5 counterintuitive lessons I learned building a hardware startup

When we first spoke to people about building an intelligent standing desk, we knew that integrating both hardware and software would be hard. We set ourselves to the challenge and now a year wiser, we thought we would share some major (and rather counterintuitive) lessons learnt along the way.

1. Hardware development cycles doesn’t take days but weeks

While we respect the lean methodology that encourages rapid iterations on a minimum viable product (MVP), the reality is that hardware development cycles typically take a longer time compared to pure software counterparts.

For hardware, developments in 3D printing and prototyping boards such as the Arduino or Raspberry Pi means that prototyping small plastic parts and electronics is much faster and cheaper than before. However, unless your product can be wholly printed via a 3D printer or made on a board, you have no choice but to resort to conventional prototyping.  This means a painful 2-3 weeks wait for the next version of your prototype to arrive from the factory. This is the reason why hardware development takes 3 times as long and often costs 3-5 times as much compared to software.

Also, hardware is still a physical product. While prototypes can be made out of cardboard or foam, people are accustomed to products they use daily, and naturally apply their expectations of usage, looks and quality, to your prototype. It is no fun (and believe me, we’ve tried) asking users to imagine that a hacked-together aluminium frame will eventually automatically elevate and bear the weight of a 27-inch iMac.

As an example, pictured are our first prototype (left) and our final prototype (right)

2.    It is not about the data science (not at first)

People will physically interact with your product on a daily basis. It is something that they will see, touch, use, and feel. Therein lies the beauty and also the challenge of hardware.

If you had to focus on only one thing, it should be building standalone superiority within the product itself. This excellent post from Calvin Chu explains it well, but in a nutshell, it’s about building into your product a design, quality, and brand that people can emotionally relate to. This is why Apple’s products have outperformed those of other brands on many levels.

Data science is the holy grail of most IOT startups. However, the data is meaningless without a large set of data from numerous products used on a frequent basis. There is really no point collecting data upfront if you do not figure out how to get customers to use your product habitually. And let’s face it, people will only buy and habitually use your product because they love the way it looks, feels, and works. So, focus on building a superior product first but keep in the back of your mind that the eventual data streams you can build your analytics on.

3. Conventional methods work the best

Technology conferences are fun to attend. But they are not where the two most important sets of people would be – your contract manufacturers and your customers.

On the advice of friends, we attended furniture fairs in Guangzhou and Kuala Lumpur. It was an eye-opening experience that taught us the industry dynamics (spoiler: outright copying and price competition is rampant), gave us insight into typical minimum order quantity (MOQ) and pricing, and enabled us to shortlist the contract manufacturers whom we subsequently looked to work with. Nothing beats seeing 3,000 furniture companies side-by-side, assessing the quality of their exhibits and having face-to-face conversations with the factory owners.

You should also be where your customers are. In our case, we ran an ecommerce business selling third-party gas-spring standing desks imported from the USA. We made it a point to personally deliver and install every single desk we sold, so that we thoroughly understood the customers’ needs and their pain points. We also had the privilege of attending HR fairs organized by our partner and spoke directly to decision-makers. Aspirus was developed from what we learnt in those customer interactions.