Foot for thought
March 27, 2024
I just finished the final project for Microsoft PowerUp. Given a list of shoes in a spreadsheet, our job was to build a solution for a shoe store.
My approach was to start from the back and move to the front, in this order:
-
Back-end database tables and a model-driven app
-
Automated flow for notifications on inventory purchases
-
Front-end app for employees to set discounts on products
-
Front-end app for customers, including shopping cart
-
Invoicing and purchasing flows for inventory and accounting
-
Customer-facing chatbot to query the database
-
Management dashboard including reports on popular items and slow-moving items
If I were to do it all over again, I would take my Grandpa's advice to start from the front and move to the back.
Only by working on the management dashboard did I fully understood what I needed to calculate inventory turnover rates. Although model-driven apps make it relatively painless to retrofit tables with new columns, it would have been cleaner to do so from the start.
And it was only from building the customer-facing chatbot that I realized that I needed a better approach for shoe sizes. Initially, I had started with just a simple text field in the SKU table. Then, after learning all sorts of exciting trivia about shoe sizes, I created a system that can assign any SKU to multiple standardized sizes from multiple sizing charts (US, UK, EU, Mondopoint, plus width charts). The data structure also allows a knowledgeable salesperson to record sizing data to offset the posted size for shoes that run too large or too small.
I like the bot-first approach to software development.
By imagining the questions that someone might ask a bot, you'll think hard about how you're going to reply.
"What do you have in my size?"
Simple question. And it's not easy to answer.
And it's an important question, one that goes to the essence of personalization.
Personalization isn't knowing everything about the customer.
It's about knowing your own business, front to back, top to bottom, and inside out.
Only then can you put intelligence in the middle to fulfill the promise of personalized service.
Want to see what I built? Here's my 10-minute demo.