I was Employee #4 at shomi. RIP shomi.

I think I was the fourth employee at shomi when it all began in 2012.  I was saddened to hear yesterday that they’re closing up shop at the end of November.  I have friends still working there, and given my previous role there as Software Development Manager, I had high hopes for shomi.

I was the Application Development Manager at shomi from the very beginning – early 2012, until May 2014, when I chose to leave.  When I left, we had been up and running for six months and had just released our first post-launch major upgrade.  I left behind a team of 15 developers that I had directly hired and managed, and a lot of additional colleagues I also liked and respected.

The leadership team at shomi had some rocky times (I’m sure all leadership teams do), but this was made more difficult by some thorny personalities on that team.  I couldn’t agree with the new direction they were proposing for the development team in 2014, so I thought it best to get out of the way and let them remodel development how they wanted.  Over the following 18 months, almost all of the original developers drifted away to other gigs as the changing leadership direction impacted their work.

I haven’t worked there since 2014, but I can offer my opinion one why shomi didn’t make it.

In a word, the problem was “cost”.  Content costs big big bucks, and when shomi is down south bidding against Netflix and CraveTV for Canadian rights for movies and shows, the prices get high. If your subscriber base isn’t big enough to distribute those costs widely, you can’t make money as a video service.

Meanwhile Shaw (www.shaw.ca), the partner with Rogers in shomi, sold all its media assets to Corus, leaving its 50% stake in shomi it’s sole remaining media asset.  Reportedly as part of that move, Shaw wrote off it’s stake in shomi (https://cartt.ca/article/analysis-steady-summer-shomi-speculation).  While the write-down is largely an internal booking matter, it clearly signals that they won’t be throwing any more cash to shomi to cover ongoing operating losses.

Another area where cost may have been a factor was in software development and deployment.  When I left, the development team was 15 people (QA not included), and our public website ran on about 30 mostly smaller VMs in a data center we were leasing space in. (Note that this does not include the video processing and streaming infrastructure – it’s only includes the shomi.com web site and the content management applications).   

Since I left, they’ve added five or ten positions to the development team (many non-developers), and they’ve moved the original 30-VM infrastructure into Amazon’s EC-2 cloud, where they now use well over 100 VMs – beefy VMs that can cost as much as $5 or $10 an hour.  Sounds cheap, until you have a hundred or two.  I have no solid numbers, and the “beefiest” VM I see on EC-2 today costs $13/hour, so let’s use 150 VMs at $5/hour just to see where it goes:  $750/hour for a day gives us $18,000 / day.  I wish I knew what the costs were for the original hardware, but I’m guessing it wasn’t eighteen thousand dollars per day.

So, I’m pretty sure they didn’t save any money by moving the cloud, and I recall warning them of that before leaving.  The reading I had done at the time had led me to believe what the experts from Google’s cloud team had told us in the early days:  If you don’t design your application for cloud, you can’t save money by moving an existing design the cloud.  In fact, it will most likely cost more.

Expensive content, not-exactly-booming membership, growing bloat in personnel and infrastructure make the math pretty straight forward – shomi wouldn’t turn profitable for some time, if ever.  And when Shaw stopped contributing money for operating costs, Rogers was left with The Decision.  And it was a sad one for me.

RIP shomi, I had a wonderful time in those first couple of years!