Chapter 2: The Microsoft Band Team
On October 29, 2014, it was the day before the launch day of Microsoft Band. Our manager told us to skip work and greet other customers at the Microsoft Store at Bellevue Square. During the same week, marketing team had finished planning their seeds and tech bloggers had already published their sneak peak articles for the hype. By the launch day, it was a scene that most Microsoft fans had not seen for a long time — People lining up in front of Microsoft Store. Not Apple Store. It didn’t matter which part of the team we were from. We were all gratified for working on a project that meant a lot to these customers.
CHAPTER 2: The Microsoft Band Team
I used to play a game a lot in my head whenever I was tasked with an unfamiliar subject. During the second phase of my career in the company, I had the opportunities to work as a variety of roles — tech lead, project management, DevOps, design, etc.. Am I doing what my manager once did for me? Are we making the same mistake? Am I being that engineer who was so patient and always helped me with software design? Am I doing beyond what was expected to consider flexibility in design change? Am I becoming that Senior Engineer who openly called me an idiot in a group e-mail? When in doubt, picture a similar situation back on the Band team, what would the team do in this situation? And then think why they did it in the first place.
Microsoft Band was a mix between a smart fitness band and a smartwatch that debuted in October 2014. The device supported all popular mobile platforms at the time including Windows Phone, Android, and iOS. It was noteworthy for the number of sensors that the product team could jam into a tiny wearable device. It was a small-time release within an incubation team in Xbox Accessories that very few people knew about. The product enjoyed huge interest from Microsoft enthusiasts but ultimately suffered a critical quality issue in its electronic component. It doesn’t look good on Wikipedia, but I had a different experience within, and I believe many startups share similar stories.
The tides against us
The Band organization was small, but the product team had to reach externally for a talent pool that was unconventional in Microsoft. They needed people in mobile, cloud computing, firmware, health, fitness, algorithm, hardware, embedded system, etc. This was a lot for a team of less than a hundred when it first started.
Under Steve Ballmer, CEO of Microsoft, this was unheard of, as the company had been on its crusade and at the peak of the Windows Phone war. In the same year, hundreds of employees in Windows Phone were working around the clock shipping new features to compete against iOS and Android. Terry Myerson, the executive vice president of Windows, disclosed the acquisition deal of Nokia’s phone division. Microsoft won its Windows Phone platform battles in a few countries and regions, but The Band team was building a product for our competitors! (Tech jargons alert! feel free to skip the bullet points if technical jargons aren’t your thing.)
- From the company culture, Microsoft had not set its mission to “Empower everyone on the planet.” Although the company had its infamous peer review recently removed, the general alpha-male culture still existed. There was no Standard of Business Conduct training like Nelson’s TV drama. Microsoft later transformed its employees’ training into several seasons of TV drama to boost engagement. It was Netflix-worthy and the best excuse to eat popcorn at work.
- In technology-wise, Microsoft was the furthest away from the idea of open-source. Microsoft was a close-source company, meaning If the company wanted something, they build its own. Git version control was unheard of and we relied mostly on Source Depot and Team Foundation Server to maintain our source code. The worst thing was that they only worked on Windows! The plugins on Mac barely worked, so the few iOS/Mac developers had to duo boot in Windows just to push our code. It was a productivity nightmare.
- Serverless DevOps was years behind us so we also managed stacks of Windows servers and Mac-mini for CI/CD in our offices, and each of those machines had USB cables attached to Nokia, iPhone, Samsung Galaxy, all kinds of smartphones, and our smart wearable devices. We needed these setups to automate our tests. Our few build and test engineers single-handedly built the entire infrastructure that more than nine teams depend on.
- The tiniest version of Windows OS itself was half a Gigabyte running on Intel x86 or x64, and our wearable hardware had a ARM chip and only had 64MB flash storage, so we had to custom-make the whole OS and its shell.
- Neither Xamarin, React.Native, nor Flutter had been debuted to the public. These are some of the modern frameworks that allow engineers to create applications on multiple platforms with speed(iOS, Android, Windows). We did not have any of that in-house yet we were developing on Windows, Windows Phone, iOS, Android, and Mac OS X. The worst part was we could not ship features that made Windows Phone look bad. The best feature had to be on Windows despite Android and iOS platform SDK maturity being at least two years ahead of Windows.
From the bigger company culture, politics, resources, infrastructure, hardware, software, etc., the tides were against us. We overcame all those obstacles and shipped our product. That’s how difficult it was, and that’s why we had so much pride.
The team in flow
The people on our teams were extremely talented, diverse, and energetic but laid-back at the same time. We had an .NET engineer with only one hand, and he still wrote code faster and more robustly than most of us did. Android app team had a pastor that could tell you that your idea is bad while making you feel great about yourself. The iOS app team had an engineer who had been a car mechanic until he hurt his back and decided to build mobile apps instead. He is now a Principal Engineer. Our test team had a guy who wrote the first Starbucks iOS app. The firmware team had a young college grad, and she self-taught Bluetooth and single-handedly implemented our Bluetooth stack. The firmware UI team loved snacks so much that their office had two tables full of snacks that are there for anyone to pick it up and go.
My mentor was a veteran mobile development consultant. He was so good that Microsoft attempted multiple times to retain him when his contract expired, and he declined the offers simply because he preferred the freedom of being an independent contractor. One time, he just shoved iOS development in my face while he jumped to Android team to rewrite their Android Bluetooth code. The Android consultant was fired, and once my mentor had finished his rewrite, the memory bugs were gone, and the performance tops the benchmark among other platforms.
We had a budget cut on the entire product, the first to go was us contractors. But instead of laying off contractors, our managers stood their ground and defended us. Together, they found an alternative solution to retain all of us. Most contractors switched to a new firm that was transparent how much they make on commission. And the contractors ended up getting more take-home pay, and the company also meet their budget goal. It was a win-win! By then, my manager had already known I was given almost nothing after the cut from my firm. My manager renegotiated the price-cut with the representative of my firm and told them “you do not cut Chao‘s pay!” I had always thought my manager was kind, but he leveled up to simply “badass” at that moment.
The management team worked very hard to allow the development team to achieve their maximum productivity. It wasn’t until two years later, that I learned this methodology to be called Agile. We had clarity in our tasks, the teams were adaptive, and we were open to communicating with anyone on the teams. We focused on the features delivered and not on who takes the credit (at least for a while). It felt top-down at times, this will always be inevitable if you work on a hardware product, but half of our software features were once ideas from someone on the team, and people continue to bring their ideas to life.
Even managers could not resist hacking around projects to make something cool and useful. It was like a whole organization achieved in Flow. People who had transferred from other teams in Microsoft to the Band team had made the same comments — “I have never seen this in Microsoft.”
The Price of Hit Refresh
In 2014, Satya Nadella became the new CEO of Microsoft and later published Hit Refresh in 2017. The whole company was at its dawn of cultural shift and the new focus on “Mobile first. cloud first.” MSFT stock surged, and Microsoft had a massive layoff cutting more than 12,000 jobs. Software Development in Test was no longer a position offered. Algorithms selected test engineers to be let go, and the remaining test engineers were converted to Software Engineers, Software Engineers were then expected to own testing as well.
Testing your code? That’s a no-brainer to many software engineers today. It was a great move I applauded because I genuinely believed in that principle, but back then, engineering managers had little time to adapt, and engineers had an even worse time adjusting especially if their former test engineers were laid off! I was lucky, my test lead was still nearby, and I had him on my code reviews.
It may be a godsend in Microsoft’s cultural-shift overall, but it also created a ton of quality backlash. When you make a mistake in a software product, chances are, you can patch it over the air, and customers love it again when you fix the bugs. When you make a mistake in a hardware product, customers want their refunds, and you not only lose money but also create a colossal train of branding and post-sales service problems. You simply can’t make an unrecoverable mistake. That’s why hardware quality requires completely different sets of expertise in engineering, and it is often so complex that might even involve robotics or something beyond our imagination. Within two years of change, Windows’s quality tanked, and Surface had its infamous battery swelling problem. And the head of Surface re-established a quality engineering org to rebuild the quality talent it desperately needed. That was the price of “Hit Refresh.”
Retreat in wearables
By 2016, the Band team was on its way to its third hardware iteration. We believed we had fixed all the problems that consumers complained about in the last two iterations. Apple had already invested billions of dollars in this industry. Google, Samsung, Fitbit, Pebble, etc. followed with their millions and millions of investments.
Peter Thiel’s ZERO to ONE had already been New York Times Best Seller. In his book, he examined a case study in the Search Engine war between Google and Microsoft. The result was that both Google and Microsoft lost and let Apple create an unforeseen monopoly opportunity in iPhone. In his startup notes to entrepreneurs: avoid copying; avoid competition.
Microsoft had learned its lesson in clashing with competitors unprepared, and the company needed to switch to a different mindset — fail fast learn fast. Although Microsoft Band had cost the company a relatively small amount of investment, it was still the time for C-suite to decide where future leads. They made a difficult strategic decision based on the fact wearable tech did not align with how they had expected the consumer market would behave, which was the cold-hard truth we all had to face.
A few years later, the smartwatch hype had cooled. The big names remained. Apple leads the market share, Samsung follows, Pebble was swallowed by Fitbit, and Fitbit was acquired by Google. Microsoft focused on their Azure bet and saved themselves potential millions of dollars of losses in wearable tech.
HBO, Silicon Valley Replayed
In the summer of 2016, we were invited to a large theater-styled conference room with everyone on Band team. Panos Panay, the CVP of Surface at the time, was giving a spiritual speech on how our team came together and deliver such amazing products and culture that the company had not seen. He praised the Band team as the team was running its own company.
All of a sudden, I could not help but remember the hit HBO series Silicon Valley episode that I had just watched the night before — Gavin Belson, the founder of a fictional tech company Hooli, had just fired the entire Nucleus team in a conference room. When I came back to reality, Panos had just delivered a statement that the Band would be disbanded.
The next day after Panos’ meeting, the usual calm and warmth in the office of Building 86 had gone cold and silent. People started gossiping, they stopped working, some started looking for an internal transfer or outside the company, and others began waiting to receive their severance package so they could leave with financial stability.
It was the first time I had witnessed a whole organization axed in Microsoft. We had interns who had just flown over to Redmond and onboarded their roles a few weeks ago and were ready to start changing the world. We had people recently transferred from other big-name teams like Windows, Xbox, and Hololens. We were all upset. It was the most depressing day I had seen in the office. The Band team was dead.
Note: As a satire, if you have never experienced a whole organization shut down, here is a recap of Silicon Valley:
Silicon Valley — Gavin firing Nucleus — YouTube
Final Save
I was more in shock than anything else. I could not face the reality; instead, I hid in my office designing a solution on how a user can interact with notifications on our smartwatch sent from iOS, Android, and Windows phones. It was the first design document that I have ever written, and I had spent almost a month working on it.
I figured out how to offer this as a public API, allowing other app developers to support their app to design interactive notifications such as the simple canned response of a messaging app. This feature didn’t exist at the time, let alone an interface design that could work on all mobile platforms, and I was in my proudest moment to celebrate this achievement in my office.
I walked to the next office to grab one of the firmware engineers to consult him on whether the firmware payload structure I designed would be enough to be sent from smartphones. Everyone was awed, but not because of my achievement. One of them eventually asked “Chao, why are you still working? There is no more Band.” “I know but I still wanna know,” I said.
The firmware guy was dumbfounded by my genuine curiosity, but he still helped me to confirm the final few assumptions. I was finally able to finish the design specification on interactive notification. I went back to my office, updated the document, and hit the final save. I proofread my document page by page, again and again until a defeat pounded in my heart. I stopped and gazed at my monitors for as long as I could remember.
That was the last time I worked on something that I truly had a passion and purpose. Everything else turned to addiction, and more on that later.
To be continued…
Update:
I was very surprised and happy at the same time to see others who’ve worked on the product also felt nostalgic reading this article. The article was briefly shared on Reddit where fans of the Band were recalling how they loved or hated the product, and I was able to reconnect with a few familiar faces on LinkedIn. Thanks for the support!
If you like this story, please clap or follow me for more. I plan to publish one chapter every or every other week.
Chao S.
Ex-Microsoft Senior Software Engineer
