The 6 Month Product Process Behind The Launch Of Parsec
Chris Dickson, my co-founder, and I share the product role at Parsec. As we got started we tried to employ all of the lessons we’ve learned building other technology companies and products for startups. It was really important to us that we eliminate all of the technical risk as quickly as possible due to the challenge of building an interactive streaming application. Chris started building the product by asking, “if bandwidth, compression, and hardware continue to improve, it should be possible to build a gaming experience that feels native, why not try to build it ourselves?” If we could build the technology, why wouldn’t everyone who is building or upgrading their gaming machine play on a cloud machine? As the founder of Hover Hound and PC Hound, which have about 70,000 users building PCs inexpensively, Chris had a lot of experience with people building and upgrading their computers with these tools.
As product people, we wanted to figure out two things early on:
- How can we eliminate our biggest product risks as quickly as possible?
- If the technology is possible, will people use it?
Getting Started On Parsec
The inspiration for Parsec came from a Hacker News post about one year ago. If you could almost play a game on AWS using a VPN workaround for Steam-In-Home Streaming, you should be able to build something purposefully built to solve the problem. The goal was to find the quickest way to test the business, so we initially turned to open source projects. We tried using Moonlight and Gaming Anywhere, but those proved to be dead ends in terms of performance. Following that, Chris built a rough project, taking shortcuts where he could, to build a test product for early users. We got our friends to test it out, and the feedback was overwhelmingly positive, but many users caveated their enthusiasm with some issues we needed to resolve around crashing, latency, and usability.
Hooray! It’s Possible — Do People Want This?
Now that we knew it was possible, Chris started building a clean version that eliminated a significant amount of the latency in the product. Although we know of many ways to improve the latency even today, we had to get something into users’ hands to prove number 2, that people actually wanted our product.
We released an “Alpha” version of the technology on Reddit, hoping to get a couple users to test it out. We got an incredible response and lots of sign ups. With very little effort on our part, we had 50 testers of our product. We weren’t exactly proud of user experience of the application at this point (our users had to launch a terminal application and change their setting in a config file buried in the application package), but the excitement from the Reddit post showed that there was pent up demand. With a terminal application as our front-end, one of the biggest issues we had early on was convincing our early customers that we weren’t hacking into their computer. Mr. Robot has really gotten people scared of their terminal :).
We weren’t worried about the technical hurdles to build the web application. We knew that was possible, so it wasn’t worth testing earlier in the summer. We had validation that people wanted the product and that the technology was possible.
Testing With Early Users
Our earliest users found LOTS of bugs, and we learned a ton about the use cases and the different technology setups we’d have to support. We’re still expecting to run into a lot of technology issues after this release, but we’re confident we learned a lot by getting our product into users’ hands early.
After learning a bunch from early testing, we also built a scaled down version of our cloud product. Friends could log-in, play a game, and report back any issues or feedback directly to us. Again, we took a lot of shortcuts. We didn’t build user-specific settings or give people access to many games because our goal was to find out if the cloud product was viable and if people who didn’t have a gaming machine wanted to use it.
Improving The User Experience
After the positive feedback here, we focused on the technical hurdles of turning our product into a web application and developing a web launcher. User expectations have shifted in the last decade from desktop applications to web products. We needed to fit the current mental model, but it also saves us from building different GUIs for every device. For usability, we started moving all of the settings and features into the web app. Things like login, VSync settings, resolution settings, and max bandwidth were in our terminal application, but none of our early users knew how to change them because it was impossible to find the config files.
The web app wasn’t pretty, but it gave us an opportunity to test and collect feedback in June and early August.
From this, we learned that people really wanted to use the software, but were having a difficult time navigating and getting started. Recently, we hired an awesome designer, Ed Moss, to help build out an onboarding flow and improve the overall look and feel of the Parsec brand.
Additionally, there were a couple more user experience issues with turning a gaming PC into a server that we had to hash out. The most important of which was eliminating the need to open ports on your router. No one wants to spend time doing this, so our software does its best to open the ports automatically with UPNP. Unfortunately, some routers don’t allow for an automatic process opening ports, so some users will still have to follow the manual process of opening ports via our documentation.
Always Eliminate Your Riskiest Assumptions First
It was our goal from the beginning to eliminate our riskiest assumptions as quickly as possible and then move toward overall improvements later on. From a product perspective, I recommend that you always think about your Minimum Viable Product being the next significant iteration of your product. I believe it’s imperative to build products like a science experiment, focusing your development work on the next most significant hypothesis in your business model.
We’re now ready to test a huge assumption in our business. Will a large audience of gamers adopt our technology and use it to consume their favorite games? And if we reach a large number of gamers, will they keep using Parsec? We hope you’ll try it out and give us feedback, so we can continue to improve our product and business.
Thanks!
