Four Reasons Most Service Robotics Companies Fail
Wouldn’t it be nice if we had robots to clean the office, greet our clients or put away the groceries at our house? In this day and age of advanced technology we have come to expect – and embrace – the advances that can make our lives simpler. Yet why hasn’t the vision of robotics as dutiful helper come to fruition happened?
The use of robotics is relevant in the manufacturing industry – namely automated factories – but they have failed to make it into our homes and businesses in a meaningful way.
It turns out the challenges facing their adoption outside of the factory are plentiful. Chief among them is the difficulty in aligning the ability to deliver on a consumer or business need without creating operational issues or inadvertently setting unrealistic expectations of what the robot can do. To do that, and at a price point that a business or consumer can actually pay, is a sweet spot few are able to find. And, unless all are aligned, the likelihood of success for a service robot is minimal.
While there are numerous problems robotics companies face in order to succeed in the market, the four main ones continue to be: product market fit, lack of product features or attributes primarily due to software, operational hurdles in market and the inability to set realistic expectations to those that interact with the robot.
1) No Product Market Fit
We’ve all seen those YouTube videos where robots, can do amazing things like climb walls or do back flips. These are robotic accomplishments for sure, but not particularly useful in everyday life. Said another way, sometimes product engineers aren’t being guided by a market need so much as the desire to solve a mechanical problem. The result is an amazing machine that no one wants to pay for. By the time they realize this product development sequence is exactly backward, it’s too late.
Theodore Levitt, economist and former editor of the Harvard Business Review, was famous for telling companies to focus on their customer’s needs and not their own capabilities: “Too many companies focus on how to produce more goods and services—or how to do it better—to sustain the business. But what every company needs is customers, and to get them, you must meet their needs. That’s why marketing or understanding the customer should come first, and the product should follow—not the other way around.”
What would you pay to not have to vacuum your house? This is the fundamental question iRobot answered with its Roomba. It isn’t perfect, but its value to you the consumer is clear. For as little as $200, a tedious chore is effectively perpetually done for you. The value is clear; the product solves a consumer desire to never have to vacuum again.
This is an example of a company understanding a need in the marketplace that they can then solve with robotics. But currently there are not a lot of products – or companies – that seem to understand market needs. Unlike the industrial robotics space, service and consumer robot companies still struggle to either identify or deliver on a value proposition in order to be successful. Think of a welding arm in auto manufacturing. Making a car is an incredibly scripted, repetitive process where the name of the game is manufactured throughput, so robotic welding arms that can meet a car frame at exactly the same three dimensional point in space every time are uniquely suited to adding value to this environment. They can work longer and harder on endless shifts in ways workers cannot to maximize throughput. Thus their cost is compared to the labor savings and low defect rate they deliver to determine an ROI.
Service robots, however, still have a long way to go in showing economic value for the tasks they can actually do. Too often young robotics companies set about solving a robotics problem as opposed to a meeting a customer need and the result is a lackluster performance in market, if they can sell a unit in the first place.
2) Utility Issues - Lack of Product Features or Attributes
Related to market fit is the next robotics issue: the robot lacks sufficient product features or attributes to function effectively in the market. And often the issue is not one of hardware but of software. We have robots that can go anywhere and do just about anything. Where they still have a lot of room for improvement is in the software.
A robot that can’t distinguish between an object and its background won’t be very good at picking objects up. Similarly, a delivery robot that gets lost, or a hotel lobby robot that can’t understand you, won’t perform well either. Now, these are examples of very dynamic machines moving around and performing tasks as opposed to having a programmable coffee maker prepare your coffee each morning (yes, by some definitions your programmable coffee maker is a robot), but this is where we see companies trying to launch solutions and coming up woefully short.
Lets take the case of a robot designed to greet you in a hotel. To be practical in that setting, it needs to be able to see, hear and understand you and then deliver on your needs in that exact moment in time. Probably something to the effect of checking you in or out, telling you about the surrounding area or offering you restaurant options and directions. The benefit to the business would be in freeing up the staff to handle more complex interactions with guests. But to be able to delight the guest and replace a front desk worker, a concierge, a bellman or any combination of the three, requires massive computing power and data sets to work from. Speech and facial recognition software, noise cancelling software, cloud computing, device security software and API access to reservations and other company systems all need to be packaged in a form factor with which you want to interact. This is just part of the software stack that needs to be incorporated for this usage scenario to even operate. That is an incredible amount of work for an engineering team to do even with third party software.
So what is the robot company to do?
Likely, they will try to minimize development time by focusing on just one of the use cases, say, check in, to narrow the product scope, add speed to market and maximize performance. This, in turn, reduces the economic value the robot can provide to the business because it now only performs one task. In doing so, the cost-benefit analysis (the cost of the robotic system vs. the cost of a front desk worker) comes perilously close to zero or even negative ROI rather quickly. Further, if the machine fails more than two or three percent of the time, it won’t make sense for it to be in the lobby of the hotel. So by not having enough product features, the economic value added decision to invest in the service robot becomes insurmountable. Why buy a robot to do what the staff can do better and with less friction?
3) Real World Operational Hurdles
Further complicating matters of adoption rates, at least for business focused service robotics, are the operational hurdles these machines face out in the world. Think about a typical business setting, a retailer perhaps, or a warehouse. Each of these settings has already been optimized for efficiency and effectiveness, sometimes many times over. Adding new robotics technologies that somehow add cost or inefficiency to adopted processes is usually a nonstarter for the businesses.
If you are retailer operating on single digit profit margins, you likely won’t risk adding operational complexity. If a robot can’t deliver a package or move around the business without someone having to follow it, then its use creates operational hurdles. Further, someone has to be responsible for the robot, maintain it, keep it in good working order and make sure it is charged. If that requires an extra headcount, then the value equation is likely thrown off and a business will choose not to adopt the technology. Burger flipping robots work best when the kitchen is built around them, not by bringing them into an existing kitchen and trying to retrofit. This is one of the most difficult things for robotics companies to overcome in the sales cycle, the operational disruption their solution creates.
4) Expectations
Finally, though software is advancing rapidly, it is still not yet able to deliver on expectations formed by pop culture. Rosie the robot from the Jetsons, C-3PO and the machines in ExMachina are the benchmark set in people’s minds for what robots should be able to do. But the reality is just to have a robot understand you in a public setting is an incredible challenge. In the hotel lobby scenario above, the guest wants the robot to be able to check her in, tell her about the property, recommend food and shopping options around the hotel and potentially book appointments. That is what she would expect a robot to be able to do, especially if it is autonomous and built to look like a member of staff (i.e humanoid).
But to get a robot to do all the things we want it to do based off what we’ve seen in the movies is not technologically possible, at least not at a market affordable price. And this is the final dagger in most service and consumer robotics company’s hearts – performance rarely lives up to the expectations. This is one of the main reasons the home robots – like Anki’s Vector, Kuri and Jibo – ultimately failed. You can’t just have a conversation with them, but that’s what we wanted. Cute robots like these don’t ultimately work well in home or business settings because they don’t deliver on the need. They either need to do chores for you or they need to be a companion, in which case they need to be able to interact with you and converse with you. If they don’t, well then Amazon’s Echo is all you need. As Matt Simon wrote in Wired “Vector, Jibo, and Kuri show that you can’t sell a home robot on adorableness alone.”
This is not to say robotics companies won’t continue to chip away at the problem, but the likelihood your job is at risk from a robot is not high.
Bottom Line for Success
It is incredibly difficult to get this all right; however it has been done. iRobot is arguably a good example. A more recent one would be Simbe Robotics’ Tally. Tally is a shy robot designed to blend into its surroundings moving slowly and methodically. It is designed to tackle the massive problem of out of stock items in retail stores. Tally checks the store three times a day and relays the information back to the staff to help them make sure the items we are looking for are all out and priced correctly. This frees up the staff to focus on stocking and helping shoppers. The important point is they identified a massive industry problem, keeping on top of the inventory position in store, and delivered a robotic solution that works flawlessly.
So what can other companies learn from Simbe? What do they need to do to be successful?
First, they need to do a better job finding customer problems they can solve well with the technology available now, like industrial robots do. Second, robotics companies need to narrow the scope of their offering to something both the hardware and the software can deliver on 100 percent of the time with little or no intervention. Finally, these companies need to do a better job of level setting the expectation that the form factor they choose (e.g. humanoid) sets. A robot that can move around autonomously, has a much bigger expectation put on it then your coffee maker does.
