What makes Home robots succeed or fail?

Reviewing a research paper in the context of Anki Vector

Researchers led by Prof. Selma Sabanovic in Indiana University recently published a paper in the International Journal of Social Robots, which they kindly gave me access to. The authors did great market research and analysis on what makes home robots succeed or fail based on data from 448 products sold via Kickstarter/Indiegogo over an eight year period (2011-2018). I found this paper to be an excellent epitome of the kind of background work that entrepreneurs ought to do while trying to bake a new product design. The authors categorized and labelled the data, performed a variety of statistical operations, and did sentiment analysis on comments left by product users to generate some very interesting results. Here are some interesting conclusions from this paper on what specifically does and does not work when it comes to home robots.

• Having the ability to take videos with a camera is one of the most important features people expect in home robots. Integration with other cameras such as home security cameras adds value. People who wanted camera features were also very concerned with privacy. This is interesting in the context of Anki Vector as we all know how much Anki invested in the security features of Vector. Anki actually hired a very experienced security expert, Talha Tariq, to design the security of Vector.

• Performing the functionality equivalent to Amazon Alexa is one of the basic expectations people have of home robots. Again in the context of Anki Vector, we witnessed how Anki engineers hurriedly implemented and released Vector’s Alexa integration as their first feature after Vector’s launch in October 2018.

• People like different color options and upgrade options. Black, blue, and white were the preferred colors. People not only desired upgrade options but were inclined to pay for them. This is an area which Anki may have missed, Vector came in a single color and without any upgrade or subscription options which could have paid for the services that Anki was hosting to run Vector. On the other hand, Petoi Bittle does a great job in this regard by providing lots of upgrade options, although it did make manufacturing much more complex and packaging much more tedious.

• The study found that heath and fitness, education, and home security would have the largest number of backers. The study also found that education specifically for computer programming had the highest number of projects launched but didn’t secure high number of backers. Instead, people expected education focussed home robots to help children with language, science, social development, and general knowledge. In the context of Anki, we saw that Anki Cozmo had a very high adoption and success, possibly because of it was better in helping children with the above skills.

• The authors found that single person use robots are likely for successful compared to family use robots. Similarly, single function robots would have more backers with multi-functional robots. The latter is an important conclusion because it indicates market preference for single purpose cheap robots as opposed to multi purpose expensive robots. This conclusion may be biased by the fact that most backers in Kickstarter/ Indiegogo are men. Designers of new robots may therefore devote stripping down features to the simplest minimal viable product (MVP), that could be priced cheap.

• The authors suggest that animal form robots have more success of adoption than human form or machine form robots.

• While analyzing for product failures, the authors found many failures were due to the products not satisfying the needs of early customers (e.g. Jibo) or due to products not having sufficient non distinguishable features compared to non robotic products. This was certainly true in the case of Anki Vector, as it lacked a concrete killer use case. Most of the functionality provided by Anki Vector, could actually be done in a much easier way with other devices.

In my opinion, the authors definitely succeeded in showing how crowd-sourced data can be used to supplement other avenues such as consumers surveys when it comes to designing products. This approach has some distinct metrics and advantages because it captures a customer action (in terms of actually buying the product and potential using it). Similarly sentiment analysis of comments and definitely yield a trove of information, which can be used for future product design.

If you get a chance to read this work, please provide your thoughts and comments below. Thanks!