Engineering9 min read

Building Real-Time Dashboards for Logistics Operations: Lessons Learned

Moving a logistics company from spreadsheets to a live operations platform revealed surprising complexity. Here is what we learned building the system — and what we would do differently.

Ishanix Team

Full-Stack Engineering · June 5, 2025

The Problem With Spreadsheets at Scale

When a logistics company is small, spreadsheets work fine. When you are managing 300+ active shipments across 12 carriers with 5 people sharing a Google Sheet, things break down fast.

We were brought in to replace the spreadsheet entirely. Here is what that journey looked like.

Lesson 1: The Data Is Messier Than You Think

Every carrier had a different API. Three of them only had FTP exports in CSV format. One had no API at all — just email notifications. A significant chunk of the initial project was just normalizing data into one schema before we could even think about displaying it.

Takeaway: Always budget 30–40% of a data platform project for data normalization and ingestion pipelines. It is never as clean as the client thinks.

Lesson 2: Real-Time Has a Cost

True real-time (WebSockets, server-sent events) is expensive and complex. For most logistics dashboards, "near real-time" — polling every 30 seconds — is indistinguishable to the user and dramatically simpler to build and maintain.

We started with WebSockets. We moved to polling. It solved 90% of the complexity.

Lesson 3: Operators Want Actions, Not Just Information

The first version of the dashboard showed a lot of data beautifully. The ops team loved it for about a week, then the requests started: "Can I reassign a route from here? Can I flag a shipment?" They wanted to act, not just observe.

The second version embedded actions directly into the view. That is when adoption really took off.