In 2007 Paul Graham, the founder of Y-Combinator, penned an essay titled Microsoft is Dead. The thesis of the article was that Microsoft had become irrelevant thanks to the combination of the internet, Mac computers and Google’s services. In fact at the time, Y-Combinator didn’t bother inviting Microsoft to its demo days to meet any of their startups, but did invite Yahoo which at the time was a more relevant company.
I was at Microsoft from 2002 until 2019 and pretty much agreed that Microsoft was essentially dead as an innovative tech company in 2007. Under Steve Ballmer, the CEO at the time, the company was good at extracting money and building software for business customers. However the company also tried and failed to build any compelling products in every new space it tried. The company tried and failed to build a compelling mobile operating system, consumer email service, social networking app, search engine, or music services. It also tried and failed at building hardware with the failure of the tablet PC or Zune to take hold.
That entire time period ended up being described as Microsoft’s Lost Decade. I remember as an employee during that decade noticing that the stock price was always $27± $2. It was never much lower than $25 nor did it go past $29 for my first decade as an employee. The stock price was around $34 the day Steve Ballmer announced he was stepping down on August 23rd 2013,
Satya Nadella took over the company on February 4th 2014 and immediately began working on changing the culture and business performance of the company. The market, employees and customers responded positively to his changes. Today Microsoft’s stock price is $410 and has outperformed the NASDAQ Composite by 3x since the day Satya took over. His transformation of Microsoft has made many describe him as the most successful CEO in the tech industry.
So the trillion dollar question is how did Satya transform Microsoft? What follows are my observations of the key changes he made at the company that completely changed its trajectory.
Lesson #1: Culture Eats Strategy for Breakfast
The first thing Satya did was immediately start working on changing Microsoft’s culture.Satya believed strongly in the mission of the company as shown in his first letter to employees on becoming CEO but he also realized that the culture of the company was holding Microsoft back.
The defining characteristic of the culture under Steve Ballmer was complacency and a sense of hubris. Despite failure after failure, many people at the company acted like we were the best thing since sliced bread because we had Windows & Office, two of the most successful software products ever built.
Satya worked directly to challenge that mindset by asking everyone at the company to embrace a growth mindset. We were to be a company of learn-it-alls not know-it-alls. People were expected to be curious, embrace change and try new things. There were stickers in every conference room that contrasted a meeting that was a fixed mindset meeting versus a growth mindset meeting.
Everyone at the company was encouraged to try new things, learn from failure and refuse to be complacent. Satya led by example here by changing a number of ways we did business as the company had become stuck in its ways. This was a transformative cultural change that started at the top but was felt by every individual team at the company.
Lesson #2: Forget Strategy, Do What’s Best for the Customer
Microsoft is a company that believed a lot in strategy. Everything corporate move was a chess move to drive towards creating some durable competitive advantage. This had worked brilliantly for years. Office and Windows had become a virtuous cycle where companies bought Windows because they needed it to run Office at its best and bought Office because it was the best productivity software for Windows. The enterprise business grew like a vine as once you got Active Directory and Exchange in the door, you could pull in all sorts of other Microsoft products such as Sharepoint which rapidly grew to becoming a billion dollar product despite being widely panned both within and outside the company.
However a lot of this strategy was also user hostile. Windows 8 which was a failed attempt to turn desktop Windows into an iOS style operating system was driven by the idea that we had to prevent the iPad from disrupting Windows and so a touch tablet that could also run Windows desktop apps would be the best way to stave that off. Being able to or even trying to actually execute a usable product was never a part of the equation.
Microsoft also eschewed bringing its productivity apps to the web or mobile despite the fact that it was already quite clear that they had both won and the Windows desktop was losing.
Satya signaled to the company and our customers that those days were over within a month of becoming CEO by announcing Office on the iPad. This was a product the Office team had been working on for a while but thought they’d never get approval to ship because it was “off strategy”. Other customer centric moves followed such as Satya declaring that Microsoft Loves Linux a few months later as a clear signal that Azure would now treat hosting Linux VMs as a priority as opposed to being focused primarily on hosting Windows VMs.
Microsoft, a company that made popular the concept of strategy tax, where integrating one’s products at the cost of making it more difficult for one product to be individually successful abandoned that philosophy under Satya. Office no longer had to ignore the web or iOS because it made Windows less valuable. Azure no longer had to ignore Linux because it made Windows Server less valuable.
Lesson #3: Sometimes You Just Have to Cut Your Losses
Steve Ballmer was obsessed with competitors and no price was too steep to pay when chasing after a competitor that Microsoft was focused on. The company lost billions for years on Bing chasing after Google, spent over $6 billion buying aQuantive after being spurned by DoubleClick which sold to Google for $3 billion only to have the aQuantive acquisition written down to $0, and acquired Nokia for $7.2 billion to chase after the iPhone.
When Satya took over the company, it was quite clear that Windows Phone and Nokia Lumia devices were a distant third to iPhones and Android phones. Satya wrote down the entire cost of the acquisition and laid off almost 26,000 Nokia employees over two years, 18,000 in 2014 and an additional 7,800 in 2015.
It was quite clear that there was no path to meaningful success for Nokia’s smartphone business and so Satya ignored the sunk cost then jettisoned the business. Throughout my time at Microsoft, I observed similar dispassionate winding down of businesses where there was little value in continued investment.
I was a big fan of the Microsoft Band, which was Microsoft’s competitor to the Fitbit which was wound down in 2016. I remember asking Satya about this in a leadership Q&A and he responded that the business model wasn’t great and even Fitbit the market leader at the time, didn’t have a great business model. Satya turned out to be right as Fitbit was acquired by Google for $2.1 billion amid uncertainty about the company’s future.
Lesson #4 Every Investment has an Opportunity Cost
This is more of a lesson from Amy Hood not Satya. Amy isn’t well known outside Microsoft but she’s the most influential person at the company besides Satya. Amy is CFO and runs a very tight ship when it comes to spending.
I used to work on Bing Ads and a constant question our teams used to ask was why despite being able to go head to head against Google search relevance in the US, our product performed so poorly in other markets like Latin America or Europe. It turns out that Amy Hood was the reason.
In a leadership Q&A Amy shared that her philosophy was to always look at the opportunity cost of every business request from a business group. So when the Bing team asks for $1 billion to increase search crawling and train AI models to improve search relevance in Europe & Latin America, she asks how much money that would make versus spending that $1 billion on data centers for Azure. Given that acquiring advertisers and tuning the ad algos would lag any increase in usage from improved relevance in the consumer search experience, the ROI has always been better to spend on Azure.
As a person on the Bing product team, this sucked and my friends who still work there today still complain about it. However as a business decision and a Microsoft shareholder, I am quite happy with Amy’s decision.
It should be noted that this is an approach that should be used sparingly since in some cases investments that do not return as quickly as other opportunities could be strategic.
Lesson #5 At a Software Company Engineers Are Your Most Valuable Resource, Listen to Them
During the Steve Ballmer era, Microsoft forgot its roots as a software company as its CEO came from sales and not the product teams. This meant that certain practices that seemed quite obvious to product teams never got greenlit under Steve Ballmer. When I worked on cloud services like Windows Live and Bing Ads, I used to marvel at the fact we used Microsoft software such as SQL Server which at the time was designed for B2B use cases instead building or leveraging open source cloud based solutions like everyone else in the industry was doing at the time.
The company was also hostile to open source which combined with the above usage of Microsoft B2B software to build cloud services was like trying to run a race with your legs tied together.
Under Satya some of this slightly improved at first. You could use Open Source software if you got it audited. I went through this process when I was the group product manager for Bing Ads where I got approval to use services from WalkMe to create walkthroughs for our advertiser tools. However I quickly found the process to be Kafkaesque. I needed to pay $20,000 out of my team’s budget for the software to be audited and then it was only for that specific major + minor version combination. So if a new version of the software package was released next week, I’d need to pay $20,000 to get that audited and certified as well or hope some other team at Microsoft had.
After spending $20,000 to get JQuery certified to use on Bing Ads because WalkMe depended on it and realizing I was going to need to find another $20,000 whenever the library was updated, I sent an email to Satya about how ridiculous the process was. He asked members of his team to follow up with me and they mentioned they’d already heard some complaints about the process. I gave them a bunch of feedback and shortly afterwards the software auditing process went away and Microsoft started leveraging Open Source to build its products just like every other tech company.
This greatly improved the productivity of Microsoft’s engineering teams. There are all sorts of examples big (e.g. migrating Windows source code to git) and small (e.g. hackathons) that occurred during the early years of Satya’s tenure as Microsoft CEO which count as listening to developers and making them happier, translating into more productivity and higher morale.