Brad Wardell's views about technology, politics, religion, world affairs, and all sorts of politically incorrect topics.
Practical realities of adaptive software development
Published on March 31, 2006 By Draginol In Game Journals

Stardock's business model has been predicated adapting to what users request.  The company's motto is "Innovation on demand".

The idea being that we release the best product we can and then we rapidly enhance it based on user feedback.  Users of Object Desktop, for instance, are very familiar with this.  We have also introduced this model to our game development which has had very positive results.  Our conclusion isn't so much that Stardock is "great" at customer service but rather most PC game companies are simply not geared up to do the kind of development model we're used to.

The PC utility market is extremely competitive and so rapid updates are one of our critical advantagers.  Since 1999, a huge percentage of Stardock's R&D budget has been put into building an infrastructure to crank out user requested updates quickly.

Central to that strategy has been Stardock Central which allows, literally, a developer to put up a build onto the system and have it available to users without creating an installer.  In addition, a custom set of forums that interact with the user's account information was created over the past few years that lets them interact with all our sites, Stardock Central, and access things quickly and easily depending on which site they're on.  It also has allowed us to implement various "community" features that helps make users part of the team -- because they are.

Where the model is starting to break down is as we've become more mainstream, the sheer numbers of users has started to show weaknesses in our system.  In addition, the demographic has dramatically changed which has significantly hampered our ability to do what we're best known for doing -- crank out rapid updates.

In the past, especially with Object Desktop, our demographic was almost completely all power users.  As a result, we could release alpha-level updates to those users who, in turn, would let us know what changes they wanted or any bugs they found. Power users who weren't comfortable with alphas would wait for betas and power users not comfortable with either would wait for releases.  Bugs, features, etc. would be queued up to be addressed and the system flowed nicely.

Over the past year or so, the utility market we're in, the desktop enhancement market has had two odd things happen: First, the market has increased in size dramatically. Secondly, the number of other providers of such software has strangely declined (despite more opportunity).  What happened to all these developers is subject for a different topic. But the net result is that, Stardock's site for supporting Windows customization, exploded with far more users than had been expected. Even now, it's basically limited by server levels.

The market size increasing though has brought in users who are not power users. Users who don't read manuals, don't understand the difference between alpha, beta, RC, release. Don't know how to resolve problems on their own.  The net result has been that the same updates that we used to just throw out onto Object Desktop (or ThinkDesk) now have to go through an internal QA or else we'll suffer in massive support calls (phone, email, more phone, more email) that drowns out truly serious support (In the last week of February, a sampling of tech support email resulted in the conclusion that roughly 2 out of 3 emails would have been resolved had the user simply looked at the readme in the given product or gone to the knowledge base -- two years ago, that would have been 1 out of 10 emails).

Of course, we're not set up with that kind of internal QA, not in terms of a long-term infrastructure.  When we do a major release like WindowBlinds 5 or Galactic Civilizations II we can put together a pretty impressive QA team using people from other parts of the company. But those people are far too expensive to have doing QA for the long term.  So now we're in the position of having to build a full time QA staff that can rapidly handle the updates that come from the various development teams and get them out there.

Similarly, we're going to have to start hiring people to do forum support. More people I should say.  In our rapid development model, we would put up an update that would list what's in there and users would respond with what they found worked/didn't work.  Now, we put up updates and people will ask "Did you fix <bug X>?" or "When is <Feature Y> going to be put in?"  That is, assuming the user even reads through the change log at all before asking questions.  Then there's the users who read the change log but don't try the update and complain about theoretical things.  "Oh, I don't like that new WindowFX feature, I think it might slow down my computer." or "I think the economic change in GalCiv II is bad" (without having tried it).

Of course, right now, we simply don't have those people to beef up a permanent large scale QA department or to act as a day to day buffer between development and support. They have to be hired and trained which we are, but it's a slow process.

So the business model of developing niche software (desktop enhancement utilities and turn based strategy games) in order to attract power users and hard core games has started to break down.  They have started to break down because those markets are no longer niche markets. Turn based strategy games have become mainstream and desktop enhancements are mainstream. And so we will have to evolve.  But it's been painful and likely to continue to be painful for awhile.

Comments (Page 1)
2 Pages1 2 
on Mar 31, 2006
I sure can sympathize, even though I may not understand the finer details. Yet, you all have been doing an amazing job, especially in keeping us informed. Allowing us to see inside all that is developing is what makes this such a great community.

The only thing that bothered me is the way the article ended. Where is the light at the end of the tunnel? It had an air of 'things may get worse before they get better' to it. It seemed as though there was some sadness to the prospect that all of this is or has gone mainstream. I hope no mayyer what that the 'personal' touch that makes these sights so great, never diminishes. No matter what, I will stick it out as a lot of other people will. I'm patient and highly addicted to all these sights have to offer.

If I only had some talent to offer...and lived in Michigan.
on Mar 31, 2006
you've got a lot of support out there, brad - maybe you can rely on some more 'vollies' to monitor the forum while you bed the QA in? not me. i iz crap
on Mar 31, 2006
I'm betting your policy of listening to feedback and providing rapid updates also has the effect of creating more vocally demanding/critical customers by encouraging them to think about the product's flaws and raising expectations that they will be improved/fixed in the near future. That's a two-edged sword because if the fixes/changes don't come (or just not quick enough), those heightened expectations are deflated and you get upset customers who might have been otherwise content if they hadn't in the first place been encouraged to start thinking so much about what they'd like improved. But on the other hand, it's a real plus for many people that the developers are taking a really active role in post-release updating. Personally, I've bought at least 3-4 pieces of software I was on the fence on buying after seeing its developers in a forum interacting with customers and quicking releasing updates/fixes, so I'd say that the benefits of your rapid update model are worth the risks. Keep up the good work!
on Mar 31, 2006
I think your current business model is great. Incredibly demanding on yourselves though. I was wondering how you and your team could possibly keep up the pace since the release.

I personally don't play with beta's because I know I would not help in it's development.

Thanks for all your hard work. I cannot wait for next weeks release. It will almost be like getting a new game again. It will even have new features I have asked for, and what could be wrong with that?
on Mar 31, 2006
My sympathy. It suprised me how many people didn't know what a beta was. I thought any gamer should've known that.
on Mar 31, 2006
Welcome to the horror that is mainstream
on Mar 31, 2006
You may want to look at the as far as utilizing this growing community of users both as testers, and moderators.

in a recently released mod for RTW - the modders put the beta to the community which responded with zeal to both test, and document bugs, CTD's, and nice to have features - as far as I can tell its been working a treat

of coarse only a few community members are going to have the knowledge to do this effectively although if its structured well all users even simple ones (who dont read readmes) can easily find a place for constructive input, where it can then be filtered by more experience members of the community.

and as far as forums self moderation by senior and responsible members of the community who volenteer their time seems to work fine

evolution can be a good thing - a strong community of fans is an asset - utilizing this asset well will make your lives easier and make the community feel that have an integral role in the future of the game they love - in the case of totalwar it took the modders to do what CA could/would not
on Mar 31, 2006
but rather most PC game companies are simply not geared up to do the kind of development model we're used to.

Overly modest, it seems.
on Mar 31, 2006
That transition from samll to medium organisation can be one of the most vulnerable times in a company's life. But you have great products so I'd say the odds are in your favour.
on Mar 31, 2006
Ironically, the on demand/beta and release design process that doesn't work so well here is, more or less, the software engineering model Google uses and what Microsoft is trying to get into for Vista. As you've stated, the backend is most problematic, ie the Vale/Steam method of broadband software delivery. Best of luck to you and Stardock.
on Mar 31, 2006
I was for one was quite suprised to see post's like this one and others offering relatively rare insights into how a software devloper has progressed over the years. And its refreshing.

I think the real challenge for stardock will be (one that has probably all ready being addressed, if it hasnt been so all ready) "How do we not alienate existing 'hardcore gamers' or power users while at the same time appealing to wider interest, to further increase the market for our games", a more mainstream market. For example, talking about games here, If anyone compares Oblivion to the earlier Morrowind its not difficult to see not only the cosmetic changes (which were not interested in) but also changes to the fundamentals of how the title is represented to others. Quick example, the increasing abscence of statistics and accurate figures (abscence of anything other than a whole number in item stats for example) to appeal to a more mainstream audience. With the Elder scrolls series appearing jointly on console and PC for the first time ever, its something that will appeal more to casual console owners, rather than the more traditional 'hardcore' PC rpg players. Im sure many of these console owners would be far less PC savvy than a power user, so we can compare similar situations at stardock and Bethesda.

Even though it has been stated earlier that Stardock wont cater to the casual gamer specifically, with the kind of sales figures coming in, it would be silly to pass up on this oppitunity and not make a sequal to Gal Civ II that doesnt increasingly cater to the a more mainstream market of which Gal Civ II suprisingly found itself plunged into. But i fear in doing so, features that made the series popular with 'hardcore' gamers in the first place will be sacrified or dumned down.
on Mar 31, 2006
I'm actually very impressed with the direct relationship you all have with your customers and fans alike. Nevertheless, when loads increase, I guess it's not easy to keep up. Good luck with it anyhow, i'm sure you'll find bunch of trustable members of the community willing to lookout on forums for free!
on Mar 31, 2006

A victim of your own Success!

And yes, Users are ecpecting more immediate gratification and fixes.  I think this is the Microsoft syndrome where they got so blasted for security holes and were rushing them out to the tune of a couple a week. 10 years ago, you patiently waited for the next dot oh x release to patch your software.  Now, you want it patched today.

on Mar 31, 2006
If the wife would let me move to Michigan . . . . I'd be there to help formally.
Let me know when you open the Houston office.
Until then, I'll do what I can helping with support at WC.
on Mar 31, 2006
I would not recomend having volunteers do Forum Administration, unless you give them a very strict set of guidelines that are available both to forum users and the administrators themselves (obviously).

2 Pages1 2