Dubman.com/narratives

Launching Microsoft Office into the cloud

The introduction of Google Docs must have a lit a fire in Redmond, because Microsoft launched a top-secret mission to create an online hosted version of the Office tools with a decisive urgency. Through my professional network I learned of a project management opportunity to help make that happen. Microsoft needed to hustle and produce a web-based administrative and support interface to new, hosted services for use by subscribing corporate customers. At the time, due to the particulars of how the company was organized, there were essentially no available people with the requisite skills to build this front-end in the organization responsible for the hosted services, and plans must have come together fast because they didn’t think they would have time to hire and spin up a new internal team. I’d worked on Office before it was Office, and had I been running the zoo, I would have kicked off basically the same project literally a decade earlier (1994) as the web was emerging, but now Microsoft was scrambling.

Hence Microsoft, very quietly on this hush-hush project, reached out to a design firm they were already working with (on non-engineering stuff like Windows Vista icons) to see if they could help build the UI. That design firm, FILTER, ended up reaching out to me, as this was a bigger engineering lift than their typical projects, and it turned out that I was well suited to help them. Frankly, my own coding skills were a bit rusty at this time… the rustiest they have ever been as I had come off several consecutive years of non-developer roles. At that particular time, in 2007, I would have been unsure how to crank out that UI myself the way they wanted at the pace they needed it done. But it was not my coding skills, per se, that they were after.

Launching a hiring spree

Instead, what the group manager wanted was for me to go on a tech hiring spree and staff up a team of developers and testers who could craft something they weren’t in a position to implement either, an uncompromising UI that resembled a native app, on the web, using only what could run reliably inside of the famously buggy Internet Explorer 6. Fun!

A tech hiring spree is something I had sort of done before at Microsoft, when it wasn’t my entire job. I had years of college recruiting experience involving resume screening and interviewing, plus years of interviewing more experienced candidates.

The size of the team (16) I had to build was similar to the software engineering class I taught at Yale. I had to sell candidates on the role without telling them what it was about. For the first time in my career, I didn’t totally understand all the code I was looking at over the shoulders of my team, and was too busy at my job to learn it on the side, but my developer background remained a superpower I was able to employ on occasion to good effect. I did have to constantly resist the urge to jump in and just solve problems myself instead of the more scalable and effective solution of hiring enough developers and setting them up for success.

Growing in more ways than one

The contract became the biggest in the history of the design firm, and I had to renegotiate that contract three times. Every time that got run up the flagpole I got the sense that some higher-ups at Microsoft thought this project was really, really important, so important that FTEs would sometimes whisper to me, do you know how important this project is for Microsoft? I could only imagine; nobody told me the business side. After increasingly stressful conversations, the higher funding always got approved.

At one point, I had to let someone go whom I had hired, which I tried to do with maximal grace and humanity. Microsoft at this time (as in the 90s) was not a warm and fuzzy place to work. The full-timers were quite stressed. I attempted to insulate the contractors I hired from the worst of the stresses my manager laid on me. I was doing okay; I was working hard but sleeping at night and not going bug-eyed in front of a debugger at 3 AM. I monitored everyone’s mood and tried to avoid driving the contractors I had hired into the ground as I had experienced myself earlier in my career.

One of my contract testers left while we were in ship mode to take a full time PM job at Microsoft, which was hard for me to argue with. When I had to replace her immediately to keep the project on track I presented Microsoft with a secret weapon in the form of my brother Arthur, who ended up being invited to lead the test team on day three. Towards the end I needed one more experienced developer fast, and hired the once-and-future Microsoft developer who’d previously hired me as a PM at GiftSpot.

Launching on to the web

The team I built worked helped to invent techiques, later deployed widely, to leverage AJAX to craft an app-like experience on the web. My Microsoft manager for this project was Tom Keane, who was ace, top notch, and went on to be a Corporate Vice President for Azure Global. I always felt personally respected and professionally valued even when he was kicking my ass, and he got the best out of me that I was able to deliver.

Microsoft launched Business Productivity Online Suite, BPOS, the awkwardly-named, corporate cloud-based predecessor of Office 365 (much better branding!) that would launch three years later, which is now Microsoft 365, which billed $20 billion in FY’20. (You’re welcome!) I was proud of my own ground-level contributions, even though I didn’t write a line of code. But I missed writing code.