Inner Space

the classic game


Case Studies

world class design & results


Bill Stewart

our founder's profile


In the late 1980s, we discovered an opportunity to create a new product. This lead to lucrative success and the creation of a new product genre. Like all great UX work, we started with a hypothesis and kept making improvements based on user feedback.


Ian MacDonald and Bill Stewart were experimenting with graphics to add visualizations to boring projects. We experimented with variations. Making the graphics a randomized screensaver was an efficient way to look at variations. We shared it online for free because we liked it. We improved it and added pricing because users demanded it. Of many user requests, we charted the most requested and impactful and acted on those (removing bugs, adding a few features, password protection, site licensing). After a few dozen iteration cycles, our freeware experiment had become big business making money hand over fist with no advertising or consumer reach. We scaled it up, got a publisher to put it in stores and it quickly became the top-selling software product in the world.

Even the most trivial thing can evolve into a worldwide success if you collect good data from users, then analyze, execute and iterate. Read below for the full story.

Easy to use


The idea of paying for a screensaver may seem weird now, but it was 10x weirder back in the late 1980s. Computers were slow. Graphical interfaces were a relatively new thing which made computers slower. You could get a cup of coffee, do jumping jacks or discuss politics while waiting for a big spreadsheet to update. Why would anyone want to pay for software running in the background, slowing their computer even more, that showed animated graphics when they weren't even there to see it?

Screensavers seemed to make no sense, even as a free product. We discovered an opportunity that didn't seem to exist. We had relatively boring contract work, so we experimented with graphic visualizations to make it less boring for customers and quite frankly, ourselves.

We tried variations on charts and visual shapes. It is easier to find something good if you see lots of variations. We structured our code as a randomized screensaver to see many variations quickly. We found some variations just felt pleasing and interesting to look at, so we tinkered to improve it. Since we liked it, we thought others might, so we shared it via an online service. This was before the web existed, before most people had even heard of email. Naturally, we didn't ask for money or consider it a product, but we did include our mailing address in case people wanted to write a hand-written letter to us. Old school.

We called it Magic. It was composed of simple line segments, but we had tweaked the animation code to have rhythmic shifts in tempo, direction and color that were surprisingly soothing. It felt magical because it *seemed* to sync with music and calm the mind. Hence the name, Magic. When we posted it online, we expected nothing at all. Instead, we were quickly deluged with mail praising our work, requests for bug-fixes and features and many hand-written fan letters including cash payment. They usually included an apology for not *finding* our price list, so users guessed, sending us $20-$100 cash.

We didn't quite realize what we had, but felt a moral obligation to fix bugs and make it better. It was exciting having fans. Every time we released a new version, our fan base, amount of mail and money coming in jumped up. It was cool getting funky-looking cash from all over the world. We added documentation, an order form and a price list with the company name so people could send checks instead of cash.



We had a moment of clarity. We had fans and were making money from a fun little diversion we enjoyed working on. It took time away from doing mundane local contracts. So we said 'This is what we're doing now. Magic is actually what we do.'

Once we took it seriously, things really ramped up, but there were engineering problems to solve first. To operate, a screensaver has to run 24/7 in the background, looking for keystrokes and mouse movements. The lack of user input for a certain time indicates when animation can start without interrupting work, and user input while animating tells the screensaver when to stop. We identified several important product factors:

  • How can the screensaver operate in the background without slowing the computer down?
  • How can we ensure that the screensaver won't crash, even if other software messes up memory on the computer?
  • Who are our customers? What is their environment? How do they rationalize purchasing it?
  • What features can we add that add genuine value for our most important types of users?
  • How can we lower the cost of ownership (beyond simple purchase price)?

We collected and organized user feedback to answer these questions. Magic did save screens from burning in, but research showed that nobody paid for it because it "saved their screen". Users enjoyed it, but they used screen burn-in as a justification for purchase. The intangibility of the perceived value meant that users had no tolerance for errors, slowdowns or crashes. Nobody "needed" it for their work. Word could crash 50 times a day and users would have to put up with it. Not so with screensavers.

Solving problems isn't that bad when you realize what's really important to solve. We tweaked and redesigned it through dozens of iterations until it ran 24/7, responded instantly, but never slowed the computer down at all and never crashed. Not crashing was a HUGE deal back then. Every program ran in the same memory space, so if anything crashed, the computer flatlined until you turned it off, waited, then restarted. Crashing even once was a complete deal-breaker.

We added password protection and other features because users wanted to buy site licenses. Computers couldn't be locked down back then. When we added the ability to lock a PC down with a password, it opened a very big door. Sales increased 100x as we sold site licenses around the world to companies such as DuPont, Lily, Microsoft, Atomic Energy of France, The Kingdom of Tonga, the Federal Reserve, etc.

Operating system


Business was booming, but without marketing, we knew we were only scratching the surface. We had vision. At the time, Bill Stewart said that "If we can add more fun, keep the perfected engineering, put it in a nice box and get it in stores, I'm certain it will sell millions."

We found a publisher, carefully added a few features, created a lot of fun animations, helped develop packaging and distributor relationships, and the product sold millions.

While we were developing Magic, a UC Berkeley grad student named Jack Eastman created a Mac-based screensaver experiment he named "After Dark". Berkeley Systems was the publisher for his crude early version, so we chose them too. They were reluctant partners because they didn't believe Windows was going to be big business, especially for graphical software. After we signed with them, both Microsoft and an eventual competitor offered lucrative deals to work with them instead. We (foolishly) felt honor-bound not to entertain breaking our existing contract for better offers.

We toyed with naming the new product line "Magic After Dark", but settled on just "After Dark" because we wanted to differentiate the new product from what we'd already done. We also had some confused fans who assumed Magic was "spookily entertaining" because it was channeling dark, satanic forces instead of simply being a rhythmically pleasing algorithm.

The first version of After Dark sold like hot-cakes, topping sales charts 3 months after initial release. It helped that Microsoft Windows exploded in popularity and After Dark was the cool new product to showcase their new OS. The 2.0 release was an even bigger sensation a year later. Working the trade shows was surreal. When actual marketing people demo'd the product, things were quietly OK at our little booth, but when we (techy introverts) demo'd the product, we drew monster crowds who never wanted our demo to end. Other companies had celebrities, special effects, comedians, magicians. We aimed to concisely hook users in 60 seconds, but they kept asking for more until we ran out of things to say. Some stuck around and watched the whole 25 min demo AGAIN, then brought others over to see it the next day.

It was the biggest selling software product. Like Magic, the most satisfying part of it was that it made millions of people happier in their day. Users valued the saving of their screen as an asset, enhanced security and the animations and there was no downside. You don't have to do too many things right if you don't do anything wrong.

Saves time


When After Dark launched, we gave all Magic customers a discount to switch to After Dark since it was the successor product. Magically, very few Magic users actually switched. While After Dark boomed, Magic sales soared too. At the time, it was a pain in the neck. We had a team on After Dark, lots of work to do, trade shows to travel to. We wanted to focus on After Dark, but we kept getting tons of Magic orders which wouldn't go away. In retrospect, it was pretty cool that we kept getting Magic orders 8 years after we stopped working on it and actively started dissuading users from ordering.

When After Dark got to version 2.x, we were planning new products and features. There were Star Trek, Disney, Simpsons and other screensaver packs. We also invented the first live wallpaper as a product to be called Before Dark. It was 15 years ahead of its time but worked flawlessly.


Unfortunately, despite Berkeley Systems making over 90% of the millions in profits from our work, they felt a desperate need to get rid of us. Perhaps they resented our getting even a tiny percentage, or hated being dependent on outsiders for their livelihood. They used a legal loophole to cut us out and make After Dark 3.0 without our involvement. It was a bad call. I'm sure the team brought in to replace us did their best, but AD 3.0 failed. Bad reviews. Negative profits. Company floundered. The most loved brand in software was tarnished. We lost out on millions, our customers were mislead, the product was ruined and our ability to build on that to create more significant software was kneecapped. Not good. Ownership of After Dark traded hands a few times and the 4.0 wasn't as bad as the 3.0, but the magic was gone.


It was fun to create a pop culture item featured in comic strips, tv shows and movies as something universally recognized. It was nice to hear from designers that our approach to usability and low cost of ownership was influential. These were great things.

Our big takeaway is how satisfying it is to align a product with users' spoken and unspoken needs and wants. If we ask "How can this enhance the humane life of users?" before "How do we make the most money from it?", there is a lot of profit in doing right by users. Many sites and apps today have lost sight of authentic human needs in favor of "maximizing views, in-app purchases and cool just for the sake of cool". At the end of the day, screensavers weren't just cool and the product wasn't good by accident. Relentless attention and iteration on user feedback led us from fun tinkering to making the most profitable utility ever - a utility that protected your monitor and locked down your PC that was truly loved by fans. Users liked the flying toasters, but they loved the product because the look, feel, engineering, utility and usability felt premium, showing respect and assurance to the user at every turn.

We learned how to go beyond merely good to design and build truly great software. It doesn't make us "gurus", but it is useful experience we do our best to carry into all new projects.