Written by T. Bishop, Dispatch
Most warehouse management systems sold to small businesses cost between $40,000 and $300,000 once you factor in customization, training, ongoing support, and the inevitable consulting bills. Our entire WMS budget — including hardware — was about $2,000. Here is how we made that work and why we have not regretted it.
The decision.
In 2017 we needed a way to track inventory by location, grade, footprint and customer of origin. We had been doing it on a series of increasingly unwieldy Google Sheets and the seams were showing. We looked at three commercial WMS products, balked at the price tags, and decided to build something ourselves.
The stack.
- Postgres database on a $20/month VPS.
- A small Django app running on the same VPS, serving the dispatcher screen and the mobile-friendly grader interface.
- Two cheap Android tablets at the grading tables ($300 each), running a stock browser.
- A Bluetooth barcode scanner ($120) for pallet IDs.
- A handful of laser-printed pallet cards on cheap label stock.
What it does.
The WMS knows the location, footprint, wall type, grade, and customer-of-origin of every pallet in the yard. It can route inbound loads to receiving lanes, generate the pallet cards, log baler tonnage, and produce the diversion certificates we send to customers. It does not do payroll, accounting, customer relationship management, or any of the other things commercial WMS products try to bundle. We have separate tools for those.
What we wish we had done differently.
Two things. First, we should have backed up the database to a second region on day one, instead of waiting two years until a near-disaster taught us to. Second, we should have hired a part-time developer to maintain the codebase from the beginning instead of letting it accumulate cruft for five years before paying somebody to clean it up. The total cost of those two oversights was probably $8,000 — still less than a single year of commercial WMS licensing.