After eight years, what did IBM’s outgoing CEO accomplish?

Few were surprised by IBM’s announcement on Thursday evening that Ginni Rometty will step down as CEO after eight years in the job, to be replaced by Arvind Krishna, who oversees the company’s cloud computing and A.I. software efforts.

Rometty received accolades and praise from fellow CEOs, including Apple’s Tim Cook and JPMorgan’s Jamie Dimon, for her tenure. But while her strong leadership style made her a role model to many, she couldn’t find the right path to get IBM back on the upswing. You don’t hear much about the fifth place player in many tech markets. Yet that’s where IBM finds itself today in probably the most important market for the future of corporate computing, looking up at cloud leaders Amazon, Microsoft, Google, and Alibaba.

It’s already been widely remarked on that IBM’s revenue during Rometty’s tenure declined from $105 billion to $77 billion, a more than 25% drop. You might know the story: 22 straight quarters of declining revenue and just a few of growth. Oh, and don’t let me forget to mention all the stock buybacks.

Rometty knew cloud was the future, making acquisitions to bolster IBM’s offerings, and betting on artificial intelligence, cybersecurity, mobile computing, and the Internet of Things. The goal wasn’t necessarily to become bigger, it was to find businesses where IBM could thrive and dominate. As IBM’s then-CFO Martin Schroeter explained in 2015: “We’re not trying to be the largest of something. What we are, though, is trying to be the highest value.”

The problem was that Rometty bet too small, bet on the wrong parts of those markets, and bet too slowly.

• Take the cloud. Early in her tenure, Rometty paid $2 billion for Softlayer, a cloud hosting service that specialized in the complex needs of government agencies and financial institutions. But Amazon and Google relentlessly expanded the kinds of services offered on their platforms, quickly overtaking Softlayer’s unique strengths. By 2016, IBM eliminated the Softlayer brand name and was stuck in fifth place in the cloud market.

• Another huge miss: developers. Amazon, Google, and Microsoft all made it easier for developers to build apps on top of their cloud services, while IBM was less flexible and inviting. Rometty may have realized that mistake when she moved to buy Red Hat for $34 billion last year. Red Hat’s popular software helps developers write apps to run on various cloud platforms—a great asset if four companies are already dominating the cloud era.

• Did I mention the stock buybacks? They were even more expensive than the Red Hat purchase. Seeking to please Wall Street, IBM spent a staggering $58 billion on buybacks during Rometty’s eight years—nearly half the company’s current market value. And while that did make earnings per share look better, it didn’t help IBM compete against Amazon, Google, Microsoft, or Alibaba at all.

On top of the revenue drop, Rometty oversaw an even more severe decline in profits. IBM’s net income fell from $16.6 billion the year before she took over to $9.4 billion in 2019, a 43% decrease (and that’s calculating in the 2017 tax cut’s huge benefit to IBM). The whole point of the downsizing strategy was to make IBM more profitable, not less. Now that will be Krishna’s challenge.

I’d love to hear your views. How do you judge Ginni Rometty’s tenure? And what advice do you have for her successor? Email me below or tag @FortuneTech on your tweets. And have a great weekend.

Aaron Pressman

Twitter: @ampressman

Email: aaron.pressman@fortune.com

NEWSWORTHY

Finding my way back to you, babe. If you're an iPhone user yet to switch over to the Google Maps app, some good news. Apple said it has rolled out a redesigned Maps app across the entire United States as of Thursday, with much improved detail. Europe gets the improvements next, starting ASAP, Apple says. The company may start to have ever so slightly less resources to make these improvements in the future. Along with supplier Broadcom, Apple lost a $1.1 billion patent infringement case this week filed by the California Institute of Technology over the tech in Wi-Fi chips used in the iPhone.

A lot of zeroes. On Wall Street, Amazon impressed. Revenue for the key holiday season jumped 21% to $87.4 billion, better than analysts expected. That pushed Amazon shares, already up 17% over the past year, up another 11% in pre-market trading on Friday. And that pushes the company's stock market value over $1 trillion.

What were they thinking? Some good follow up to recent news tidbits: Avast, the cyber security firm that was using a browser add-in to track users and sell their browsing data, is shutting down the program "effective immediately." And Google decided to roll back the changes to its search results pages that made it more difficult for users to tell the difference between organic and sponsored results.

FOOD FOR THOUGHT

As artificial intelligence programs improve, the debate over their danger takes on growing importance. Some researchers are focusing on the challenge of imbuing A.I. with a sense of purpose that follows human preferences. Natalie Wolchover, senior writer and editor at Quanta Magazine, digs into the efforts to make A.I. less scary and more attuned to humans. She spoke to Berkeley computer scientist Stuart Russell.

Humans aren’t even remotely rational, because it’s not computationally feasible to be: We can’t possibly calculate which action at any given moment will lead to the best outcome trillions of actions later in our long-term future; neither can an AI. Russell theorized that our decision-making is hierarchical — we crudely approximate rationality by pursuing vague long-term goals via medium-term goals while giving the most attention to our immediate circumstances. Robotic agents would need to do something similar, he thought, or at the very least understand how we operate.

FOR YOUR WEEKEND READING PLEASURE

A few long reads that I came across this week:

The iPad's original software designer and program lead look back on the device's first 10 years (Input)
We chatted with design power couple Imran Chaudhri and Bethany Bongiorno about the tiny tablet that could.

Our Predictions About the Internet Are Probably Wrong (The Atlantic)
It’s easy to forget how unforeseeable the “unforeseeable” really is.

The Essential Clayton Christensen Articles (Harvard Business Review)
Clayton Christensen died on Jan. 23, 2020. Here we present some of his seminal HBR pieces through an adaptation of the introduction to the book The Clayton M. Christensen Reader.

If Everyone Hates Spirit Airlines, How Is It Making So Much Money? (Marker)
Inside the loathed company that defies every rule of branding—and logic.

IN CASE YOU MISSED IT

As coronavirus spreads, facemasks become a ‘symbol of protection’—and a lighting rod for fear By Naomi Xu Elegant

The next frontier of road rage is the parking garage By Andrew Moseman

In the age of Twitter and group chats, seconds-long streaming delays can make or break live entertainment By Aric Jenkins

Cisco’s experiment with radical transparency By Claire Zillman and Emma Hinchliffe

Why Waymo’s alliance with UPS is a big deal for autonomous vehicles By Aaron Pressman

BEFORE YOU GO

Luxury carmaker Lexus holds an annual design contest that attracts thousands of entries. They named the 2020 finalists this week. Among them, Lick is a crazy-looking portable body cleaner for people who don't have access to a bath or shower. Pursewit is a super-smart sewing machine for the visually impaired. Check out the whole list and you'll be smiling all weekend at the ingenuity.