Learnings From the Sudden Death of the World’s First Social Robot

October 16, 2019

Jibo was a brilliant idea. It would be the world’s first social robot, bringing to life years of academic research in the form of a product that would live in a consumer’s home. It was an amazing invention, lovingly designed and carefully crafted by people passionate about the product and its persona. Nothing captured the imagination of millions of people the way that this little robot did. It landed on the cover of Time as one of the best inventions of 2017, the year it launched.  

Despite this promise, Jibo the business failed after raising more than $80 million and delivering a few hundred robots to early customers. The servers were turned off in 2019, and Jibo bade a sad farewell to its owners. “Maybe someday,” it said, “when robots are way more advanced than today and everyone has them in their homes, you can tell yours that I said hello.” 

Jibo was an invention, but I’d argue that it never actually became an innovation. It failed to create value for its employees, investors, and most of all, its customers. Its failure had little to do with poor product-market fit, and Jibo had plenty of advanced robotics inside. Instead, Jibo failed in its own special way, in ways that illustrate how other creative inventions also fail.

The Beginning of Jibo

When Jibo started, it was born from a great place: an MIT researcher had proven that people interact with physical robots in very special ways, and that same research and vision made investors believe there was a clear (albeit ambitious) path to creating the world’s first social robot for the home. 

The early team worked to create a prototype that showed how a physical robot, while remaining stationary on a desk and only articulating its body, could yield a magical experience that people wanted. And boy did they want it! The tiny team took on a CEO and pushed out a concept video on the crowdfunding site Indiegogo. Before they knew it, it was the most successful Indiegogo campaign ever. This proved that people wanted the fantasy of the social robot they saw in the video—even if they didn’t really know what it was going to be like. The public’s understanding was overwhelmingly grounded in robots from the movies and TV: Rosie the Robot Maid from The Jetsons, R2-D2 and C-3PO from Star Wars, TARS and CASE from Interstellar, and every other robot in between. However, unlike other pieces of personal technology for which people have good mental models, people didn’t know what it meant to have a robot companion, because no one had ever owned a social robot. 

I was the Vice President of Design at Jibo for two-and-a-half years, and my incredible internal (and extended) design team created everything from logos and packaging to movements, screen animations, speech interactions, touch screen interactions, hole patterns for the speakers, and more. Other teams believed that design would be critical to the product’s eventual success in the market, and they worked tirelessly with us to give their very best. But their best didn’t make up for the unique set of failures that we, as a company, made.

Four Strategic Failures

Sketches of the Jibo design.

Failure #1: Creating a Category Is Really Hard. This is where the heart of the problem really begins. Since there was no dominant design for what a social robot is supposed to be, and since no one knew how to successfully make one that lived with you at home—including the team that had the initial vision—there were few starting points to work from. This was the Wild West of human-machine design: no rules, no reference designs, no best practices. And instead of being developed in the well-funded R&D arm of a corporation, it was being built under the brutal pressure of a startup. We needed to hire our first employees, design everything, write all the code, manufacture complex hardware, ship it, and support it. It was a Sisyphean task with a classic innovation problem: first-mover disadvantage.    

There’s a good reason why most successful startups don’t create new categories, but rather build on existing ones. Being a first-mover is much harder and riskier than being a fast-follower. Even one of the most innovative companies in the world, Apple, is generally a fast-follower (and sometimes it is a slow one.) Apple didn’t invent the digital smartphone, the graphical user interface, the MP3 player, Bluetooth headphones, or the smart watch. It’s not that they didn’t bring incredible innovation to these product categories—but the categories at least existed prior to their entry. They could “reinvent the phone” because smartphones already existed. They could put 1,000 songs in your pocket because portable MP3 players were already on the market. Apple just figured out how to do them better. And when Apple was first to market, it often had rocky starts or outright fails, like the Newton MessagePad, which preceded the PalmPilot by four years. (If you’ve never used Newton MessagePad, I can tell you from personal experience that it was a feat of incredible design, and it included inventions that the iPhone still uses to this day.) And AppleTV is finally getting traction after a slow and prolonged deployment.  

Solution: Create transformational innovations in an existing category, or expect that you’ll need the horsepower and protection of a larger organization to weather the storm of uncertainty.

Failure  #2:  Creating  a  Shared  Vision  is  Hard. As we grew the staff from 13 to more than 100, each new employee was incredibly committed to the concept and to helping reify the vision—as they understood it. Of course, not everyone understood the vision the same way. Some interpretations of what the product should do were incredibly specific and easy to render: “I want it to twerk,” said one executive, and the next day Jibo was twerking to the song “Turn Down for What?” While nauseating for some of us, it was an important step because it was something tangible we could react to, experience, and discuss. 

However, at other points, it was much harder for the teams to form an aligned vision. As the VP of Product once said to to me:  “I believe Jibo is alive” Me: “Alive, alive? Or just like, makes you feel like it cares?” VP of Product: “Actually alive.”

Not only did people have different beliefs about the vision, but they also differed on the specifics of what we should spend our limited time working on. Amazon had begun shipping its Alexa intelligent speaker in 2013, the year after Jibo was founded. Our product team said, “Amazon’s Alexa has a timer, so we need a timer,” while the design team said, “We should only create experiences that convey the social character.” When we weren’t able to effectively resolve these conflicts, small fissures widened into chasms, and we lost our ability to jointly agree on what the next sprints of work should be.   

Solution: When teams disagree about their interpretation of the vision, be willing to:

1. Address issues, knowing that some of the right answers may mean delaying an already ridiculously tight schedule.  

2. Ensure that the team members can handle lots of ambiguity.  

3. Know that prototyping and externalizing the actual experience is the only way to inform a shared understanding.

Failure  #3:  We  Lacked  Courage  and  a  Framework  for  Transformational  Innovation. After leaving Jibo, I spent some time reflecting on the lessons I learned, and I developed a framework for creating transformational innovations that I call the Experience Centerline. It’s a simple concept that refers to establishing the core of what makes the user experience of a product differentiated and valuable. 

Elegant solutions to complex problems are incredibly difficult to create, because it means that a lot of time and effort must be spent defining the core of what the product needs to be without any of the extra embellishments (this is the Experience Centerline). In addition to that, you need the courage to stand by your beliefs—even when it’s easier or desirable to hedge your bets by deviating from it.  Startups face this precise challenge whenever early customers or even investors ask for features that aren’t on the Experience Centerline. Do you stick to your guns with a messianic sense of purpose, or do you cave because of outside pressures and incentives?

When it became clear that Jibo’s production schedule was slower than the executive team wanted—by a factor of three—and customers who plunked down $700 in the Indiegogo campaign and the company’s venture capital backers were getting restless, the strategy changed from “Make the best possible product” to, “Ship it now, and we’ll work on it when it’s already in the field.” 

At that point, shipping the product was probably the only acceptable thing to the investors, customers, and the team. To be honest, it would have been much harder to pause the project, endure a massive round of layoffs, rework the experience with a smaller team, and spin things up again. But making hard choices is what often leads to great innovations.  

Solution: If you don’t spend enough time up front perfecting the core of the experience, no amount of  tweaking later on will help you recover from the gaps left begind after abandoning the hard work of deeply understanding and creating the Experience Centerline.   

A 3D printed model of Jibo. 

Failure  #4:  Lack  of  Leadership  Skills. It wasn’t that we lacked leaders, or that we lacked leadership. In fact, we had a whole team of successful, experienced leaders in all parts of the company.  But we did lack certain leadership skills

I’m a Senior Lecturer at MIT, where I teach design thinking and innovation in the Gordon-MIT Engineering Leadership Program to undergraduates as well as professionals. We believe, and the research has shown, that teams benefit when all members have strong leadership skills—from the CEO all the way down to the most junior engineer. However, knowing about these skills doesn’t necessarily mean they will be implemented. Paraphrasing a colleague of mine: just because you know about skiing, doesn’t mean you can get on a mountain, strap boards to your feet, and ski.  

There are a variety of leadership skills that we teach in the Gordon-MIT Engineering Leadership Program that could have helped, including advocacy, identifying the paradox, vision creation, self-awareness, and courage. 

Those kinds of skills would have helped the team communicate more effectively and operate more cohesively, but there were more errors  that exacerbated the problem. 

Four Tactical Errors That Compounded The Problem

In addition to needing those leadership skills at various levels of the organization, we made four key tactical errors (so try to avoid this when you’re in a high-pressure situation):

Error: We Made A Software Platform Before We Knew What It Needed To Do. That’s maybe a little glib, because we thought we knew what it needed to do. But since we hadn’t yet created a version of the product that worked and which we liked, we shouldn’t have assumed we knew what interactions we should support, or the tools that developers would need at that time. We spent a lot of resources creating a robust software development kit (SDK) with the belief that a bunch of people outside the company, who hadn’t spent any time developing a social robot before, would be able to create desirable add-on skills for Jibo—when we hadn’t even solved this problem yet. 

Lesson: Until you know what the experience should be like, don’t expect that others who aren’t living with the problem as deeply as you will be able to constructively address the problem.

Error: We Didn’t Have a Method to Support Rapid Learning.The design team  couldn’t learn fast enough, because the hardware and SDK weren’t yet working. In fact, we spent months using a Samsung phone taped to a model of the robot that designers moved manually, to try to get an understanding of how the robot might feel. If we created the ideal experience first, lived with it, and then tuned it up, we would have quickly learned the key areas we needed to focus on.

Lesson: Focusing on creating a single working solution may have made it seem like progress was slower—but it would have ultimately led to faster deployment.

Error: We Tried to Motivate the Team with Gloom And Doom.At a team meeting, an executive tried to motivate the team by telling everyone how little money and time we had before going bankrupt. Of course, that technique didn’t help our already talented and highlyinvested team be more creative or develop new technologies faster.

Lesson: What the team needed was clarity about the vision, a single point of advocacy for that vision, and the courage to stay the course without distraction.

Error: We Got Distracted by Assumed Competitive Pressures. Shortly after the initial Jibo team was hired, Amazon’s Alexa launched, and we focused on their success, which was wrong and distracting. Unlike Alexa, Jibo didn’t need features like a kitchen timer or a shopping list. A social robot has one job: to be social in a way that people want to live with.

Lesson: It’s essential to differentiate between the kind of market signals that require you to change your strategy, versus market noise that you should ignore. 

I left Jibo before the robot launched, because I realized that we wouldn’t be able to deliver the Experience Centerline. We had over-invested in areas that pulled us off the course of discovering and learning what the core of the product needed to be. Prior to leaving, I put the names of everyone on the team on a metal plate inside the robot. If you take apart a Jibo, you’ll see these signatures—just like on the first Macintosh. Our signatures became the physical testament to the one thing that we were aligned on: we were trying to create a glimpse of the future.