Time to think about the long way we've come with fruux and how to go on...
We had severe downtime during the last days and currently our service is also not available, while we are trying to bring the systems back online by implementing certain optimizations. As an immediate action we had to disable new sign-ups a few days ago. We‘d love to grow our user base further, but with our current infrastructure we can‘t even provide our current users with the reliability they deserve. Unfortunately it‘s no technical problem - we could easily add a few more servers to our cloud to handle this capacity problems, however there is no budget for additional servers, but more on that later.
Right now our systems are busy executing these tasks and we're trying to relax while monitoring the whole process which will hopefully allow us to bring the systems -at least in a limited fashion- back online really soon (as in one or two days not weeks). It's one of those typical Sundays, except it is not, because this Sunday will either be our game changer and the beginning of something great or the day where we saw the end of the road in retrospective (If these lines sound familiar, you probably read the excellent blog post by Todd Agullnick, Co-Founder of Xmarks Inc., with a very similar story to ours).
Time to think about how far we‘ve come with fruux.
We started as an entirely bootstrapped student project. Initially we never thought about fruux as a product. We just started developing fruux because we desperately needed such a piece of software ourselves. During this time fruux eventually spread amongst family, friends, fellow students and ultimately people we didn't even knew in person. We started getting tired setting up accounts manually so we finally implemented a self-signup into the application and since then our user base keeps growing at crazy rates which makes it hard for us to constantly scale our infrastructure to keep up with this growth. First and foremost it's hard for us because it's a non-technical issue. We've really learned squeezing every single bit of computing power out of our cloud and whenever we hit conferences and talk with people from the industry they are completely blown away when we tell them how little computing power we actually have, but the time has come where we have to scale up big time to keep up with the load and provide a reliable service. Unfortunately nothing in life is for free...
As many of you know, we are only able to work on fruux in our free-time since we graduated. Our full time jobs help us funding the infrastructure for fruux, but also prevent us from investing serious time into the further development, which is kind of a chicken-egg problem.
The first step towards a solution for this issue is to ensure that we have the resources to scale our infrastructure (i.e. deploy more servers) up to a level where we are able to provide a very reliable service. We calculated that this infrastructure will cost us about 700 EUR (approx. 950 USD) per month, but we are unable to fund this personally and we won‘t be able to fund it through donations or ads, so the only option we have is to generate revenue by actually start charging for our service.
Our idea is to provide paying users prioritized syncing as the first "pro" feature. This way our paying users will always enjoy a reliable fruux. Once we actually start selling subscriptions for "fruux pro" it will also become much easier for us to plan the future development of fruux, implement great major new features, jump ship from our current jobs to full time fruux work and ultimately even hire additional engineers in order to bring the project up to speed. Based on your feedback, we discovered that about 25 EUR (approx. 34 USD) per year should be a fair deal (like paying for one "fruux caffè latte" per month) for our service and this way we would need roughly 400 subscribers to cover the infrastructure costs including related taxes (this pricing is based on your feedback and on a cost per user approach to ensure scalability). If we are able to achieve this number of subscribers, we'll be able to provide you with the most reliable fruux ever and also have an excellent argument when acquiring investors which will allow us to transform fruux into our full time jobs (=devote much more time to the project then we are currently able to do), hire additional engineers and provide you with the kick-ass features you have been waiting for for quite a while now (hint: iDevices support, a webapplication, a developer API, you name it...).
We thought about starting this by setting up a binding sign-up form for "fruux pro" and once 400 of you signed-up we'll setup the new infrastructure and start charging your accounts. If we are unable to convince at least 400 of you, that a reliable fruux is worth paying for, we'll probably have to sit here and wait until the load continues pushing the cloud over capacity and ultimately move on and pull the plug as a last resort, because in it‘s current state the project is neither a pleasure for you (due to the capacity issues) nor for us (because the solution is technically very easy, but not fundable).
We'd love to hear your feedback!
We're trying to achieve a "hybrid solution". In order to bring the service back online as quickly (and as reliable) as possible we'll pony up the upfront costs to scale up (at least a bit) right now and then start working on the implementation of the billing infrastructure. Then we'll hopefully be able to convince enough of you to become subscribers so that we are able to stay online (=build a sustainable business) with the additional computing power (and scale up even more). If you feel like supporting us in the meantime, you are happily invited!
We're super stoked about your overall positive feedback and the dozens of emails you sent us so far!
The following questions came up multiple times:
Q: Is your pricing final? A: No. We just calculated what would work to cover costs and also allow us to scale, while estimating a percentage of users we might be able to convince to become actual customers.
Q: Where is the sign-up link? A: Our first priority is bringing the systems back online (work in progress). Starting tomorrow we'll dive into the ugly paperwork for PCI compliant credit card acceptance and also start implementing the billing system.
We're making good progress with the tech part of our challenge. If nothing unexpected happens, we should be back online in a few hours.
We're up and running again (@fruux_status).
Unfortunately there is a service disruption at Amazon Web Services (our datacenter) causing downtime for fruux. We're working closely with the respective engineering team at Amazon Web Services to resolve this situation. Check @fruux_status for updates.
Update 2010-10-11 #1
Thanks to the nice guys at Amazon Web Services (who are obviously working like us on Sundays) we are now able to access and migrate our data, so we should be back online really soon. Also we scaled up even more for reliability reasons. Due to the increased costs this however sets us even more under pressure to start generating revenues asap because we don't have the budget to keep the service running at this scale. Check @fruux_status for updates.
Update 2010-10-11 #2
Initially we were set for a quite Sunday in order to prepare for the paperwork stuff and billing implementation during the coming week, but obviously Murphy's Law struck once again... Nevertheless we'll work hard during the week to overcome all remaining obstacles on our way to sustainability. We'll keep you posted on the implementation of the billing system, the planned pricing (we're reading all your comments) and the progress of the negotiations with our bank for the credit card merchant account etc.
Update 2010-10-11 #3
We are up and running again. Thanks a ton to the nice guys at Amazon Web Services for helping us fixing these issues - even yesterday (Sunday) evening! Now back to the billing implementation and getting the necessary contracts for credit card acceptance signed.
Update 2010-10-15 #1
After a week of negotiations and talks with credit card companies, banks, payment gateway operators, and outsourced billing providers we've come to the conclusion that credit card acceptance is a total mess and utterly expensive. We strongly believe that we shouldn't burn (to much) money just to make money, because then we'd have to charge a higher price for the additional processing costs and involved work. Thats why we've decided to implement with PayPal for the start. We hope to complete this implementation (we're about 50% done now) over the weekend and launch paid accounts for new subscribers starting next week. We'll stop offering a free version of fruux for new users for a while and instead concentrate solely on the paid version and additional features, but we'll relaunch fruux free soon, once we've delivered some major new features to our subscribers. Once we are confident with the payment process, we'll implement it for existing users and allow them to migrate their accounts.
Update 2010-10-15 #2
We aren't entirely sure about the final pricing yet, but we decided on the product structure. The good news is, that we won't offer fruux free, fruux starter, fruux home basic, fruux home premium, fruux business, fruux enterprise and fruux ultimate. Instead we'll just have fruux pro and (soon) fruux free. Pricing will be around 20 € including VAT/sales tax.
The bad news is that we didn't manage to complete the implementation of the billing system over the weekend, but the good news is that we've made good progress and shouldn't be a whole lot behind our initial schedule. We'll probably need a few more days, but the implementation will be better (completely automatic account activation, deactivation and renewal based on data from the payment processor), too.
We've finally implemented the billing system during the night. At the time being it only applies to new users signing up with fruux. We'll evaluate the technical aspects for the next couple of days and also keep an eye on the signup metrics. Our current pricing model for this test is one week free trial and then 19 € per year. About one hour until sunrise, time for a short nap... :-)