One of the first technical decisions we had to make when starting Joymode was whether we were going to build our own warehouse management system, or go with an existing one. My cofounder Keith put it best when he pointed out that we were essentially the customers of our own system, and that system could either be homegrown or bought.
We decided to build our own Warehouse Management System because most WMSes are built for e-commerce systems where companies are shipping out product. But at Joymode, our operational flow is unique, and that led to two key realizations:
- When you’re building a system that focuses on shipping out products, the return flow becomes more of an afterthought, which makes sense for most e-commerce companies. After all, returns make up a single digit percentage of their transactions, so focusing on the shipping flow makes sense from their perspective. But for us, 100% of our items that go out, end up coming back in, and then need to be processed, cleaned, and put back into inventory. Building our own system gave us the flexibility to integrate both physical and digital processes so that we could be more efficient in cleaning and minimizing item turnaround to get it back out to customers faster.
- Most existing systems will track the product count of a SKU, but not give you item level details. Again, when you’re focused on selling items, there’s no difference between one unit and another since they’re functionally equivalent. For us, though, we can determine how many times each individual projector has gone out, what condition it was returned in, and what its expected lifetime will be. All this data helps us make more informed decisions throughout the company, from merchandising to inventory to cleaning, and without a system to track all of that, we wouldn’t be able to make any of those optimizations.
There were a few other considerations (attracting an amazing engineering team, being able to work closely with our operations team to spec out exact requirements, etc.) but those were the main ones we considered in building out our system.
In future posts, we’ll talk more about other architecture that we chose as a result of this key decision.
