The Case for Steve Ballmer

There's a widely accepted story about Microsoft's history: that the company stagnated under Steve Ballmer and was rescued by Satya Nadella's visionary leadership. This narrative shows up in just about every discussion of the topic, and it's rarely challenged. Nadella deserves credit for his tenure, but this framing badly underrates what Ballmer accomplished. Microsoft's financial performance under Ballmer was strong by any objective measure, and the company made deep, long-term bets during his tenure that positioned it for the success it enjoys today. Those bets were widely criticized at the time, which suggests they weren't obvious moves, but hindsight shows they were the right ones.

What's often overlooked is that Ballmer also cleared the political obstacles that could have crippled any successor. He set the company up so effectively that it's easy to look at Nadella's success and conclude Ballmer was a failure — the same dynamic that makes a well-structured technical talk seem trivial to listeners who didn't have to work through the problem themselves.

Why Ballmer Became a Punching Bag

By 2007, Microsoft had lost its reputation as the most feared company in tech. Paul Graham's essay "Microsoft is Dead" captured the sentiment, noting that young startup founders no longer understood why anyone would have been frightened of the company. Graham's thesis was that technology companies are projectiles — once they lose relevance, the balance sheet takes years to catch up.

Ballmer became the focal point for this criticism. He came from the business side, starting at Microsoft as the first business manager before becoming EVP of Sales and Support, and later CEO. Tech culture has never been kind to non-technical leaders, and Ballmer's stage persona didn't help. He was portrayed as a sales guy who didn't understand technology, could only copy competitors, and cared only about the bottom line. When he pushed out Windows chief Steven Sinofsky in 2012, online commentary was nearly universally negative. One anonymous insider comment recommended firing 40% of the workforce and resetting Windows, while claiming the company's online services were hopeless.

Ballmer's defense was that the market undervalued Microsoft relative to its fundamentals, especially compared to Amazon, Google, Apple, Oracle, IBM, and Salesforce. That assessment turned out to be fair: Microsoft has outperformed all of those companies since Ballmer made that argument.

The Wins That Get Written Off

The conventional narrative says Ballmer missed every important shift in technology. The evidence says otherwise. Several financially material developments happened under his leadership:

  • 2009: Bing launched. Bing is treated as a laughingstock, but the bar for "failure" here is extraordinarily high. Bing reportedly generated $1B in profit in 2015 and $6.4B in profit on $12.6B of revenue in fiscal 2024. By most standards, that's a wildly successful business unit — the kind most companies would love to have.
  • 2010: Azure was created. Microsoft is now one of three companies — along with Amazon and Google — that are head and shoulders above everyone else in cloud infrastructure. Microsoft is a solid #2 in cloud, and the enterprise sales organization built under Ballmer was critical to Azure's success.
  • 2010: Office 365 was released. Microsoft moved its enterprise software suite from boxed products to subscriptions with online options. This transformation was risky at the time, but it created a business that could rival Azure in value if Microsoft ever split into separate units.

The enterprise sales arm deserves special attention. Ballmer built it before becoming CEO — he was EVP of Sales and Support, and he continued to invest in it. The sales organization was so effective that when I was at Microsoft, Google would offer enterprise customers Google's suite for free, and Microsoft salespeople would still close deals for the paid Microsoft product. Google literally couldn't give its enterprise software away in some markets.

At a startup I knew, an Azure salesperson opened a pitch with "You're on AWS, the consumer cloud. You need Azure, the enterprise cloud." That pitch sounds ridiculous to people who work at tech companies where "enterprise" is synonymous with overpriced junk. But Microsoft's enterprise sales organization was effective at every stage. When I worked in Azure in 2015 — under Nadella, but in an organization whose culture and processes were built under Ballmer — there were months when Microsoft hired and onboarded more salespeople than Google employed in total.

Misses and Portfolio Thinking

The list of Ballmer's failed bets — Bing, Zune, Windows Phone, HoloLens — is often presented as evidence of incompetence. But this misunderstands how large technology companies work. Microsoft runs a portfolio of bets. One big success can pay for a dozen failures. The relevant question isn't whether every bet succeeded; it's whether the total return justified the portfolio.

By that standard, Ballmer's record is strong. Revenue grew from roughly $14B or $22B — depending on whether you count from when he became President in 1998 or CEO in 2000 — to $83B when he stepped down. The company was recording $27B in annual profit, which was more than the entire revenue of the company he inherited. Microsoft's total return under Ballmer was very good.

And it's not as if there were no hits. Ballmer presided over Azure's creation, the transition of Office to a subscription model, the development of an unmatched enterprise sales organization, and the building of Microsoft's gaming business — he was CEO when Microsoft acquired Bungie and made Halo the Xbox's launch flagship. Even Bing, the canonical "failure," was generating billions in profit. Critics point to Bing as proof Ballmer failed, which says more about the scale of his successes than his failures.

What Ballmer Inherited

Nadella inherited a company primed for success. Ballmer inherited a company in a much harder position. By the early 2000s, Microsoft was widely considered a boring has-been, next in line to become "the next IBM," largely due to decisions made during Bill Gates's tenure.

Graham cited Google and the web as the primary causes of Microsoft's death. Both threats trace back to the same root cause: the antitrust case against Microsoft. Microsoft knew the internet mattered and planned to dominate it, using its Windows monopoly to destroy Netscape. The company technically lost the antitrust case, but the remedies were widely considered useless — the original breakup was reversed, and the remaining penalties were mostly meaningless. Netscape was already doomed by the time the case ended.

What's less known is that Microsoft had a later plan to kill Google. A proposal was discussed to redirect users who typed "google.com" into their address bar to MSN Search. This was before Chrome and mobile existed meaningfully. Windows had 97% desktop market share, and IE dominated the browser market with anywhere from 80% to 95%. Executing that plan would have killed Google before it could launch Chrome and Android, and Microsoft would likely own the web today.

Microsoft chose not to do it — not out of fear of antitrust action but out of fear of the PR disaster that would follow. Google "supposedly killed" the 2011-2012 FTC investigation with lobbying, and has been more careful with how its recent antitrust case plays in the media. Microsoft, by contrast, was treated harshly for actions that Google gets away with or is even praised for. One example: in 2011, Google officially called out Microsoft for unethical behavior, and the media piled on, but at Microsoft, employees said they'd gotten the idea from watching Google do the same thing. Reputations are sticky.

This asymmetry — where Microsoft received severe criticism for actions that Google could take without scrutiny — constituted a real constraint on the company's maneuvering. The same move, executed by different companies, produced completely different outcomes in terms of public perception and regulatory attention.

Internal Politics and Succession

Ballmer also dealt with Microsoft's notoriously brutal internal politics. He bore some responsibility for creating that culture, having been a senior figure since nearly the beginning. But he cleared out the worst actors, leaving Nadella a much more functional company.

Microsoft had working competitors to Google Docs and NetMeeting in 1997, but they were killed for political reasons. This pattern was common at Microsoft, which was famous for intense internal struggles — more so than most other companies. When I was at Microsoft and asked about promising projects that were sidelined or killed, the biggest recent sources of those problems had been shown the door under Ballmer.

One concrete example of leadership making structural changes: when Azure needed features added to Windows networking, the Windows team would respond with "we'll put it on the roadmap," which everyone understood meant "we don't have to do anything you say." So leadership moved Windows networking into the Azure organization, giving Azure full control over the features it needed. Compare that to Qualcomm's server chip effort: when that group threatened to become larger than the mobile group, the mobile group had the server group killed before it could grow large enough to defend itself — and when the CEO supported the long-term opportunity, he was fired and removed from the board.

The Long Game

Ballmer navigated a period when Microsoft was perceived as doomed. Predictions like Graham's — that Microsoft's revenue would decline within five to ten years — seemed credible at the time given Microsoft's political constraints, both internal and external. But no dramatic turnaround happened. Microsoft kept executing on enterprise products and kept making — and heavily subsidizing — new bets: Windows Phone, Bing, Azure, Xbox, and HoloLens. These bets were panned as reckless, with many arguing that Microsoft should focus on its most profitable lines, like Windows.

It's rare for companies to use successful businesses to fund new ones when the writing is on the wall for the old ones. Even with clear data showing that change is necessary, most leaders won't do it because they look foolish if it doesn't work. Ballmer was willing to take that risk.

It's also rare for companies to successfully expand into radically different markets. Most companies are locked into their core business. When Google launches a consumer product that fails — a common occurrence — it's usually dismissed as expected: of course Google failed there. Microsoft, though, made major transitions multiple times. Xbox was Microsoft — a boxed software company — learning to compete in hardware against Sony and Nintendo, both of which had decades of hardware experience. Azure was Microsoft learning to compete in online services against web-native companies like Amazon and Google. Both transitions succeeded at an enormous scale.

These transitions weren't smooth. Azure's early days were chaotic — before the team had an incident management process, major global outages prompted hallway conversations asking "is it Azure down?" The company that built Windows didn't originally know how to run cloud infrastructure. At one point, an engineer noticed that AWS's retail price for disk was cheaper than Azure's cost to provision it. Microsoft later hired senior people from Amazon with supply-chain and datacenter expertise. It's easy to say that companies should recognize their weaknesses and hire experts, but most fail at this. The old guard typically shuts down outside expertise, especially at a company as fractious as Microsoft was.

Google engineers were dismissive of Azure while it was rising, pointing out Microsoft's operational incompetence — comical for engineers at a company that pioneered large-scale online services. Despite starting in a deep technical and cultural hole, Microsoft built a business worth roughly a trillion dollars.

Beyond Finance: Relevance to Programmers

People dismiss Microsoft's relevance by pointing out that it attracts criticism as a financially successful but boring company — the new IBM. Under Ballmer, however, Microsoft shipped things that matter quite a bit to working programmers:

  • 2007: LINQ, which remains a well-regarded language feature by practitioner standards.
  • 2011: Sumit Gulwani of Microsoft Research published "Automating string processing in spreadsheets using input-output examples," which was named a most influential POPL paper a decade later. This work likely explains why Excel's autocomplete features work far better than Google Sheets'.
  • 2012: TypeScript, now the most widely used programming language released this century, and a plausible candidate for most widely used overall.
  • 2012: Surface. Even as a so-called failure, it was roughly a $7B/year business in 2022.
  • 2015: Visual Studio Code — released after Ballmer's tenure but building on work from it, including the addition of Erich Gamma in 2011, and now the dominant programmer editor.

Rounded to the nearest Bing, IBM might be worth a full Bing, or might not. Among today's trillion-dollar tech companies, both Apple and Nvidia hold developers through ecosystem lock-in — iOS and CUDA respectively. Microsoft holds developers through deeply entrenched platforms like Windows and .NET, especially for areas like AAA games. But Microsoft also retains relevance among developers who aren't stuck in an ecosystem, mainly because of developer tools like TypeScript and VS Code that were created under Ballmer's watch.

A Missed Opportunity

The strongest criticism of Ballmer's era is that Microsoft failed to learn from its own antitrust case the way its competitors later did. A sufficiently prescient executive might have lobbied hard to prevent the case, or managed to minimize its media impact like Google has done. Microsoft also failed to recognize that the U.S. was entering a two-decade period of reduced tech antitrust enforcement after it was the subject of the last major case.

Not killing Google was perhaps the clearest example of this: the decision was driven by fear of bad PR, not antitrust risk — and bad PR was arguably a worse enemy than the actual legal system. The DOJ case against Microsoft was regular front-page news in the 1990s, and the company could not take actions that Google would later take without severe criticism.

But is that an indictment of Ballmer? It would be, if we expected CEOs to be infallible. As criticisms go, this one boils down to "Ballmer wasn't the greatest CEO of all time by a huge margin." That's true. It's also not much of a criticism.

The Track Record in Context

It's also worth noting that in the broader history of the tech industry, sustained multi-decade excellence is rare. Microsoft has executed well from its founding in 1975 to today — a roughly fifty-year run that's arguably unmatched. Intel has been around a bit longer but stumbled badly around the turn of the century and has had serious security issues over the past decade. IBM's long history hides the fact that it was small in its early days, was hamstrung by the antitrust case that ran from 1969 until it was dropped in 1982, and was nearly dead until Lou Gerstner's turnaround. Companies like DEC and Data General vanished. Even Apple nearly died. Oracle has had a comparable run of financial success but with far less growth in adjacencies — and Oracle is worth perhaps two Bings, making it the 20th most valuable company in the world. That's not nothing, but it's not Microsoft.

I joined Microsoft in 2015 and made a private bet against the company: that Google would reach a trillion-dollar market cap first. I agreed with the big bets Microsoft was making, but I thought internal dysfunction would hold it back. I was wrong. Microsoft beat Google to $1T and is now worth roughly a trillion dollars more. I wouldn't have made that bet even a year later, after seeing firsthand how effective Microsoft's sales force was, how well it shipped products that appealed to enterprises, and how that compared to Google's cloud efforts. My own misjudgment was similar to what Ballmer's external critics have been doing for years, except that I corrected my view after seeing how Microsoft actually operated. Most of the public criticism has no equivalent corrective experience.