Psychochild's Blog

A developer's musings on game development and writing.

16 April, 2007

How low can you go?
Filed under: — Psychochild @ 11:24 PM

I saw an interesting post over on MOGBlog entitled, Wanted: Two Million Dollars. It’s an interesting post about what it would take to make a small project, and the conclusion is that you could do it fairly inexpensively. If you could fit the project into 2 years, the post says, you could do it for under $2M.

Note: I’m not saying that you cannot make a game cheaper than this. Rather, I’m pointing out some of the missing expenses from the post I linked above. You can make a game for much cheaper, and several people have. But, if you want to hire 14 people and pay them the salaries listed, you will end up spending more even if you don’t hire anyone else. A true “indie” game will have a few underpaid and overworked developers that are doing multiple jobs simultaneously.

So, as much as I like such a can-do attitude, I have to show some gentle corrections in the figures.

First, a snark. a game is still “massive” if you don’t have millions of players. The original definition usually indicates that the number was supposed to be in the triple digits. That means if you attract over 100 players, you qualify as “massive” enough. You still have to worry about issues like customer service, maintaining servers, etc, whether you’re dealing with 1,000 players or 1,000,000 players.

Now, I’m going to consider this from a U.S.-centric position. Your milage may vary in other countries, or even other states. :)

$300k is probably enough to handle most of the other other costs. We spend about $200/month for website, bandwidth, and co-location after buying a used 1U server for Meridian 59. A project like this will probably spend a bit more than we do, but if you’re making a modest-sized game you can probably get away with about $1000/month in hosting during development and closed testing. You’ll probably need one of your programmer-types to also do IT support for your servers, and expect the other people to be smart enough to fix their own computers if things break. And, bandwidth during launch can get pricey if you don’t want to have the same experience as every other MMO on launch day.

For work hardware, you don’t need cutting-edge machines, so get some reasonable machines in bulk. If you’re smart, you might be able to get a good discount (or even free equipment) if you agree to put some logos on your website or in the game. You don’t see the “NVIDIA: The way it’s meant to be played!” screens in games because the developers think NVIDIA needs free advertising! ;) But, it might be hard to get serious attention from these people if you’re a small-scale developer.

If I were planning this size of a project, I wouldn’t have a tech lead. I’d hire 3 competent, self-motivated programmers: a client programmer, a server programmer, and a network/tools programmer. I’d split up your budget between them and give them $80k each. This will be a lot of work per person, though, especially if you don’t license any engines. Also, this means you can’t expect a cutting-edge 3D client with an infinitely scalable server technology.

For the artists, you might want to consider more outsourcing. You can get a fair amount of art out of a Chinese (or Eastern European) outsourcing company. You can’t eliminate the whole art department, because you’ll still need artists on your side to ensure the deliverables are what you need and to fix up the inevitable problems that would take too long to have the outsourced company fix.

For designers, I’d probably look for people that could do light programming/scripting and level layout. This might raise the prices a bit, depending on who you can find. Remember that you’re going to be doing a lot of content, which could be a sizable amount of work for just 5 people. In addition, be careful with Junior Designer! They may come cheaper, but they will require more training and investment; this could overwork your full-time designers in this type of project. But, you might find a brilliant designer in the rough.

You’re probably overpaying your leads for this type of project. I think the art and design leads would be happy with $80k given the relative salaries. Making double seems a bit excessive for an indie project. Salary figures will be shared eventually, and paying the leads so much more might create a bit of resentment with the other people. And, as I mentioned with the programmers, you’re probably better off not having a lead and just having self-motivated workers that are in charge of a specific area.

But, here comes the bad news: You have a lot more costs to consider.

This list is missing a QA team. You’ll want at least one internal tester to do quick turnaround testing on your stuff. Let’s face it, we tech guys sometimes suck at testing our own stuff. Some days we’re happy when it just compiles. :) If you’re working with a publisher, you can get some more testing resources from them, but if you’re staying lean and mean you have to figure on paying for more people. Especially when it comes to compatibility testing! With Vista out but some people clinging stubbornly to XP (and even some 2k holdouts!), you have to make sure your game runs on a variety of configurations. Your programmers will not have the time to do this.

You might want to hire on someone to focus on the business side of things full-time. It can be a lot of work dealing with tax paperwork, getting feedback from the lawyer, etc. Ideally, this person would also have project management experience to keep things running on schedule and on budget. The could also fill in other areas as needed, such as rounding up external testers, proofreading text, etc., during the lulls in their work. Figure another $80k-100k salary if you want quality here, and you do. ;) The challenge is to find someone competent that won’t use their position as “the biz guy” to try to dictate design. If you plan on filling this role, don’t expect to be able to lead the design, too, otherwise you’ll get overworked and something will get ignored.

When paying all these people, you need to handle payroll taxes, which means you’ll need to add about 10-20% on top of the salaries for the company’s obligations. Want to offer benefits like health insurance? That’ll be additional costs. The rule of thumb is that you’ll add another 50% for taxes and benefits on top of salary. So, the base rate of $850k/year burn in the post now becomes about $1.25M if you want benefits.

You’ll need an accountant and lawyer to handle some of the basics. If you don’t want to overwork someone with bookkeeping, that’ll be a significant cost to have a bookkeeper. You’ll want a lawyer to at least look over your EULA, etc, and review any major contracts, such as hiring agreements, founder’s agreements, etc. You will also be served by filing some simple IP protection, such as trademarks on your important properties. And, if you’re not renting offices and have employees in different states, you’ll want a payroll company to avoid the headaches of too much paperwork. All this stuff is not super-expensive, but it quickly adds up.

Finally, this includes no marketing/PR budget. This doesn’t have to be a lot, but you should still get someone to do the basics. The cheapest thing to do is to contract a good PR person to help you write and send out press releases to announce your game and organize press contacts to get the word out about your game. There’s a great chapter on PR in a certain book I’m familiar with. ;) Your biz person might be able to pick up some of the slack on this side, too, by organizing the screenshots to send out, etc., if they aren’t overworked already.

Depressingly, I’ve pushed the cost of this project way beyond the $2M goal. You’re probably looking at least at $3M, if not $3.5M for this type of project. You might be able to cut some of the cost if you lower salaries but offer a well-defined profit-sharing program for your employees. I know I would be willing to work for less if I could get some percentage points on the back end. But, you have to figure this into your operating costs.

Unfortunately, it’s easy to overlook these costs. The biggest surprise is usually when people forget the payroll taxes. If you’ve never run a company before and had to do payroll, it’s very easy to overlook.

Hopefully this helps people avoid getting into a mess. :)

Edit: I adjusted the tax rate figures above. My tax rate estimates were way too high. Also added the note in the second paragraph above since people seem to be think that I am advocating that this is the cheapest you can possibly make a game for.


  1. Wow. 30% payroll taxes. Canada’s payroll taxes look a little more favourable. The employer’s contributions are: CPP (Canada Pension Plan) 4.95% to a max of roughly $2000, EI (Employment Insurance) 2.52% to a max of roughly $1000, and EHT (Health tax) 1.95% of total wages in excess of $400,000.

    Using the MogBlog example of 14 people and 850,000K wages, that works out to something like 45K in payroll taxes or about 5-6% of the 850K in wages.

    I suspect (but I don’t actually know) that benefit rates might be a little less expensive in Canada (due to the socialized health care).

    Personal taxes are reputed to be higher in Canada, but there are LOTS of Canadians who are a:\ accustomed to the higher personal taxes and b:\ would dearly love another game development studio …

    I recommend something in the vicinity of Toronto, Ontario. Further, I recommend you hire me! ;)

    A summary of Ontario payroll taxes here:

    Comment by Tuebit — 17 April, 2007 @ 7:38 AM

  2. Sorry. More like $50K in payroll taxes on the $850 in annual wages. I guess with math skill like that, I wouldn’t be top on the list of potential hires.

    Comment by Tuebit — 17 April, 2007 @ 7:46 AM

  3. MMO Production Costs

    [...] Brian Green has a post up asking “How Low Can You Go?” talking about the minimum amount it takes to make an MMORPG. He’s commenting on a post on a very new blog that claims a new MMORPG (that isn’t trying to compete with the giants) could be made for $2 million and take 2 years doing it. Brian raises the stakes and claims it would take $3-$3.5 million, though he allows that some savings could be made by trading equity for salary. [...]

    Pingback by The Forge — 17 April, 2007 @ 2:17 PM

  4. Especially when it comes to compatibility testing! With Vista out but some people clinging stubbornly to XP (and even some 2k holdouts!), you have to make sure your game runs on a variety of configurations. Your programmers will not have the time to do this.

    Although it’s fun to bash Vista, I’ve found that the biggest problem with Vista is that no-one else’s software/drivers work on it. My own software seems to work no problem.

    The bigger/biggest problem is graphics cards. To do them right, you need a large computer lab with 50-100 computers, each with a different graphics card (and driver).

    One (partial) solution to this problem is to give each member of your team a very different machine, with different amounts of RAM, CPUs, OS, mice/graphics tablets, types of graphics cards, and size/aspect of monitors. (I’ve played too many games that assume my monitor is 1024×768!)

    In other words, the WORST thing you can do is go to Circuit City and say, I want five HP 5354 desktops for my five new employees. Such action guarantees that your software will only work on the HP 5354 destkop.

    Comment by Mike Rozak — 17 April, 2007 @ 3:46 PM

  5. Tuebit wrote:
    Canada’s payroll taxes look a little more favourable.

    Actually, I engaged in a bit of bad math about taxes by incorrectly using a rule of thumb figure. The actual tax figures for one month in 2005 for Near Death Studios, Inc. was 11.35% for employer taxes, including federal and two states. The rule of thumb is that taxes and benefits will add about 50% more to your salaries. This is reasonable, I think, because if you’re paying $80k-$100k then your employees will probably want benefits. I’ve fixed this mistake above. If someone wants more exactingly accurate figures, we can talk consulting rates for writing your business plan. ;)

    Also, to address a few points from the trackback from The Forge above: I’m not saying that this is some magical minimum amount you have to pay, rather that it’s easy to underestimate your costs. Even an 11% difference due to payroll taxes can throw off your estimates significantly. You can make a game for very little money if you’re smart about your spending. But, don’t expect to be hiring a group of 15 employees and paying them great salaries and expect to keep costs low.

    Also, be careful in trying to avoid payroll taxes by having contractors instead of employees. The government hates that, and an audit will likely end in crying for you if you try to do this. Plus, the payroll taxes just get shifted to the employee instead in most cases in the form of “self-employment taxes”. So, in the end, you’re just making the employee pay them instead of the company, effectively lowering their salaries.

    Mike Rozak wrote:
    Although it’s fun to bash Vista, I’ve found that the biggest problem with Vista is that no-one else’s software/drivers work on it.

    I didn’t specifically bash Vista (although I’m not touching it, personally). Rather, I pointed out that this is going to confound your compatibility testing. As you point out, drivers are much more hit-and-miss on Vista, causing additional problems.

    In other words, the WORST thing you can do is go to Circuit City and say, I want five HP 5354 desktops for my five new employees. Such action guarantees that your software will only work on the HP 5354 desktop.

    On the other hand, this will drive up the costs for your machines. You can often get a discount by buying in bulk, as I said. And, it can confound the issue if you need to fix machines because you have to buy such a variety of parts. Finally, I don’t think your employees will be happy about this situation; who wants to be the poor sucker running 600×480 on 13″ CRT monitor? ;)

    Some more thoughts,

    Comment by Psychochild — 17 April, 2007 @ 5:46 PM

  6. Machine and driver testing, to a large degree, should be what your alpha and early beta stages are for. It really would cost a ton to try and test a wide variety of machines… just have people apply and include their dxdiag. Choose alpha and beta applicants both in terms of experience with beta and based on their hardware. Or heck, games have open stress tests… why not an open tech test?

    Comment by Jason — 17 April, 2007 @ 8:26 PM

  7. Public compatability testing – While public hardware/driver compatability is necessary, it’s better to catch the problems before the public sees them because (a) most users who encounter these problems won’t be able to properly tell you what’s happening (aka: You’ll get “The graphics sux!” and not “Your app has selected the wrong aspect ratio and is rendering using 8-bit color instead of 24-bit on my random graphics card XYZ version number N”, (b) Unfortunately, even if they can tell you, you can’t do much without actually buying the hardware, which is either expensive or impossible to find (in the case of random graphics/audio hardware).

    Variety of machines – When I worked on speech recognition in Microsoft, every developer had a different machine (which undoubtedly costs more). These included differnt processors (DEC alpha, MIPs – no longer around) and multiprocessor machines. It worked really well from a test standpoint; many compatability problems were quickly discovered. In my own game, when I found a few bugs when I bought my dual-core computer. I found different bugs when I installed my widescreen monitor, and when I added a second monitor.

    So, the moral of the story is: Not only should one of your development computers be Vista, but it should have the 64-bit version of Vista installed. You should definitely be testing on a dual core machine, and probably a quad-core.

    Random comment about Linux – If testing on several flavors of Windows weren’t bad enough, a team is developing a windows-compatability layer for Linux, called Wine. I didn’t know about this until some people tried my app under Wine and reported some bugs to me.

    Oh, and don’t forget Windows running on the Mac…

    Comment by Mike Rozak — 17 April, 2007 @ 8:59 PM

  8. It’s interesting the variation in salaries you can get just state-by-state. My small IVR/computer-telephony company here in Minnesota, we generally take only about $60K a year. Bonuses and the like do bump that up a bit (a few K), a certain people have commissions, etc., but base wages for our 12 employees is only about $700K total.

    Of course, a fair number of us (all the devs, for example) do have ownership as well, which tilts the playing field. However, it should also be pointed out that “ownership” in any company not making money is usually far more of a burden than a blessing. (Just take my word on this.)

    Example: at one point (during the tech crunch), we owners took half-pay (or none) because of extended cash flow problems and dipping too far into our reserves for a project… luckily, I live cheap. You are also regularly providing personal financials to bankers and the like on behalf of the company if you’re pursuing any sort of leasing or loans, and and signing personal guarantees to pay off the company’s debts if defaulted. Just remember, especially in a small company, ownership implies responsibility as well as privilege.

    /sermon (sorry)

    Good point on contractors. They can be lifesavers, but be sure that they are self-directed, have their own facilities, etc. If they act like an employee, they’ll probably be considered one, and you’ll have tax authorities knocking on the door eventually.

    Especially if you are in a lower wage market, don’t forget potential overtime provisions either, both federal and state. Most of the positions being talked about will be exempt, but the one time you don’t verify that… frankly, I’m not sure if the 6.5xMinimum Wage exception for computer-related fields still matters/applies… anyone know for sure?

    Profits from some forms of software development should qualify for the Domestic Production Activities Deduction starting last year, by the way, in case anyone overlooked it. Line 35 on the 1040, form 8903. It was a couple thousand off my taxable income this past year (we’re an S-corp). Definitely have an accountant look at your situation before claiming it, tho… the rules for what qualifies are written up in some of the most grossly generic legalese I have ever run across. Ugh.

    In short, depending upon location and persons involved, I’d think an annual burn of “only” $700K-$800K would be achievable. The challenge is getting a reasonable rate of return. Say you average only 1000 subscribers, and only get the standard $15/mo. out of them… that’s only $15K/mo, or $180K a year. $180K

    Comment by Craig Huber — 18 April, 2007 @ 5:15 AM

  9. Allright, i wonder if my local bank will give me a $3-5,000,000 loan?


    One day :^)

    Comment by DanielStone — 18 April, 2007 @ 7:26 AM

  10. [...] Iron Realms’ Matt Mihaly has posted a new post discussing the minimum cost of an MMO, seguing off Brian Green’s post on the same subject, and it’s neat to see dueling specifics as far as cold, hard cash numbers. [...]

    Pingback by Pro Game News — 19 April, 2007 @ 8:07 AM

  11. Cost of Building an MMORPG…

    I recently wrote an article that mentioned how the cost of building a commercial MMORPG now involved budgets in the tens of millions. Psychochild discusses the minimal cost of building a commercial MMO in his blog, and pins it down to around $3.5M doll…

    Trackback by GuildCafe Favorites — 19 April, 2007 @ 9:40 AM

  12. I could be wrong, but wasn’t DAOC done in about 18 months with a $2.5M budget? Of course, this was a while ago, and things were cheaper, but still.

    One metric that I try to use when making project estimates is what they call the “$10k per man month”. Basically, it will cost the company about $10,000 per month for every employee. Everything averages out (like salaries) and all of the other costs. A tight startup can usually get that $10k down to about $7k or so, but still, it is a good number to get a cautious estimate.

    I take the same approach for financial planning as I do when I travel. I throw everything I want to bring on the trip on the bed (the big feature list) and all of my money (estimated costs). Then I remove half of the baggage (trim the feature list!) and double the money. I travel light and rarely run out of funds…conversely, a smaller feature list makes for a quicker development schedule (and less stuff to cause bugs) and if you over estimate on your costs, you won’t find yourself short of funds due to unexpected events.

    Just my two cents.

    Robert / Nicodemus

    Comment by Robert Rice — 19 April, 2007 @ 10:54 PM

  13. I agree with Matt.

    [...] Actually, what’s been going on is a blog-to-blog topic, spreading around like a virus, on the topic of the lowest end cost of a MMO. Someone said two million. Then someone else said 3.5. Matt Mihaly, in the latest incarnation of the topic, says less than one. [...]

    Pingback by Azaroth’s Blog — 22 April, 2007 @ 3:04 PM

  14. DAoC used funky accounting for their costs. They had pre-existing technologies from previous technologies that and spread the cost over several projects to lower “apparent” cost of each title.

    It cost more than the $2.5m. About twice that much if I remember correctly, in line with AC…

    Comment by Ophelea — 26 April, 2007 @ 11:59 PM

  15. Pricey, pricey

    [...] “Wanted: two-million dollars” post. Brian “Psychochild” Green set him straight. I didn’t contribute to this thread as I had my own take on it a year earlier with a [...]

    Pingback by Mischiefblog — 22 May, 2007 @ 2:13 PM

  16. So, I get the $3M and get a game finished. What will it cost to launch and support this? How many boxes will need to be sold at the $50/$15 current model to make a profit within a reasonable time? It would take 60,000 copies to pay back the $3M but do you need that many? Is that enough monthly income to support the game?

    Comment by Silvermink — 2 March, 2008 @ 6:30 AM

  17. As with most things, the answer is, “It depends.”

    Launch costs depend on what type of launch you want. If you are doing self-published distribution off your website, then you’ll have less expense than if you’re putting boxes in stores. Support costs depend on the type of game; a hardcore PvP game is going to demand more support than a low-keyed game. A buggy game is also going to need more support than one that is relatively stable.

    For a smaller game, you’re probably going to have lower support costs. However, you do need to carefully consider your business model. For example, having a few thousand players means that subscriptions may only bring in enough money to tread water. Item sales, on the other hand, may add to your development and launch costs.

    Also note that if you distribute boxes and charge $50 per box, you don’t get all that money. Stores, distributors, publishers, etc. all want their cut. As the developer, you might expect anywhere from $0 to $10 per box; some publishers might also try to take some of your ongoing revenues as well. So, you’d need to sell more than 60k boxes to make back $3M.

    Hopefully that gives you some info. Feel free to post if you want more.

    Comment by Psychochild — 4 March, 2008 @ 11:56 AM

  18. I read through this post the other day, then saw in Dave Allen’s post about QoL that they made the game with a virtual company for about $3.5m.

    Just a data point, I guess! :-)

    Comment by UnSub — 20 September, 2010 @ 7:51 PM

  19. Hah! I love a good necro posting. ;) A very interesting data point!

    It’s been fun reading through these old comments. Unfortunately, I suspect Matt Mihaly wasn’t able to stick below that $1M figure for Earth Eternal. How the world changes in only 3 years.

    Comment by Psychochild — 21 September, 2010 @ 10:49 PM

Leave a comment

I value your comment and think the discussions are the best part of this blog. However, there's this scourge called comment spam, so I choose to moderate comments rather than giving filthy spammers any advantage.

If this is your first comment, it will be held for moderation and therefore will not show up immediately. I will approve your comment when I can, usually within a day. Comments should eventually be approved if not spam. If your comment doesn't show up and it wasn't spam, send me an email as the spam catchers might have caught it by accident.

Line and paragraph breaks automatic, HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Email Subscription

Get posts by email:

Recent Comments


Search the Blog


November 2019
« Aug    



Standard Disclaimer

I speak only for myself, not for any company.

My Book


Around the Internet

Game and Online Developers

Game News Sites

Game Ranters and Discussion

Help for Businesses

Other Fun Stuff

Quiet (aka Dead) Sites

Posts Copyright Brian Green, aka Psychochild. Comments belong to their authors.

Support me and my work on