Interview with Orbus Gameworks' Darius Kazemi

January 2, 2008

Darius Kazemi, President of Orbus Gameworks, was a popular speaker at the first 2007 ION Game Conference, discussing the dirty details of data mining in online games to a full house. In the first of several interviews with ION Alumni, we will be discussing what some of our past attendees and speakers have been up to since the first ION last year – starting with Darius. Read on for more details, keeping in mind how amazing it is that so much business can be done in just over seven months!

PAUL PHILLEO: Thank you for spending some time with us to answer a few questions for ION Connection! Give us a quick introduction to yourself, Darius, and what you do for a living.

DARIUS KAZEMI: I run a gameplay metrics middleware company called Orbus Gameworks. Our software helps developers measure how people are playing their games, and came from my background doing just that for D&D Online and Lord of the Rings Online. We’re based in Boston, MA.

Since the first ION Game Conference (ed: previously known as OGDC) ended last May, what have you been up to in business over the last six months or so?

Since ION, we hit our 1.0 milestone on the middleware and since we’ve been licensing our metrics software, and my engineers are hard at work on 1.1.

How did attending and speaking at ION affect Orbus Gameworks this year? Anything you can share with us?

While I can’t share specifics, I can say that most of our biggest sales leads came from ION 2007. I attended about 15 conferences in 2007, and ION was by far the best in terms of business opportunity.

What were the most memorable qualities — and moments — about ION 2007 for you?

I really enjoyed the show’s laser-sharp focus on online games. Other conferences may be unofficially about this topic or that topic, but almost every attendee at ION was directly involved in the online games space, which made networking very easy. It also meant that the session material was a cut above what I normally see: nobody had to dumb down their talk for fear that someone who’s never worked in the online space might not understand.

I suppose the other thing that really stood out about ION was that as an attendee, I felt like the organizers really appreciated my presence. It was the little touches, like the excellent food and the fact that everyone got a chance to meet and talk to the organizers. I didn’t feel like I was just another registrant feeding a machine.

In terms of your business at the basic level, who can benefit more from metrics analysis: marketers or developers?

That’s a great question! I think that when it comes to gameplay metrics by themselves, the developers definitely benefit more. The value for marketers really comes when you cross-correlate gameplay data with demographic or billing data. This often involves the harrowing process of integrating your billing and web site or forum databases with your gameplay metrics database, so I think that for marketers to really benefit it takes a little more work. On the other hand, marketers benefit when the game is better, so what helps the developers helps the marketers.

Reading your blog, you frequently comment on metrics applications in online-capable games like Half-Life 2 and Halo 3. Speaking in broad strokes, these days what are developers getting right and wrong with metrics collection and analysis?

What developers are increasingly getting right is that they’re doing metrics at all. Nobody was really doing this in an organized fashion even five years ago. However, I keep seeing metrics being implemented poorly by developers who are often trying metrics out for the very first time. A huge beginner mistake is to record stuff to a logfile rather than to a well-designed database. To take your examples, I’m pretty sure that Bungie and Valve have database-driven metrics, but in their recently published post mortem, the BioShock team at 2K Boston wrote that it was almost impossible to get decent metrics for their game, and I think it was partly because they were dumping data to text log files.

How difficult is it to correct these issues with metrics in online game development? For that matter, how important is it that they are corrected? How much of a difference can they make in the commercial success of a game?

While I’m tempted to say, “buy my software and everything will solve itself,” the truth is that mid-project it’s very hard to shift metrics collection methods. I think the natural course will be for developers to implement metrics poorly on one project, and then reassess and do it the right way on a future project.

And I do feel that for any game with any sort of deep online gameplay, metrics are key to both commercial success and live operations success.

On the commercial side, you can really fine-tune your gameplay during closed beta and before launch, to the point where you can track down anomalies and ask individual beta testers what they were doing to cause such-and-such an event you’ve recorded.

On the live ops side, metrics are crucial to the maintenance of a virtual world. When the design team wants to nerf a weapon, they can literally ask, “How many of players does this affect? Oh, only 0.5%? Well then, let’s nerf the weapon.” You can track down gold farmers, and exploiters. There’s a lot a live ops team can do with a good metrics system.

I can’t let you go without asking you: Do you plan to attend ION in 2008?

Absolutely. In its second year, ION is already on my short list of must-attend conferences.

Free Newsletter

Join the ION update list and receive our free weekly newsletter with all the latest conference and online games industry news.

Premier Sponsors