a large following. And then one will have to scale up without having earned a penny yet. In Clash Royale most probably the paying users started paying earliest on in arena 5&6 (and thats very early on, so I guess this would be the minority of the users).

So I guess this brings me to the point of @FMark92 who states that Supercell did earn some money before clash of clans with their games and Accel Partners invested 12 Million before the launch of clash of clans. Many thanks for that answer by the way

So in my case, if I wish to understand and know if it is possible to start a successfull game similar to clash royale (200'000+ users) without a investor? Or if not possible, is there a game type that has a very low need for a server infrastructure?

Did for example Angry Birds have the same issue with server cost?
Is there any possiblity to have the multiplayer part being done with low cost server infrastructure? For example the software is so secure that the server is one of the smartphones itself (if two player compete against one another)?
Or does it need to be a kindof "offline" game? In that case it would be difficult to market, because I think a lot of people wish to play against the human rather than the AI, then again there was candy crush saga....

What are your thoughts on that?

 

Of course it's possible. But I would be reasonably skeptical about my ability to do it.

If we're talking minimum communication then I guess the games where the only communication with the server includes sending new high scores.

I do not know much about AB but from what I've seen it's mainly a single player game, no?

 

I think you might be ignoring that clones of Supercell games are all over the app stores already, mobile users are used to them to the point where even games that just happen to look a bit like one their games are just label "CoC clone" or whatever.

The reason Supercell has managed to build the business they have now is that (ignoring that CoC was itself a clone of another company's facebook game) they innovate and come up with new styles of play for their new games.

 

Hey @LeftyRighty I wish to do a game like clash royale but its nothing like clash royale. Its tower defense but in a new way. Not comparable to the tones of games out there that really play like clash royale

That I have taken care of.

 

Neither clash of clans nor clash royale seem to have any physics involved and you can't directly control the movement of units, meaning that the server can easily calculate the outcome and you don't event have to send many messages over the network.

I'd estimate that you can start by hosting your game server on a cheap (~$10 per month) server.
It should be able to handle a big number of concurrent players, if you have an experienced programmer.

And if you happen to actually get a huge number of players you can either upgrade your server plan or get multiple servers, which is generally better but requires more programming work.

On the other side you have services like Unity networking and Photon that will handle most of the server side stuff for you, but if you use their servers then you are very limited by what you can do on the server side and they tend to cost significantly more than hosting your own servers.

 

Hey everybody and @zoran404

I do not think that the simple way of purchasing lowcost server or working with photon for example and then upgrading as soon as there are many people wanting to play the game is the way to go. Most likely you would need to upgrade to a very expensive plan with a lot of users wanting to play or you would need to purchase a lot of servers. With the money being earned primarly with ingame purchases this will not be dowable unless some angel investor sees the potential.

The worst message you can send to users is a game that is lagging, that could kill off even a brilliant game. And I wish to not take that risk (unless I would have an investor somehow).

So I would want to go back to the inital thought creating a excellent game without the need of an investor.
This would mean if I understand a game that involves minimalistic server communication.

So an angry bird for example that just sends the high scores to the server, is that right?

But what about a Clash Royale kind of game where the server is on the smartphone of player 1 or 2 (that play against each other) and not on the server offered by the game developer. I guess its prone to cheating but if programmed in a smart way, cheating could be prevented, right? Would that be possible?

Or are there some other game types that could work and involve minimal server communication?

Many thanks for your answers, you guys help me a lot here

 

Cloud computing services can rapidly scale themselves to meet demand. You don't have to start with an expensive plan or an expensive array of servers. You simply need to choose the services that will scale on demand. Photon, for example, has both fixed and scaling plans. Fixed plans are typically used for alpha/beta development cycles while scaling is for release.

It's entirely dependent on the game and how much you wish to go hands on with the solution you choose. Services such as GameSparks will handle the bulk of the server-side details for you allowing you to focus on your game but you'll have to pay more for these services. At the same time though choosing a more hands on approach may cost you more than these as you will have to build and maintain more of the implementation yourself.

 

The big costs in mobile is user acquisition not server costs. Because companies in this game have gotten really good at their analytics and getting the most out of each user. So it's a volume game which also means a high cost to entry. To get anywhere close to your 60k server cost would require several times that in user acquisition costs, as a general rule.

There are always exceptions, but if you look closely at them you will find reasons for them. Like the people involved have a lot of industry experience. Or the game has something new or novel (ie not basically a clone). Or something that set them apart.

Also being new at this, development is going to be a bigger challenge then you think.

Not trying to discourage you, it's just that any mobile startup now days has a rather hard time and is generally fairly well funded to have a chance. Just the nature of the beast. If you have something new to offer that helps. If you are just out to make a clone basically, you won't even get seed funding.

 

I think you went a little too minimalistic here. If you're only sending the score then this isn't a multiplayer game anymore. Besides people can send fake scores and which you can't validate.

If you want a minimalistic multiplayer functionality you can do what turn based games do: only send the message that the player has made a move.

Even if the server-side logic is on the player's machine you'll still need a server to send messages between them, which is what most of the networking services do.
And no, there is no way to prevent cheating, which is why this is mostly used when the player can only play against friends.

I think you are overestimating your costs.
You should consult someone who has worked on similar games before you make a decision.

 

The original Clash of Clans battle mode was a semi-passive mode. Only the attacker did anything, and the defender didn't even need to be online. This sort of typical mobile MP isn't that much harder on a server than matchmaking and chat is, and a modern server (individual computer) could handle 300k to a million total users (going by what the developers of Photon and similar apps have claimed).

The number of concurrent users depends a bit on bandwidth and style of game, but for this style there are only a few tiny information exchanges every few seconds to update what the client displays, and probably very simple calculations on the server side to determine battle outcomes.

The new battle mode is just two players syncing up and watching one each of those battles against each other, so still not terribly active. They just do their best attacking the other player for a final score to compare. I have no idea how this specific game does it, but I've tested many similar games, and the basic outline for networking seems to be the same.

So tl;dr: A few thousand users, which you won't have initially, are not going to break the bank with AWS or similar. After reaching hundreds of thousands you might find it cheaper to rent a few servers and handle them yourself. But whether you're running on a typical telescopic cloud service or renting whole servers directly, you're still going to need some expertise in making and running the server software, since you want it to be the authority.