Stackmob announced it is shutting down this afternoon. I know a few folks that are using their services and are now going to be forced to look for a new provider for their backend services. I thought I’d share my experiences.
It has been nearly two months since I took over operations of Glassboard. The experience has been interesting, enlightening, challenging, and fun all at once. I’m hoping to talk about these lessons and experiences more here both as a way of documenting my progress and to hopefully garner feedback as I go forward.
I had an eye on last week’s announcement of Microsoft’s new CEO. Obviously because I follow the technology industry, but also because I now have a vested interest in Microsoft going forward. You see, Glassboard’s entire backend is powered by Microsoft technologies: Windows Azure to be specific.
Satya Nadella, the new CEO at Microsoft, was the head of the Azure side of MIcrosoft’s business before ascending to the top of the food chain, which is why I’m optimistic about the company’s future. I truly believe Azure is the best thing the company is doing right now.
Most people are surprised when I tell them that since I inherited an Azure backend rather than choosing it on my own accord, but I think that is what makes my thoughts more interesting. If I thought it wasn’t on-par with what Amazon is doing, I’d take the time to migrate off the platform over time. At this time, there are zero plans for that. I’m extremely happy.
The Glassboard API Architecture
Glassboard’s Android and iOS apps, like most mobile apps, are powered by a REST API. The API app is an ASP project that uses WCF to distribute the JSON payloads. If there’s one thing I hate, it is WCF. A lot. My next big backend project is converting from WCF to the new Web API 2 technology that is a lot easier to work with. It’s all written in C#, which I have found to be a pretty nice language.
The API is deployed to a Cloud Service, which is similar to a virtual machine but with less of an administration headache. I have the Glassboard Cloud Service set up to automatically scale between 2 and 4 Cloud Service instances depending on how much of a load is on the service at any given time.
Your messages are encrypted and stored using Azure Table Storage. Your photos, videos, and files are stored in Azure Blob Storage. Azure’s storage system is familiar to anyone who has used NoSQL. There’s also a little bit of SQL Server mixed in there for some stuff where a relational database made sense. This is by far the biggest cost point of my monthly bill, but it’s also pretty close to what Amazon is offering.
I also am using storage queues to communicate with Glassboard’s attachment processor and notifications cloud services. This works pretty well, but the Service Bus may be a better choice going forward. I’m honestly not sure yet. I need to read into it more.
Oh, and I also deploy the Glassboard marketing site to an Azure Web Sites instance through GitHub. Whenever I push an update to my web site to its public GitHub repository it instantly deploys the new version to Azure.
Other than the Glassboard marketing site, I manage all of this and deploy it through Visual Studio 2013 on a Windows 8 VMWare instance on my Mac. I came into it with zero experience and in two months have gotten to the point where I am comfortable with nearly every aspect of it.
Is it the right tool for you? That’s up to you, but like Brent I am a fan and am glad to see Amazon have a legit competitor in the cloud services area.
With Stackmob shutting down, I know a lot of developers are looking for a new backend as a service. If you’re doing Node, Azure is worth looking at as it is a first class citizen. If you’re looking at Parse as a Backend as a Service replacement, Azure is worth looking at too. Their Mobile Services product is a direct competitor. I haven’t used that portion of the platform personally, but I am happy enough with the Azure offerings I already use to recommend it.
As more apps are requiring a server-side component, cloud service providers are becoming more important. My hope is that Microsoft’s Nadella continues to recognize that as CEO and keeps funding the best product (and only product I regularly use) they have going today.
A few months ago I wrote a post titled What I’d Do With Glassboard that outlined what I thought the potential buyer of the service should do with it going forward.
Funnily enough, this post put me in touch with the NewsGator folks and I acquired Glassboard from them last fall. In the time since, I have been working nearly nonstop to get the acquisition and transfer completed on the business and technical side, while also working on a complete refresh of the iPhone app for iOS 7.
Today marks the day that I dub Glassboard to be a minimum viable product. What the Sepia Labs team built is a fantastic foundation for a slew of amazing ideas and opportunities I hope to build on in 2014 and beyond.
A lot of that first blog post remains true today, though I won’t be killing the web app. In fact, it’s getting a fresh coat of paint thanks to a good friend and his team.
In the short-term you’ll continue to see more updates to the iPhone app as well as bringing the Android app up to the same level of quality. The Android refresh is entering beta in the next few days even.
I’m excited about the future of Glassboard. I hope you’ll come along for the ride.
Download Glassboard 3.0
Glassboard is for sale. If you want to get past all the businessy corporate speak of the official blog, here’s the main nugget:
We want Glassboard to not only continue running but also to evolve and flourish. So we’re looking for a good home for Glassboard. The backend service for Glassboard runs in Azure, and it has wonderful iOS and Android clients. It took a team of talented people to create it, so we want Glassboard to go somewhere that has enough resources.
Brent Simmons has a pretty good summary of what it is and why it’s worth saving:
The problem of persistent, private, and trustworthy group sharing is still an open problem. Glassboard represents a couple years of work by a six-person team, and it’s a great start. I believe that it can be very successful, given the right home, given resources and commitment.
Brent has a bias in this since he was the former iOS developer over there, but there’s little doubt that the team that did work on it cared about it.
Glassboard has become the ‘conference’ app for better or worse. Every Macworld, WWDC, and independent conference seems to set up a board for evening activities. I’m also using it for other things such running betas for Second Gear apps.
I doubt Second Gear has the resources to acquire Glassboard, but if I were the one to acquire it, here’s everything I’d do:
Change The App’s Mission
Brent said this:
The problem of persistent, private, and trustworthy group sharing is still an open problem.
It’s not a problem that needs solving, at least not in the way they are marketing it. Nerds care about privacy. Users care about reliability, features and usefulness. Privacy isn’t a feature. It’s an expectation.
Market it like this:
Glassboard enables you to connect, coordinate, and collaborate with a group of people independent of platform or location.
That’s the usefulness.
Annotations
As part of that new mission statement, I’d put way more focus on collaboration and coordination. Add support for annotated messages beyond just images and video. Make it easy to embed a Google/iOS map with a specific place or your current location. Allow the user to link to a file that’s stored on his or her Dropbox account.
If you want to be super trendy, add some fun stickers that can spice up a conversation.
There are a million apps out there that do group chat already. None of them really do collaboration on mobile that well.
Rewrite the iOS app
The iOS app is serviceable, but not great. With iOS 7, I’d argue it needs a complete overhaul. Hire a designer, get to work and crank out a 1.0 visual refresh in 30-45 days after acquisition that targets iOS 7 and breathes fresh air into the product. Once you have the new visual design out there, focus on adding in the features like annotations or other metadata.
The Android app is pretty good. I don’t think it needs much of a visual refresh other than to sync up with visual identity.
Kill The Web App
The Glassboard web app is atrociously bad. They “needed” it, but the product they gave users was half baked at best. Kill it with fire and then bring it back when it’s fully baked.
Change the Business Model
I have been using Glassboard since it was first released and I have given them zero dollars. They came out with a “pro” plan, but the boundaries on it are all wrong. Unless you are a heavy Glassboard user, you will never hit the limits set. Good for users. Bad for business.
I’d charge a flat monthly fee per board. Base it on how many active users there are in that board on a given month and charge $x per head. Give away the first three users for free as the ‘trial’.
In the conference model that Glassboard is traditionally used for, the boards are throwaway so it’d be a yearly cost of $10 or $20 (I’m making up numbers). For people like me who use it for a beta with 25 or so people, I am putting money into the business on a monthly basis to keep it afloat.
Change the Name
For an app that touted itself as being heavily into privacy it was always curious to me that the name was prefixed with a see-through object like glass.
I’d instantly rename it Backchannel, but you could name it whatever you wanted. Glassboard doesn’t have enough of a following that you lose much awareness with the rename, but it does give you a fresh start out of the gate.