Where does the money go?

Everybody knows that building an MMO involves a lot of people, takes a lot of time,  and costs a lot of money. Fewer people know how much money, how many people, and what all those people are doing for all that time.  I thought I would share a bit of my experience on Pirates and give you some idea what those big MMO budgets are spent on.  All of the data behind this post is up on Zoho Sheet; feel free to do whatever you want with it. (Also, Zoho is much cooler than the Google Docs spreadsheet.)

There are a few things I should mention up front: These are not the numbers from Pirates.  We had a gradual ramp up of people and project scope over the course of five years and that’s definitely not the right way to do it. I also rearranged people to compensate for some of the staffing shortages we had on Pirates.

This “budget” also doesn’t include anything but people’s salaries. Most costs scale up with staff, including desktop hardware, office space, software, office server capacity, taxes, benefits, etc. You would need to add 40-50% to the dollar figures to take that into account.  I mostly care about the percentages, so I didn’t bother trying to include any of those items. The budget also doesn’t include any server, hosting, or bandwidth costs. Those can be pretty significant in the beta and live phases.

Most of the salaries were drawn from the 2007 Game Developer magazine salary survey. Those are:

Programmer $83,383
Artist $66,594
Designer $63,649
Producer $78,716
Tester $39,062

There were three functions I didn’t have any salary surveys to draw from, so I just made up some numbers. I assumed Community people make about the same as Designers and that Operations people make about the same as Programmers. I put Customer Service people down at $30k on average, which is much more than front-line CSRs and forum moderators, but less than supervisors and managers.

Community $63,649
Customer Service $30,000
Operations $83,383

I broke the project down into four phases: Pre-production, Production, Beta, and Post-Launch. Pre-production is a period where tools are being built and the game is being designed, but work has not yet started on much (if any) final content. Production is the Big Expensive Part when tons of content people crank out all the final assets for the entire game. Beta is the period at the end of production where your game is exposed to external players in a significant way, and includes closed and open beta. Post-Launch is obviously the period after the game has launched.I also divided the effort into three major areas: World, Systems, and Infrastructure. World is the construction of the game world, quests, dungeons, etc.  Systems is the development of character classes or skills, character customization assets, UI, and game systems.  Infrastructure is the less-glamorous stuff behind the scenes like core server code, operations tools, scheduling, and IT. I broke it down this way for the benefit of a future post that currently exists only in my head.


The purpose of the pre-production phase is to figure out the answers to a bunch of questions before the team grows to its full size and things get really expensive. There is an emphasis on programmers and system designers in this phase, because they have the most questions to answer. As expected, the programmers take up the biggest chunk in pre-production. They are both numerous and expensive. System development is also a big focus in Pre-production.

Main title - http://sheet.zoho.comPre-production by Function - http://sheet.zoho.com


This is where most of the money on the project will be spent, and most of that is spend on building the world:

Production by Area - http://sheet.zoho.com
Production by Function - http://sheet.zoho.com

During the beta phase other significant costs (such as operations and customer service) start to come in, but the content team is still going full bore on building out and bug fixing the world, so it doesn’t affect the numbers too much.

Beta by Area - http://sheet.zoho.com
Beta by Function - http://sheet.zoho.com


The Post-Launch phase of the project is represented in this graph primarily because it takes a while to get money out of your players and through the various intermediaries involved, and into your bank account.  Even if you have 100k players on launch day you won’t see revenue from those players for a while.  If the game isn’t at least self-sufficient beyond that, then you’re in trouble.  The Live phase has a smaller number of world artists since you are not building the main world anymore at this point. Depending on the (paid or unpaid) expansion pack strategy of the game, this will vary.

Live by Area - http://sheet.zoho.com
Live by Function - http://sheet.zoho.com


Here are some graphs to break down the total amounts spent. See how world is the biggest chunk?  That’s why everyone is talking about user-generated content.

Totals by Area - http://sheet.zoho.com
Totals by Function - http://sheet.zoho.com
Totals by Phase - http://sheet.zoho.com

How closely to these numbers reflect the budgets on MMOs you’ve seen? I’d love to hear your thoughts.


3 Responses to “Where does the money go?”

  1. Little Joe replied on :

    Very interesting to see the differences in QA, both up front and over time, between the game world and what I’ve worked on to date. Ouch.

    As a comparison: for QA, I was earning $70k at Quark when I left in 2002. One of the people I worked with there just left to come work at my current company. She was the QA lead for one of their major products, with a large team of testers under her. She was on on the high end of 5 figures when she left. (Of course, working on desktop publishing software isn’t nearly as cool as working on a game).

    Also interesting to see the involvement and ramp-up of QA over the life of the process. In my current company requirements gathering/writing, product schedule prototyping, release criteria, etc. end up being tasks for QA as there isn’t a real product management role. We have a 6:5 developer to QA ratio and QA is constantly struggling to meet all its document deadlines, and write and execute tests on a rotation schedule that still meets an acceptable code coverage percentage.

    Of course all that up front work by QA means that we’re releasing products on schedule that are in the neighborhood of 1 defect per thousand lines of code instead of the 7-15 defects PTLC at previous jobs where they were more lax on procedure.

  2. Tim wrote on :

    Joe, I’m curious what your thoughts on are on the impact of technology selection on this cost. Basically if you invest in technology, how much does it modify your staffing needs? IE, if I bought Big World’s game engine, do I really need as many people for infrastructure and systems?

  3. Joe replied on :

    Buying middleware definite impacts the staff you need. You effectively transfer some of your engineering time to your middleware partner. Usually this is more about saving time than it is about saving money. Major middleware (like Hero or Big World) gives you a huge jump start and lets you skip most of the engineering in Pre-production, but it’s also pretty expensive.

Leave a Reply