top of page

Why "selling robots" isn't a very profiting idea, and what are more reasonable business models?

Figure 1. The companies responsible for Jibo and Cozmo are both out of business.

In 2019, we've witnessed the fall of many robot-selling companies, including Jibo responsible for the family robot with the same name and Anki that made the AI toy named Cozmo.

While we’re mourning for them, the more important thing for the rest of us is to learn from their mistakes. And after a careful inspection, one should derive the following conclusion: “Selling robots” is actually not a very good business model, especially for startups (note that the “robots” we’re referring to here are those made for individual customers or regular families, not industrial robots).

In this article, we will explain how we obtain the conclusion said. Then, some more reasonable ways that a company can profit with robots will be discussed. Hopefully, you will find the information in this article useful in respect of building your own robot business.

Why is “selling robots” not a very good business model?

The major problems for a company to take robot selling as their main business are as follows:

  • These products cannot compete with existing AI assistants or smart home products released by tech giants:

Many companies try hard to blend their robots into regular families by providing functions such as online shopping, phone call making, information searching, etc., which can help one to go about his or her day. The problem is that those functions have already been achieved by the AI assistants or smart home products (e.g., Google Assistant, Amazon Alexa, and so on) that are created by the more renowned companies. And due to the differences in fame, resources (e.g., amount of data, the quality of engineers, capitals, etc.) and experiences, it’s really hard for a startup to compete with them, if not impossible.

If you want to know how fierce the competition of this area is, considering the following fact: Even a company as big as Microsoft has closed their Cortana apps across multiple markets.

  • It’s unlikely that people will continue to buy the newest version of these robots:

Like electronic devices such as iPhones, these robots are difficult and expensive to manufacture. However, unlike iPhones, it’s improbable that people will keep buying the newest version of these robots, and that will bring serious problem to the companies’ revenue. This is especially true if these robots are promoted as toys (it means that they are dispensable) or “family members” (if you truly treat a robot as part of your family, it will be hard for you to replace “him” or “her” with a new one).

The reasons for this to happen are as follows. First, they are not something like smartphones or computers that you’ll have to use every day, so their failure rate may be lower. And even if they really fail, since they are not essential to life, you won’t feel the urge to purchase a new one.

Second, devices like a smartphone or laptop can gratify your vanity because you’ll bring them outside while everyone can see them. And for that ground, some individuals may feel the need to purchase the newest model (of a smartphone or laptop) from a premium brand. Most robots today, on the other hand, stay at home where nobody can see them, so you will not feel the necessity to do the same.

  • Most robots today are replaceable:

As mentioned previously, most robots are totally replaceable by other things at least for now. For instance, the entertaining values they proffer can be substituted by the Internet, computers or video game consoles, and their role as a family member can be substituted by a real pet, etc. If any company wants to make robot a dominant electronic product in the future, finding the irreplaceable position(s) of a robot in people’s lives will be one of the most imperative tasks.

  • The design of most robots lacks reason:

As all good product designers know, the shape of a product should support the product’s function. In other words, a product should be designed in a way that its function can be easily accessed by the users. With that in mind, we must ask: What functions do the shapes of robots, especially “humanoid” robots, support?

In our opinion, thus far, this question is poorly answered. That is, most of the functions that one put on a robot are better realized by other devices with an appearance more reasonable for carrying out these functions. For instance, want to do online shopping? We’ve already had computers or smartphones for the job, and their designs allow us to better inspect the things that we’re interested in before we actually purchase them.

With that being said, there are some robots whose designs justify their functions perfectly, such as robot vacuums, drones with camera and so on.

Figure 2. ECOVACS DEEBOT robot vacuum.

Figure 3. DJI Mavic Pro drone.

  • These robots are not smart enough yet:

Although, thanks to the development of machine learning and other AI techniques, the robots today can interact with its owners somewhat smoothly, but these interactions are far from perfect.

As a matter of fact, once a device is dubbed “smart” and is given the ability to interact with its users in a humanlike fashion, as many robot companies will do to their product, the customers tend to expect human-level intelligence from them, and that usually leads to disappointing experiences because the artificial intelligence nowadays is just not there yet.

Granted, AI has already outperformed humans in some very narrow areas, such as image recognition or detecting breast cancer. Overall, however, we still have a long way to go to make them actually humanlike, which requires them to possess some degree of “general” intelligence. And before that happens, people may rather purchase a regular machine (like a laptop) specially designed to get certain tasks done than actually buying a robot.

What are better ways that robot companies can make money then?

So, if selling robots directly doesn’t work well, how should a robot-making company earn money then? Below, some business models that we consider to be more reasonable and profitable are mentioned. Hopefully, some of them can inspire you to come up with your own idea.

  • Educational programs:

With a well-constructed app or online application, robots are a very good teacher for educating children how to code. So, instead of selling robots straight, it may be more remunerative to sell programming courses, in which robots only play a facilitating role (e.g., to attract children’s attention or motivate them).

There are multiple benefits for doing so. First, parents may be more willing to pay for a course than simply a “toy”. Second, programming allows children to release their imagination and create a robot that truly belongs to them, making them less likely to get tired of these machines.

Note that the business model of the company named Sphero is similar to the idea aforementioned (though not exactly the same since it’s still focusing on selling robots). You can see this page to learn more information about it.

Figure 4. The Edu app made by Sphero that can teach children how to code, which significantly increase the usefulness of the company’s robots.
Figure 4. The Edu app made by Sphero that can teach children how to code, which significantly increase the usefulness of the company’s robots.

  • Robot playground for children:

Instead of enabling children to bring robots home after (their parents) paying, maybe keeping these robots in certain locations, “playgrounds” to be precise, will be a better idea. In other words, people no longer pay for the robot, they pay an entrance fee so that their children can play in the “robot playground (much like the robot-version of Disneyland)”.

In fact, this idea can combine with the idea previously mentioned to create a gaming experience with educational values. For example, children can realize their own fancies of how the future lives should be like by choosing and programming different types of robots (with the assistance provided by the instructors on the spot).

Such business model can be better than robot selling due to the following reasons. For starters, children will have to be at a certain location to play with these robots, and such fact can prevent children from getting fed up with a robot toy their parents have purchased.

Second, the manufacturers do not have to mass-produce their robots; they only have to prepare the number sufficient for satisfying the children who visit their playgrounds, which can be relatively easy to estimate by considering “how many playgrounds that the company owns” and “how many children visit the store at a given moment on average”.

Finally, it can save the company a lot of money to make their robots extremely smart. In the robot selling model, where people will take home the robot they’ve purchased, such robot probably has to be intelligent enough to keep its players entertained. But in an area like the playground described here, people are entertained not only by robots but other factors (e.g., role playing experience, showing off the robots they’ve programed, etc.) as well, and that may decrease the level of intelligence that people will expect from the robots themselves.

  • Robot for rental:

Following the logic of the above section, comparing to robot selling, robot “renting” may also be a good idea because people no longer have to spend a great deal of money to purchase a robot that he or she may or may not enjoy in the future. With rental service, they can try out a robot first with a relatively lower price, and return it if it fails to satisfy them.

Also, people are given the freedom to adopt a robot only when they need one, which may significantly increase their willingness to make use of this technology.

  • Focus on a specific task:

Speaking of robots, people naturally anticipate them to have the ability to process natural-language commands and carry out humanlike behaviors. Simply put, people expect them to be “smart” (especially when the robot’s appearance is android). However, as mentioned before, the AI today is not quite there yet, and such high expectation usually brings nothing but disappointment.

Thus, instead of doing so, it can be better to explicitly tell the customers that the robot is only good for one certain task (or a few certain tasks) and drop the adjective “smart” in the marketing language. That way, people can have a more accurate anticipation towards your product, and you may also be able to cut a large number of expenses for perfecting its intelligence.

Note that products such as robot vacuums and camera drones are two representative examples of the kind of robots described here.


With the development of artificial intelligence and other technologies, non-industrial robots may become more and more prevalent in our lives. But before that can happen, we think it’s crucial for companies that make robot for regular users to revisit the question: How can they better profit through their robots? The following are some dos and don’ts in respect of such question, and we wish they can help you to refine your business model:

  • Don’t make humanoid robots. Instead, giving them a shape that they can best carry out their designed functions.

  • Don’t emphasize that they are “smart”. Instead, explaining what values can your robots provide.

  • Don’t make robots that can only be used as a toy. You should give them extra merits, such as the educational values aforementioned.

  • Do find a market place or market places where Google assistant, Amazon Alexa, Apple Siri and other smart assistants from tech giants are not your competitors.

  • Do ask yourself the following question: Why your product has to be a “robot”? Why can’t it be an app, a computer, or anything else? If you can’t answer the inquiry, your idea of a business may be suffering from some serious weaknesses.

Footnote: Neurozo Innovation provides viewpoints, knowledge and strategies to help you succeed in your quest. If you have any question for us, please feel free to leave a comment below, or e-mail us at: For more articles like this, please join our free membership. Thank you very much for your time, and we wish you a wonderful day!

1,179 views0 comments


bottom of page