Before Apple acquired the program, Siri couldn’t speak. Users could ask their questions by voice or text, but Siri only answered via text. The developers figured that the information was already on the screen, and people would be able to read the answer faster than Siri could talk.
Apple bought Siri in a deal that TechCrunch speculated to be worth around $200 million.
Once Siri was absorbed into Apple, the company added some capabilities such as more languages and the ability to speak. Apple also integrated Siri more deeply into the phone. Instead of just being a Web-based service that could only be accessed by opening the Siri app, users could hold down the home button to summon Siri instantly.
But Apple stripped the technology’s abilities to a skeleton of its original self. The company removed key features, including the option to type in questions. It substantially reduced the number of websites from which Siri could draw answers.
Now Siri was entirely reliant on imperfect voice recognition, and when it did parse a question correctly, Siri had less information at its disposal to formulate an answer.
Apple’s love of simplicity made things difficult as well. To encourage users to phrase their questions in a way that Siri could understand, the product used to show sample questions on the screen when activated. Apple’s user interface team insisted on a cleaner design and hid the questions under a tiny black button with an “i” for information. Overlaid on a charcoal background, the tiny gray-on-black button was so subtle that most users didn’t notice it. Without guidance, many iPhone users phrased their questions in ways that stumped Siri.
Part of the problem may have been that Siri was being managed by Forstall’s mobile software group, which wasn’t experienced in working with technologies like Siri. That team was used to secretly building operating systems that resided inside the devices, were beautiful to interact with, and didn’t change very much once shipped. Siri was a different breed of product. When users asked Siri a question, the audio of their voice was sent over the Internet to data centers full of servers that deciphered the question and then sent back an answer for Siri to say out loud.
Unlike iOS, which could be tested by a limited number of engineers, Siri required broad and extensive testing that took into account various accents, traffic volume, and types of questions that could potentially be asked. But Apple’s secretive culture constrained the company’s ability to adequately test the app and iron out major bugs before the launch. This was a recurring issue for Apple, but with software like Siri, the consequences were more profound.
Possibly the biggest mistake that Apple made was setting expectations too high. Software as complex as Siri was never going to work perfectly from the beginning. Because the servers depended on large amounts of data to refine the quality of the answers, this kind of software started out good enough and got better over time with rapid and constant improvements. By necessity, Apple launched Siri as a beta product that was still in the test phase. But the marketing for Siri glossed over the app’s experimental status. Instead of preparing the public for the fact that the software was still under development, the company sold Siri as a helpful presence living inside the iPhone, emanating serene wisdom and a hint of mischief.
The result was a mess. Apple had grossly underestimated the number of queries Siri would have to deal with when the iPhone 4S came out, crashing the servers and resulting in hours of downtime. When Siri did work, the answers were often wrong or nonsensical. New iPhone 4S owners, eager to show off Siri, were frustrated.
The problems poured in as Siri inadvertently waded into controversial topics. The media had a heyday when users discovered that if they asked Siri to find a local abortion clinic in Manhattan, it told them: “Sorry, I couldn’t find any abortion clinics.” One person found that when users asked the question in Washington, D.C., Siri directed them to anti-abortion pregnancy centers in Virginia and Pennsylvania rather than the nearest Planned Parenthood.
Belatedly, Apple reminded the media that Siri was still a test product. “These are not intentional omissions meant to offend anyone,” a spokeswoman said. “It simply means that as we bring Siri from beta to a final product, we find places where we can do better, and we will in the coming weeks.”
In the first several weeks after the 4S launch, the Siri office in Cupertino looked like a war zone. The team was working around the clock, scrambling to make improvements and beef up the servers to accommodate the enormous amount of unexpected traffic. Military cots were set up in the hallways for people to sleep while they troubleshot.
“There’s no way for developers to perfectly anticipate all the requests tens of millions of users will make before the system is launched,” explained Cheyer, defending Siri’s launch. “You just can’t. But once the system is live and people begin to interact with it, Siri’s accuracy should improve significantly as it learns new words, phrasings, accents, and request types.” He likened Siri’s development to building a complex factory with intertwining pieces that had been developed separately but now needed to run as one. From an engineer’s perspective, Siri’s launch had been a tremendous accomplishment despite the problems. Apple was ahead of the competition just by the fact that the company was well on its way to creating a strong services portfolio in addition to hardware and software.
But even after the blitzkrieg campaign to boost the software’s capabilities, Siri still had a long way to go. In a test against a similar service by Google, Piper Jaffray analyst Gene Munster gave Siri a D. Among eight hundred questions he asked, Siri understood 83 percent of them but answered only 62 percent accurately. Google’s technology understood 100 percent of the questions and got 86 percent right.
Six months later Siri’s accuracy rose to 77 percent. But Munster still rated that as only a C.
Determined to make the new iPhone a hit, Apple released a television marketing campaign that showed Siri bantering with Hollywood celebrities. The ads were fairy tales, featuring exchanges that Siri was not yet capable of carrying out. In one spot, Martin Scorsese chats with Siri in a New York cab—an obvious allusion to a scene in his a film Taxi Driver.
The ad begins with the Oscar-winning director talking into his iPhone, posing a question to Siri in the middle of Manhattan traffic.
“What does my day look like?” says Scorsese.
“Another busy day today,” answers Siri.
“Are you serious?”
“Yes,” says Siri. “I’m not allowed to be frivolous.”
“Ah, okay,” says Scorsese. “Move my four o’clock today to tomorrow. Change my eleven a.m. to two.”
“Okay, Marty. I scheduled it for today.”
Scorsese then asks Siri to check on traffic heading downtown.
“Here’s the traffic,” says Siri, showing him an update on the iPhone screen.
“It’s terrible. Terrible!” says Scorsese, speaking in his familiar staccato. “Driver—cut across, cut across. We’ll never make it downtown this way.”
The commercial ends with Scorsese saying, “I like you, Siri, you’re going places.”
Siri responds, “I’ll try to remember that.”
The other spots in the series, which featured Samuel L. Jackson, John Malkovich, and Zooey Deschanel, also seemed out of sync with Apple’s style. The company traditionally shied away from celebrities because it considered its products to be the star, or in Apple’s inside lingo, the “hero.” Now, it was depending on actors to relay the coolness of iPhones. Not only that, but some of the spots were simply unconvincing.
In the spot with Deschanel, she asks Siri, “Is that rain?” even though she is standing near a window and it is clearly raining hard outside.
“It may be all the rage for celebrities in iPhone commercials to have pithy exchanges with Siri,” wrote Tuesdays with Morrie author Mitch Albom, “But if you ask me, they just sound stupid.”
The TV commercials only stoked the public’s expectations. In actuality, Siri was not yet capable of responding with such
nuance and precision. The ad makers hadn’t given the developers a heads-up about the conversations they had in mind, so the team hadn’t trained Siri until later to answer such questions. This prompted a lawsuit the following March by iPhone 4S buyer Frank Fazio, charging that Siri did not perform as advertised. The discrepancy would have never happened under Jobs, who was a stickler for getting these kinds of details right.
By June 2012, Apple had made a number of improvements and was finally ready to add some of Siri’s former features, including the ability to book restaurant tables. But Siri’s skills were still inadequate enough that Wozniak gave a big thumbs-down.
“I’d say ‘What are the prime numbers greater than eighty-seven?’ and I’d get prime rib,” he said. “A lot of people say Siri. I say poo-poo.”
Nearly a year after Siri’s launch, the experience was still bad enough that Jack in the Box, the fast food restaurant chain known for its edgy ads, unleashed a television spot that demolished whatever remained of Siri’s credibility. The mascot Jack, with his round head, blue dot eyes, and yellow clown cap, is shown at a cell phone store trying out the virtual personal assistant feature in one of its phones. The device doesn’t look like an iPhone, but the ad clearly mocked Apple through an absurd exchange between Jack and a pseudo-Siri.
“Where’s the nearest Jack in the Box?” asks Jack.
“I found four places that sell socks,” says a digital voice.
“Not socks. Jack in the Box.”
The Siri clone still doesn’t get it.
“A yak,” she says, “is a long-haired bovine.”
“That’s true,” a salesman interjects.
“I like things that work,” Jack says, segueing into a pitch for a hamburger. “Like my no-nonsense All American Jack combo.”
“Sounds delish!” says the salesman.
“I found one D-list celebrity nearby,” says Siri.
The ad hacked away at yet another piece of Apple’s hipper-than-thou image. In some ways, the damage was worse than ads by rivals because this jab came from a company in an unrelated industry. Apple was targeted purely for entertainment.
But by then, most of Siri’s founders had left Apple.
Siri’s rocky start wasn’t Tim Cook’s fault. He had barely taken over the company when Apple launched the iPhone 4S. But the problems exposed the flaws in the Apple way and showed the magnitude of the challenges that the new leader faced. It also provided a glimpse into how development could fall apart without Jobs.
The Siri team wouldn’t have agreed to the acquisition if they hadn’t believed, in the beginning at least, that Apple shared their vision. Scott Forstall was their lead manager, but Jobs had been closely involved at first and had been a strong advocate. According to the original vision, Siri was supposed to have fundamentally changed the way people manipulated information and tasks. But as Jobs grew more ill and withdrawn in his last months, the Siri project lost its way. As new hires who had been absorbed into the company through an acquisition, the Siri team found it difficult to be accepted into Apple’s exclusive corporate culture—much less win any arguments against insiders. The fact that Siri fell under Forstall isolated the team even further. The executive was so controversial that some people outside of Forstall’s group didn’t want to have anything to do with Siri. What could have been the game changer that everyone was waiting for fell flat, buried under hype.
Opinions were split about whether Siri’s launch would have been any smoother under Jobs, who had had some duds in his career. In addition to every product launched by NeXT, he was also responsible for the square-shaped Cube computer and the Apple TV living room device, which he called “a hobby.” Complex software powered through the Internet had never been a strength of Jobs. MobileMe, Apple’s first attempt at an Internet-based sync and storage service, was a catastrophe as users had trouble signing up for the service and accessing their data. Jobs had called it “a mistake,” admitting that the service had needed more time and testing. The next iteration, iCloud, was ridden with bugs, though it was better than MobileMe. iTunes was successful, but it was easier to grow a small service into something bigger over time. Supporting Siri for millions of customers out of the gate was akin to turning on a fire hose.
But even if Jobs had given Siri the green light, it would have been a departure for him to make an incomplete feature like Siri the centerpiece of a new device, particularly in its ads. Google might release new apps that were still unfinished because they depended on user feedback to improve them. But that wasn’t what Apple did. Consumers demanded perfection from the company.
Having watched Apple reinvent the wheel time and time again, the public expected the parade of wondrous new devices to continue. For years, Apple had fed that desire. But the string of triumphs now made it much more difficult for Apple to outdo itself, particularly without Jobs’s reality distortion field.
At this stage in Apple’s evolution, Cook’s leadership approach held promise. Apple’s nimble and daring culture under Jobs had made for a chaotic environment. Projects changed directions abruptly, rivals were pitted against each other, and employees were forced to give up their lives for Apple, all in the name of coming up with the best products in the world. Jobs, a workaholic, made a practice of recalling his executives from vacations to deal with one crisis or another. It kept them on their toes.
That culture, however, had been showing signs of strain. The company needed Cook’s pragmatism and organizational wizardry.
“In terms of day-to-day business, he’s better than Steve,” Avie Tevanian, Apple’s former head of software who had worked with both of them, said shortly after Cook’s appointment. “He understands more about the nuts and bolts of being an executive.”
Cook was a methodical and efficient CEO. Unlike Jobs, who seemed to operate on gut, Cook demanded hard numbers on projected cost and profits. He expected teams to stick to the budget. Project managers grew more powerful.
Such hyperorganized planning was jarring for those who had never worked for Cook. In a scene that was all too familiar to the operations team, one manager who went to Cook with a plan for an Apple Store iPad app found himself hit with question after question as Cook drilled down into the details, asking him about specific costs related to a particular aspect. When the manager was unable to respond adequately, his new boss dressed him down.
“Don’t you think that’s something that you should know before you come to me with it?” Cook asked quietly as the manager quaked.
But whereas Jobs had reveled in divisiveness, Cook valued collegiality and teamwork. The dynamic among Cook’s executives shifted to adapt to his more cooperative leadership style. In addition to Ive, smart, affable types like Cue and Apple’s new Mac software chief Craig Federighi rose in esteem. Schiller, famous for his frequent screaming matches with other members of the executive team, toned down his aggressive behavior and became a calmer team player. Though some had speculated that Schiller might resist Cook’s leadership, the marketing chief fell in line quickly and was rewarded with more responsibility. Unlike Jobs, who stuck his head into anything that interested him, Cook delegated. He gave teams more freedom to run their divisions the way they thought best.
In April, Cook held his first Top 100 meeting as CEO. The location was the same as the last one—the Carmel Valley Ranch. Cook adhered to tradition, from the way the invitations were sent out to the chartered buses that executives were required to ride to the resort.
On the agenda had been a session describing the new spaceship-like headquarters that Jobs had pitched to the Cupertino City Council in his last public appearance. The attendees were shown a model so they could see how it had a mix of open and private meeting spaces.
Everyone at the Top 100 left feeling inspired.
“I was really blown away,” said one executive who attended the meeting. “In the past, people would give a presentation that would be interrupted. Steve would say, ‘You’ve got it all wrong. Let me tell you how it is
.’ Tim’s not like that. There was a feeling of more trust, more openness. I came back just pumped. My gosh, we haven’t skipped a beat.”
In the first months of Cook’s tenure, Apple’s investors were impressed with his style. Though Jobs had rarely paid attention to them, Cook was more visible and transparent. As CEO, he continued to participate in the quarterly conference calls with analysts to talk about earnings, and in February 2012, he attended the Goldman Sachs Technology and Internet Conference, where he answered questions about Apple’s future growth opportunities and labor conditions in China.
Throughout the spring, he also met with visiting groups of investors led by research analysts for Citibank, Piper Jaffray, and others. At one such event in May, Cook stayed for forty-five minutes in a cramped meeting room with thirty investors. Refreshments consisted of little more than cookies, water, and soda, but in answering questions, he provided more context and details than ever before. These meetings were typically attended by two or three senior executives—Peter Oppenheimer, the chief financial officer, and either Cue or Schiller. Cook had occasionally dropped in on them as chief operating officer, but Jobs had never participated.
Not everyone was so enamored. Though many of the changes Cook made established more order at Apple, they were also perceived as signs of increasing stodginess. The yearning for more subversive days was palpable.
Skeptics soon began expressing doubt about Apple’s future, especially after the Siri fiasco. George Colony, the CEO of technology research firm Forrester Research, was among the most influential critics, posting a stinging blog post predicting the fall of the Apple empire.
“Without the arrival of a new charismatic leader it will move from being a great company to being a good company, with a commensurate step down in revenue growth and product innovation,” Colony wrote. “Like Sony (post Morita), Polaroid (post Land), Apple circa 1985 (post Jobs), and Disney (in the twenty years post Disney), Apple will coast and then decelerate.”
Haunted Empire: Apple After Steve Jobs Page 18