Wayfair: Internal Technology Challenges

This is a snippet of the transcript, sign up to read more.

Interesting. Can you explain in more detail?

The planning systems were based on historical data, such as seasonality demands and parcel movements. Once they had that data, they would create dummy containers and provide them to execution. When warehouse personnel received such an order, they would fill up those empty or dummy containers. Many people were lobbying for a clean system where planning only does planning, and execution takes over the actuals. However, the legacy system dictated what items needed to go into a container, which became a nightmare to integrate with Manhattan.

This is a snippet of the transcript, sign up to read more.

And that's because the legacy systems that Wayfair had built divided up those tasks differently than the industry typically does, right? So, right now, what parts of Wayfair's logistics network are still, as far as you know from when you left, on home-built or legacy systems? And which parts have been outsourced to industry-standard systems?

I believe most of the systems are still in legacy. There is no clean transfer over to TMS. I know there were parts of middle mile that were moved to TMS. However, when I left, all these systems still interacted with the legacy system. Even if we didn't use the front-end legacy system, we needed the existing systems to transmit data to the yard management system, warehouse management system, and the data science team. Essentially, we just changed the UI from a legacy system to the Manhattan UI, but the backend was still using the legacy system.

This is a snippet of the transcript, sign up to read more.

Why is it difficult? What makes it so challenging?

It's a very patched-up system. If someone tries to decouple or detach the system, something would break. So everybody is scared to take that risk. Nobody wants to be the person who might be the fall guy. It's very slow and cautious, and I think it could have been more efficient.

This is a snippet of the transcript, sign up to read more.