This past month, we launched Trelent as a beta service in North America with capacity for hundreds of initial users. Trelent is, as the title alludes to, a full Windows computer in your browser. You can access Trelent from anywhere, up/downgrade your PC in seconds, and ultimately use it to speed up old and outdated machines without bearing the upfront cost of a traditional brand new computer.
You can think of our platform as a live video streaming service, where you control a computer on the other end instead of simply watching.
We want to share our launch experience with you, and provide a series of posts as our company, service, and user base evolves. To kick it off, here are a few lessons learned from Trelent’s initial launch in August 2020!
Lesson #1: Expectations versus reality.
As much as we wished we would fill available capacity from the get-go, this simply was not the case. Instead, we on-boarded three users in our first week. Despite thousands of impressions and dozens of up-votes across various product placement lists (ProductHunt and BetaPage, for example), only three out of over thirty thousand reached individuals (roughly eighteen thousand unique) opted to try out and use our service. So what went wrong? Put simply, we had expectations that severely outpaced reality.
With a service like Trelent, the general public is not aware of the benefits provided, or the problems we solve. It is very new technology, and new technology tends to follow an S-Curve adoption rate, with users slowly picking it up, then growing exponentially, and finally tapering off growth again once a certain number of people uses the technology. We now recognize that we are still very much at the beginning of that S-Curve, even if we hoped we would jump to the middle of it.
Lesson #2: Pricing model matters.
We polled dozens of people post-launch on questions they had about Trelent, and the most frequent question we received was “How does the pricing work?”. Trelent began with a combination of hourly pricing and monthly storage, where for every hour you used your Trelent PC, your account balance would decrease based on the plan you had selected. For example, if you opted to pay $0.50/hr you would get twice the performance as if you paid $0.25/hr. Storage was a separately-billed item, at $0.10/GB/Month. This is simply too complicated.
It took several sentences for us to explain our previous pricing model to you, when instead we could have been using a simple approach we are implementing now: a flat monthly rate with usage caps. For example, instead of “Pay $6.40/mo for 64GB storage and $0.25/hr of usage for 20 hours, for a total of ~$10/month” it is “Pay $10/month for 64GB storage and 20 hours usage”. It is a lot easier for customers to understand this!
Lesson #3: Even in marketing, 0x3 still equals 0.
This may seem obvious — grade school math should be after all! However, when we had zero conversions on numerous marketing campaigns we tried several times to simply increase our budget, sometimes by up to three times. This was simply a waste of resources when we weren’t sending the right message in the first place. We received comments which brought to our attention the fact that we weren’t targeting the right people — put simply, those we were reaching did not find our service particularly useful, and reaching more of those people did not help that fact.
Onward and upward: what comes next
Trelent’s launch was not perfectly smooth by any means, but it wasn’t a complete failure either. We learned a lot, and ultimately, we still did find customers (if only a few)! We’re going to spend time and put the work into improving our service and gathering data to launch new versions of Trelent in the near future, with more success. We have two large features planned to supplement our next launch.
First on our list: Linux support. We recognize that with our service starting at $0.25/hr it can still be inaccessible to some people. We are constrained in our pricing because of our license agreement to use Windows, which eats up a huge chunk of that hourly fee. By moving to Linux, we will be able to offer the same performance for just $0.10-$0.15/hr. Although pricing for our Linux plans has not been finalized just yet, the prices above should give a good indication of where we expect it to be. We hope that this will drive usage for students (specifically those studying computer science or related fields), software developers and even web hosting services.
Second (and more broadly useful to all of our users): multi-region support. Currently, all Trelent PC’s are hosted in Toronto, Canada and have backups in Montreal, Canada. This is great for users in northern North America, but outside of that region, most people are out of luck. As a result, in the next release of Trelent you can expect to choose which region your Trelent PC will be created in, and even transfer your Trelent PC between those regions after the fact if you wish. We are developing the capacity to support 10 locations across 5 continents and 8 countries: Australia (Sydney), Brazil (Sao Paulo), Canada (Toronto), England (London), Germany (Frankfurt), Japan (Tokyo), South Korea (Seoul), and finally the USA (Ashburn, Phoenix, San Jose). Our initial roll-out, however, will include 3 of those: Toronto, Frankfurt and San Jose.
Continuing the journey
We hope that this post has been at least a little bit insightful for you. It is our goal to continue posting at least once a month as we learn and grow, and to share those learnings with the Medium community. As for this post, the bottom line is that we are working every day to build a better service, and will continue to do so for the foreseeable future. If Trelent sounds like it could help you, please feel free to learn more and setup an account at www.trelent.net!
