How Microsoft Outlook Maintains A Startup Model Of Operation

This is a summary of Kevin Henrikson, Partner Director of Engineering at Microsoft’s talk at the 2016 MORE Summit. The full speech is provided at the bottom of this article.
Kevin Henrikson started out like many of us: as the co-founder of a startup. But when that startup was acquired by Microsoft and re-named Outlook, he developed a system for maintaining top-level customer service with a massive user base.
Step 1: Plan
Kevin and his team played with different sprint cycles for shipping on iOS and Android, and found that 7 days works well. On Android they now ship 2-3 times a week; with iOS it’s a bit longer due to the App Store review process.
Step 2: Sprint
What happens in those 7 days? A deck comes out at 7a.m. every Monday that breaks down the data, problems, and feedback from the previous week. Between then and roughly around 3p.m. the team fixes the things that are broken. Then they push the service to prod, submit to the App Store, submit to beta users, if telemetry looks good, they push it out. Then they do some sprint planning for the coming week at the end of Monday, and kick off. Throughout the week, they have regular standups in the middle of the day– much of it virtually through chat. On Friday they push to staging. They drink some beers, talk about what went wrong, what went right, and what they can improve (it can be personal or product-related) in a casual setting.
Step 3: Test
Where does all this informative data come from? Kevin’s team uses Testlio, a small team overseas. They give them the build at 6p.m. on Friday, and the overseas team tests Friday night, Saturday, and Sunday, and then sends them a Monday morning report. They also give Testlio all of their support tickets from Helpshift, and they then receive a report about what features people were asking about, what bugs were they looking at, and what happened with one star reviews.
Step 4: Feedback
It’s not just data that they look at, though. Outlook strongly focuses on end-user feedback. They look at reviews, feature requests, and conduct interviews with different groups of users. The three groups they focus on are heavy users, light users, and churned users, and they ask each group questions about the Outlook experience. For companies, they ask why certain employee bases are using one feature or another. They ask about product ratings. They ask about any usage variance that could be a reflection of a product issue. They also take in support requests themselves (which is unusual for an engineering team). Through all of these measures, they end up having a clear picture of who their users are, and what they want. In the same way that a startup CEO can talk to every single customer, Kevin has built his team so that through customer feedback samples and regular data he can also build a picture of his client base.
Kevin has managed to move from a startup into one of the biggest corporations in the world, and maintain all the good parts of a small team. Below is the full footage from his MORE Summit talk: