Five DO's and DON'Ts on Building Your Robotics Startup

A VC's perspective

4 min read
Five DO's and DON'Ts on Building Your Robotics Startup
System integrators like Dematic, which fully automated this warehouse in Weertzen, Germany, help large businesses with their robotic needs. Robot entrepreneurs should get to know these integrators, which can provide insight into how current technologies come short and what processes customers are looking to automate.
Photo: Dematic

It is a great time to start a robotics company. Cheap components, open-source code, and accessible options for contract manufacturing empower roboticists to focus on designing great robots instead of sweating over non-core details. Indeed, we're seeing innovative robotics entrepreneurs introducing beautiful, functional robots. Unfortunately, many of these startups will likely have a difficult journey leveraging their robots towards building attractive businesses—businesses that venture capitalists like myself would want to be shareholders of. Last month I discussed myths and realities of robotics technology. Today, I present a short list of do's and don'ts for robot innovators, based on what I've been seeing in this industry as it grows and matures.

And if you have your own collection of do's and don'ts, please share it in the comments below. 


It might sound counterintuitive, but in robotics, as in any hardware company, you’re better off building nothing vs. burning through your scarce resources to prepare a robot nobody wants. Unlike mobile apps, iterating on robot designs is an expensive endeavor. Many fledgling startups are quick to design robots without an understanding of how, or why, customers will use them. Although it is true that many customers want to see something tangible to make sure a new company is for real, many entrepreneurs go to the extent of designing and building almost-production-ready robots before understanding their customers' needs. In "The Hard Thing About Hard Things: Building a Business When There Are No Easy Answers" (HarperBusiness, 2014), Ben Horowitz, a co-founder of the venture capital firm Andreessen Horowitz, says that "figuring out the right product is the innovator’s job, not the customer’s job.” The problem is many robotics entrepreneurs tend to take this notion to the extreme, investing considerable time and money but in the end producing a robot without a clear value proposition for the customer.


Many robot startups tend to underestimate and even trivialize the rigor that goes into introducing toys and other consumer products. Focus groups, conjoint analysis, and of course, many, many iterations of a product—the majority of which fail. In fact, when we think of successful companies, their most successful products often come to mind and we tend to overlook their many failures (because these don't receive as much publicity). Mattel, for example, may conjure images of the successful Barbie franchise, but the survivorship bias may cloak the many iterations of the Barbie products, as well as the countless other attempted product lines into which significant money was invested with no success. Robotics startups offering consumer products without the big budgets or manpower need to get creative on how to engage early and properly predict what features customers will pay for, and how large of a market that translates to.


The modern toolsets available to roboticists can be as much of a curse as they are a treasure. As we’ve seen in the game console wars, specifications do not drive sales—it’s the blockbuster titles available on the platform. Everyone remembers Super Mario, but nobody asks whether the Nintendo console offered a larger sprite size than the Sega Master System, because the hardware in the devices came from a known cohort of semiconductor vendors. The same applies to robotics. The hardware and underlying software are similar, and there is a limited amount by which a small group of extremely talented entrepreneurs can push ahead of the status-quo. Therefore, and this reinforces the previous point on engaging customers early, entrepreneurs need to understand the angle they need to bring to bear to get a real competitive advantage, rather than trying to be everything to everyone, or even worse, nobody.


This point especially applies to roboticists targeting manufacturing and industrial automation. The automation market has a “barbell” profile today (big at the ends, narrow in the middle). At one end, there are the large, expensive robots that help to churn out automobiles and appliances, a market dominated by giants such as Swisslog, Kuka, Fanuc, and ABB. The GMs and GEs that make up this end of the market fork out major $$s and hire armies of engineers to deploy these extremely accurate and reliable robots whose roles are key to the fundamentals of their businesses. At the other end of the barbell, you have extremely price-sensitive, giant contract manufacturers such as Foxconn, which are seeking alternatives for the high-turnover, mind-numbing human tasks. These tasks include loading/unloading machines, assembly, and basic inspection. In many parts of the world, every labor “shift” is an $8k annual expense, so a “work cell” that operates on three shifts costs $24k annually, if operated by humans. Dramatically shortened product cycles bring the expected ROI on a production robot down to a year, so a robot that is expected to perform a function that would otherwise require humans needs to cost less than $24k, including setup, calibration, and maintenance. This is not an easy target.


Again, more emphasis here on manufacturing and industrial automation. Many robotics entrepreneurs neglect to appreciate the importance (and in many cases the existence) of system integrators that manufacturing companies and logistics companies rely on to select, procure, and maintain their automation systems. Companies such as Wynright, Symbotic, Bastian, and Dematic (a warehouse automated by the company is picture above) aren’t quite household names, but the likes of Walmart, UPS, FedEx, GM, Tesla, and Flextronics rely on them for automation solutions. Get to know them. There is a good chance they already a) know how current technologies come short and b) what processes customers are looking to automate. They are constantly on the prowl for new technology to showcase to their clients and differentiate themselves from their competitors. Use their relationships to accelerate the definition and adoption of your product by big customers, and with that, the success of your robotics company.

Shahin Farshchi is a partner at Lux Capital where he invests in hardware and robotics companies. He is based in Palo Alto, Calif.

The Conversation (0)