As an operations guy I find this really interesting. What I find interesting is the cost vs revenue of these games.
Obviously to me there is a bunch of up front costs, art assets, game play design, etc. But once you've got the basic HTML5 code going, and the art assets in place, what are your on-going costs?
Clearly there is going to be bandwidth at scale its like $2 - $3 per megabit per month, you've got power/cooling in a colo, so that's going to cost you maybe $3K/month for a 42U rack's worth of servers. And presumably you've got a couple of opsen types tending and feeding that rack. But they are split across say 100 racks so 1/100th of their salary/benefits per month, lets call it $160/month [1]. So lets give the game 500Mbits of bandwidth a month (that is 4.3TB of data transfer per day) We can put 20 beefy servers in our 42U rack with 20TB of disk space and 1/2TB of flash in each (we do this at Blekko)So 400TB of disk space and 10TB of flash, and if you wanted 3.8TB of RAM (192GB/server) and 480 threads of execution.
Now I'm making a huge assumption here, and so I'm calling it out, but my assumption is that the "game" is essentially a front end on a database, where player actions become database transactions, and those transactions show up as changes in the game 'world', further those transactions are both shared (everyone in the game can see them) and private (only you see the changes). So a system that is slightly less beefy than our setup (the Sun setup) gets 10M transactions/minute in the TPC-C benchmark [2]. At that level it suggest our stack-o-servers might support 1 - 2M players.
So our monthly cost is on the order of $6,500 and we're hosting say 1.5M players. That means we need to extract something like a half cent per month out of those players in order to pay for running the hardware. For a gross margin of 50% that is about .86 cents.
I may be completely off here somewhere, please point out my math or configuration errors! It seems like one could essentially run a game 'forever' until that average number got too low and then you would have to wind it down, but one wonders about how hard it is get get someone to pay a penny a month.
I did leave out the maintenance (warranty) cost of the servers, a recent quote from a server vendor had 'everything covered, onsite tech' for $175/server/month. So if you wanted to add that in then there would be another $3,500/month to cover.
It looks like a license to print money but clearly there is something here that sucks all that money away. Curious what it is.
[1] That is two ops people with a combined salary/benefit package costing $200K/year or 16K/month which divided by 100 (1/100th of their work) is $160.
I've worked for one of Zynga's competitors, and you are omitting one significant cost. It is vital for these games to add content monthly if not weekly, and these games are released with a large number of known bugs that are slowly fixed through the games lifetime. I'd say that any of these games, once released has at least two programmers and three artists working almost exclusively on bug fixing and adding features. Last game I worked on we had three programmers working on the client full time on it, two programmers doing occasional work on the backend and three artists creating assets for at least five months after release. I quit about this time, but my guess is this expenses went on for at least a year.
Wouldn't cutting off support and updates for the game be a better alternative than shutting it down though? If someone is hooked on the game, they won't care much for new updates.
If they get bored, they find a new game to play. As long as the game is at least profitable, run it until it isn't.
The only other reason for closing a game prematurely like this would be to minimize people complaining about the lack of support for a game, which could translate to "Zynga sucks, they haven't added anything in months, clearly they've stopped caring about their customers!" by players who either forgot about or didn't get the memo.
> If someone is hooked on the game, they won't care much for new updates.
Eh, no. It's specifically the new updates that people get hooked on. The new updates are what drive people to do the microtransactions. Without such updates, I'd expect income to drop by a few orders of magnitude: well below profitability.
I think what is being said is that the lack of support may mean the game would die in 1 year due to becoming unprofitable but why lose 1 year of profits?
No, that's not what I'm saying at all. I haven't been able to come up with a good summary; I'd recommend playing any Zynga game for a bit (about 60 minutes of actual play time, I'd say; make sure you look at the "quests" or whatever they call them and the currencies involved) and figuring out their per-game business model. You won't have to pay anything; you'll risk cookies and stuff, but if you ask I'm sure someone can suggest ways to keep clean.
If you've done that and still don't understand, I can try to explain further. It's not about support. It's about regularly asking for money. The moment you stop asking is the moment you lose revenue. Think of it like church. Every Sunday, they pass the collection plate. If they don't pass the plate, they don't get paid.
These games have to have some type of limited-time element. You have 2 weeks to get the special "one-time only" pink zebra and if you don't earn it in two weeks you'll never be able to get the pink zebra again.
Or you can just buy some credits to buy the zebra for $20 in USD.
These games require limited-time only things to make money. Every holiday (valentines, easter, 4th of july) is a big event. The cost of the game is not servers, it's 5-10 full time staff to run it.
Many mmo's don't die, they just contract to a single server of active players that are profitable long after everyone else has forgotten about the game and development and bug fixes have stopped.
There seem to be a similar tiny minority of fans that could keep these Zynga games barely alive for much longer, so why isn't Zynga doing this? My guess is they view hardcore customers as suckers and would rather have them spending $20 a month on a new title with new content then $1 a month keeping a near dead title just over break even. Carrying the example math forward even if they lose 95% of these hardcores it's going to be worth it if they can move 5% of them to a new release.
MMOs are able to continue generating content without developer input via PvP, or by allowing players to create new characters and run the grind again, or by allowing players to talk to one another inside the game context.
That's not content that's activity and Zynga games have the same ability to run the grind ad nauseum, interact with other users, etc. If Joe Customer wants to log in and keep feeding 25 cent cookies to his Zynga dog every morning he could keep doing that forever.
> Zynga games have the same ability to run the grind ad nauseum, interact with other users, etc.
No, they don't. In an MMO, if you get to level 80 as a cleric, you can go back and redo all the zones as a warrior. In a Zynga game, this isn't available: it's what MUD designers called single ladder advancement. You get to level 80 and that's it: you can only keep buying more stuff, or only maintain what you have. That's very different. There's no reason to make a new Facebook account and start over.
In an MMO, there are a plethora of interactions, mostly related to group activities. In a Zynga game, interactions are pushed into status updates. To bring in another analogy, an MMO is like a bar. You can go there, sit down, and chat with people. You can buy a beer if you want, but you don't have to. A Zynga game, however, is a beer. You can talk with people over said beer, but that talking isn't an intrinsic part of the experience of the beer. And at the bottom of the cup, it says, "Tell your friends how great this beer was to get more beer!"
> If Joe Customer wants to log in and keep feeding 25 cent cookies to his Zynga dog every morning he could keep doing that forever.
Yes. But are there enough of him to justify keeping the servers running?
In MMOs, the answers tends towards yes, because the interesting activities tend to be group-friendly, which naturally pushes up the number of people playing and paying. You can't say the same about Zynga games.
Ok, so you're saying I need to measure this stack against another game that I could ship on this stack. So leaving it generating revenue $X is a loss of additional revenue $Y-$X if I deployed a game that had $Y revenue on the same hardware. Does that about cover it?
The assumption in that though is that game $Y cannibalizes $X. Which is to say if you also release a new game $Y your returns on $X go away.
And while I see that as true with similar/derivative games "CashVille 2" is just like "CashVille 1" but with unicorns or something, I am not sure I'm persuaded that we do accrue an opportunity cost against different game. This is the root of the reasoning in my print money comment.
The reasoning is that if this config pays for the development cost and then generates marginal $ over time, you just add another stack to increase the marginal dollars you get. A writer releasing a second book doesn't 'cut into' sales of the first book, if nothing else they can sometimes increase the revenue of the first book by getting wider visibility.
I'm interested though in how we might reason about the opportunity cost here.
One aspect you aren't twigging on is that the monetization strategy is tied to the game design. Even if every game, on average, took in the same amount of money once it hits critical mass, players aren't going to invest in games that their friends aren't playing, so splitting the audience up hurts Zynga. Their growth strategy was one of complete content saturation using fast-follow tactics. "If they aren't playing Zynga game X, then we want to make sure the other game they will be playing will be Zynga game Y instead." This worked during the explosive growth period, but it's very wasteful now growth is slowing or even shrinking (I'd need to look at the numbers more closely).
As was also mentioned elsewhere, players are always hungry for new items, and my guess is that Zynga has plenty of metrics to indicate that new items is where the money goes (a hefty dose of behavioral economics like scarcity is used on those ones). Simply leaving a game open, or closing it to new players, is probably leaving money on the table. That said, I think there's a big risk in teaching players that their virtual items really do just disappear into the ether.
My guess is it's the right move for Zynga to make, but Indiana Jones and Mafia Wars 2 is a huge blow to them. They were flagship products, and both of them are only a year or so old.
Unlike books, I think the market for these games is inelastic. A player is going to devote four hours per day to some game, chosen from the available games. You want that to be your most profitable game. Allowing them to play other less profitable games is not optimal.
If you somehow knew I was going to buy exactly one book, you'd do your best to make it a hardcover and pull the paperback off the shelf.
So far, for me, this is the most compelling reasoning. In theory to incur an opportunity cost you have to be constrained in some way, clearly Zynga isn't constrained in adding hardware or people (they are laying off in this case) but I can see how the customer's attention span is a constraint. If they have a fixed amount of game time then you're competing for your place in that time slot.
The problem with your explanation is that the player is going to devote AT MOST 4 hours to SOME game, yours or otherwise. There's a risk that those who are leaving the existing game may jump to another game from another company. Reevaluation moments are dangerous.
Actually I'm saying that even the most ardent player of these sorts of games only plays a few of them at any one time. If Farmville is capturing $n net per player, but a new game is capturing $n x 1.25 net per player, your better off killing Farmville, if you think enough Farmville players would then switch. I don't know if this is the case, but that is what I meant by "opportunity cost."
Zynga had copied ideas from competitors, wasn't Cityville a Sims rip-off? Not only that but competitors were ripping off Zynga ideas as well.
I understand the need to constantly change the game to attract new customers and sell new items to the existing customers. I had played Playdom's Marvel Avenger's Alliance, but recently quit because the bugs were very serious and each new thing they added created problems in the PvP mode as well as more bugs. In order to be competitive in that game you need to buy gold to buy rare items on sale for a limited time. My problem was that they have a "Maintenance"/Refresh bug and after I bought gold and paid $20 that I was not given the gold by the game. I filed a tech support ticket with Playdom over it and gave them my Facebook account name and Paypal address. They just ignored it for weeks. Eventually I had to file a complaint with Paypal for not receiving goods that I paid for and was awarded $20 in my dispute with Playdom. Then a few weeks later Playdom had my Facebook account banned claiming "unauthorized credit card use" but that was not the case, it was failure to deliver goods paid for. After trying to resolve the issue with Facebook and Playdom, I eventually quit and removed the game from my profile. I refuse to play any more Playdom or any of its competitors because of bugs so serious that you don't get what you paid for, and they refuse to give you what you paid for or refund your money. I heard from other players that Zynga does that as well. Basically it comes down to this, if they ignore the request for a refund for 30 or 60 days, Paypal or the credit card/debit card company cannot issue a refund after a certain amount of time after the transaction.
I have never in my life seen such buggy software as these so called online social network games. I have never seen such sh*tty tech support as these online social networking games. I'll bet that this is what affects their profit margin, once users get burned in buying premium content they will either play the game without buying anything else, or they will like me quit playing from that company or all companies.
Nope I spent hundreds before that, and was a customer for life. Playing for almost a year.
Any company with shtty tech support and pss poor customer service is going to lose customers by making them unhappy and then spiral their stock out of control into a bankruptcy. That includes Facebook themselves, who even after proof that I didn't get gold in the game for my $20, sided with Playdom in banning my account for a day for "unauthorized credit card use" of my own Paypal account.
Here is another thing, Paydom's TOS says I must file a support ticket with them, I did and nothing happened and they ignored the problem, Paypal's TOS says I have to resolve disputes with them which I did by creating a ticket, and then Facebook got upset because their TOS says I have to resolve disputes with them. Which one gets priority? Playdom and Facebook refused to do anything and then banned my account anyway even if I did nothing wrong and I was the one who was punished for a bug in the game that doesn't give me gold I legally paid for with my own account. That is all kind of messed up there and if this is what happens to a typical user who needs a refund due to a bug in the game not giving gold or whatever they paid for, then this whole industry is going bankrupt into another Dotcom bubble burst. You never do this to your customers, never!
I had to buy gold to buy things in the game I needed. Like ISO8 Goo to do special missions to win a special limited edition super hero, buy better weapons, or ISO8 gems to make my super heroes better.
It is hard to explain, but in order to play the game better you need to keep buying stuff. Zynga games are like that as well. If you don't buy stuff then you suffer in the game and other players can beat you up better.
thanks for your reply!
Do you find that you get more enjoyment out of these types of games rather than on or offline games where you pay up front without "pay-2-play"?
It's quite an assumption to think that after playing petville for 2 years and seeing the plug pulled, these people will be flocking to other Zynga games.
Games have a maintenance burden which you haven't factored into the equation. This maintenance requires effort, time and money. Sounds like Zynga was stretched too thin, and had bought too many companies without really working out what the ongoing costs would be.
There is a fine line between maintaining the current player-base and attracting new players, and the money in both is fast and hard to come by (however the latter is more exciting).
It would be an unfortunate charity to give little to no attention to attracting new players but instead letting the current player-base stagnate, and it's troublesome to keep assigning hard-working developers and artists to maintain the slow and steady decline of a game.
Make no mistake; killing off the stagnating games is in the best interest of the artists and game developers in the same way that it is financially sane.
Sounds like you have some better first hand knowledge here, tell me is there no situation in which you just 'run' the game and artists and game developers are no longer involved?
This is a critical point for me. I was a huge World of Warcraft fan, played a lot, and would still play a lot if it wasn't a complete waste of time :-) I also play 'Settlers of Catan.' (SoC) I've played SoC since it came out, and while it has had a couple of expansions, I don't play them, I still play the original. Once I got the box and tiles and cards there was no additional participation of any game designers or artists. The game continues to grab followers, the marginal cost of shipping a new one is some injection molded plastic and printed cardboard.
They are doing a great job at maintaining a balance between attracting new players and maintaining existing players, particularly in the recent Cataclysm expansion and Mists of Pandaria expansion.
The Cataclysm expansion was massive in that it revamped the existing world in addition to adding new content and I believe this is highly undervalued by existing players. Furthermore they have applied the same attention to detail to an entirely new expansion (MoP), and the gold-trim is really starting to shine through -- especially with regards to how well received the pet battles are (appealing to new and returning players) and how well received the challenge modes are (appealing to existing players).
Since Cataclysm was such a huge under-taking it took an astronomical amount of time and effort and they have adopted a policy of more regular intervals between expansions and patches, and in that regard Cataclysm was an opportunity to stream-line their entire workflow (from art design to player-facing content) -- an impressive feat in itself for which they are currently reaping the benefits.
The amount of over-head involved in making making new content and engaging consumers is expensive in all fields from offline single-player games to online multi-player games and from movies to music, and Zynga's shotgun approach to new games has been expensive and it's starting to hurt them.
Blizzard is a complete mystery. I paid for a copy of Starcraft: Brood War when it first came out (1998), and I've played on Battle.net using my original serial on and off ever since. Their servers are still up and running silky smooth, 15 years later.
What is even MORE amazing is that I was in Best Buy four days ago, and I saw not one, but TWO copies of the Starcraft Battle Chest (Starcraft, the expansion, and strategy guides for both) still selling at $30.
I first bought my battlechest in 2000 for $30.
So 14 years after the release of the original game, it still moves enough copies for Best Buy to have it on the shelves.
Starcraft has player numbers that would make a two-year-old game envious. I have no doubt Blizzard will kill it off at roughly the same player count as other companies with similar games, but for now it's well worth them keeping it running.
Zynga's taking advantage of human psychology to abuse their customers who have been playing their games for a long time. I didn't realize this until working this out while responding to your post:
By killing off the oldest games, Zynga causes a significant segment of its customer base to substitute newer games for the older ones. These customers are probably the most profitable (they likely buy stuff in game), and they will likely make purchases in the new games as well.
This has another benefit for Zynga in that it lets them play games with the user metrics for their newest titles (thus making the stock holders happy). Having a long-time customer start playing a new/trending game is just as good as (or better than, see above) a brand new customer.
Was that last question ironic? Aren't you describing a traditional 'boxed'/retail PC/console game? That's something that is well designed and self contained that you buy once and can enjoy forever (if you have some friends to play it with).
(IMO, in the context of Zynga and similar, the idea of pay-once-play-forever games is off topic in this discussion!)
Obviously to me there is a bunch of up front costs, art assets, game play design, etc. But once you've got the basic HTML5 code going, and the art assets in place, what are your on-going costs?
Clearly there is going to be bandwidth at scale its like $2 - $3 per megabit per month, you've got power/cooling in a colo, so that's going to cost you maybe $3K/month for a 42U rack's worth of servers. And presumably you've got a couple of opsen types tending and feeding that rack. But they are split across say 100 racks so 1/100th of their salary/benefits per month, lets call it $160/month [1]. So lets give the game 500Mbits of bandwidth a month (that is 4.3TB of data transfer per day) We can put 20 beefy servers in our 42U rack with 20TB of disk space and 1/2TB of flash in each (we do this at Blekko)So 400TB of disk space and 10TB of flash, and if you wanted 3.8TB of RAM (192GB/server) and 480 threads of execution.
Now I'm making a huge assumption here, and so I'm calling it out, but my assumption is that the "game" is essentially a front end on a database, where player actions become database transactions, and those transactions show up as changes in the game 'world', further those transactions are both shared (everyone in the game can see them) and private (only you see the changes). So a system that is slightly less beefy than our setup (the Sun setup) gets 10M transactions/minute in the TPC-C benchmark [2]. At that level it suggest our stack-o-servers might support 1 - 2M players.
So our monthly cost is on the order of $6,500 and we're hosting say 1.5M players. That means we need to extract something like a half cent per month out of those players in order to pay for running the hardware. For a gross margin of 50% that is about .86 cents.
I may be completely off here somewhere, please point out my math or configuration errors! It seems like one could essentially run a game 'forever' until that average number got too low and then you would have to wind it down, but one wonders about how hard it is get get someone to pay a penny a month.
I did leave out the maintenance (warranty) cost of the servers, a recent quote from a server vendor had 'everything covered, onsite tech' for $175/server/month. So if you wanted to add that in then there would be another $3,500/month to cover.
It looks like a license to print money but clearly there is something here that sucks all that money away. Curious what it is.
[1] That is two ops people with a combined salary/benefit package costing $200K/year or 16K/month which divided by 100 (1/100th of their work) is $160.
[2] http://www.tpc.org/tpcc/results/tpcc_perf_results.asp?result...