Trying to invent Gmail and Slack at Microsoft in 2003
A friend told me that Eric Horvitz’s group in Microsoft Research was seeking an engineer to complete an experimental communications prototype. Microsoft Research? Experimental communications? This sounded interesting.
The Grand Central vision
The project turned out to be an app called Grand Central, whose purpose was to explore new ideas for real-time communication. A unified GUI would integrate email, chat, voicemail and other modes in real time, including conversation-threaded mail, about a year ahead of Gmail, and several years ahead of Microsoft Outlook adding the feature.
At this time, the disparate set of interfaces for the various communication modes were starting to feel overwhelming even before adding mobile to the mix. It seemed obvious that we were all missing a communications tool centered around people and their conversations, that works across orgs, is nimble with ad-hoc group membership, and unifies diverse communication modes. This was so obviously the future it was literally a dream project for me. At age 6, before I’d even touched a computer, I drew up prototypes for a “home communications center” that integrated audio, video and text into the walls of a house. By 2003, I’d been participating in threaded online conversations for over two decades at that point, starting with USENET. Communications? Count me in!
This unified communications client was to be built using C# and Microsoft’s .NET framework, which seemed wise to learn at the time. A Microsoft researcher had gotten started but couldn’t do it alone. I came on board to make it fully functional, and get everything to synchronize with real Exchange servers. This would allow “dogfooding” the service within Microsoft, using it internally, which was seen as critical to gaining traction.
URLs and links and other structured information “nuggets” were extracted and indexed. Our project had the now-familiar chat bubble interface, before that was a thing. The first such interface I ever saw was the one I was building myself.
A window into Microsoft Research
The Adaptive Systems and Interaction Group I joined was doing fun stuff like full-text Inbox search, scoping searches to “Stuff I’ve Seen” and applying Bayes’ theorem to everything they could think of. My manager was fresh off inventing the concept of the “toast” popup for a brief notification, an afterthought on his project that got adopted by the product groups within five seconds of his demo.
This research group was much more open to experimentation than the product groups, and did have a bit of an academic air, unsullied by commercial concerns. Microsoft was not quite the same company I had left 7 years earlier. It felt sort of like going home, with many familiar faces on campus, who were now senior and rising up the ranks. As I was a former employee on a temporary contract, I was no longer a “blue badge”, but I was treated with nothing but respect. I enjoyed the deep dive into C# and .NET. I used “Managed C++” - not the prettiest thing - to get the prototype wired up to Exchange.
While it was great to work with great people on a cool project and get paid to learn, the actual prize we were going for was to spur new innovations in the next version of the big commercial products used by millions of people, including Microsoft Outlook. By that measure, the project was a success and a disappointment at the same time. Everyone had hoped for more traction within the company than we got. It was intuitively obvious to me in 2003 that a Slack-type service should exist, and Microsoft was in a unique position to make it happen.
Unfortunately, even though we got everything working, the Outlook team took a look, and basically took a pass. They just had other priorities. As a lowly contractor, I was not in the room to make the case.
Unrealized visions often die on the vine, but what do you call the organizational failure to see a multi-billion-dollar opportunity when it comes with a working functional prototype? What do you call it when a company pays people to spurn innovation that it pays others to do? And did anyone think to run some ideas by some actual users? Because actual users loved every single one of the ideas we demoed, when they were eventually delivered by others.
Fumbling the future, leaving billions on the table
It took Gmail proving that customers liked the conversation threading the following year for the Outlook team to get more interested. When Gmail was announced in 2004 with conversation threading as a marquee feature, I grimaced, and then signed up. Gmail helped spawn Google Apps (later G Suite, Google Workspace), which became a bane for the Office business. Gmail also helped spur the ad-driven business model that made Google into the industry titan it is today.
Then it took years more for the rest of the ideas we had prototyped to show up one-by-one in Microsoft apps or, more commonly, on other platforms like Slack and HipChat. For me, this experience felt surreal, like hearing various artists produce a stream of #1 hit covers of songs from an unreleased pop music album I played on years before that was just sitting in the can. Slack, which didn’t even launch until a decade after we presented the Office team with a working Slack-like prototype, was sold to Salesforce for $27 billion, after Microsoft had taken a pass at $8 billion.
This dynamic, a failure to commercialize home-grown innovation, was familiar to me even in 2003. I knew the apocryphal tales of Steve Jobs touring Xerox PARC that led to the development of the Mac. I’d steeped myself in case studies, like Fumbling the Future: How Xerox Invented, then Ignored, the First Personal Computer. Now I was seeing a similar dynamic from the inside. It was validating and frustrating at the same time to see ideas we’d attempted to pioneer commercialized by others later on.
As for me, I had fun, learned a ton, and worked with great people. But in the bigger picture, waiting for others to show the way and then offering them billions to buy them out is not why I got into this biz. I’d rather be on the other side of that equation.
For all of its commercial success, I look at Microsoft Outlook today, and to a first approximation it looks the same to me as it did two decades ago. It feels like another mammoth-sized missed opportunity. Microsoft Teams was conceived over a dozen years after this research project concluded, in 2016, after Slack had shown the way.
But hey, at least the toast popup shipped.
