Move Product

• GIF

Inspiration

E-Commerce Entrepreneurs are running 6-8 figure brands as a team of one or two people. They often sell 100s to 10Ks of SKUs. Many of them are non-technical. Meanwhile, all of their business data is disparate. Ads require integrating with Google Analytics, FB Ads, etc.. Supply Chain data requires integrating with Ali Express or other product sourcing APIs. And customer transactions are in Shopify or some other app. They can't afford devs to build these integrations. In this hackathon, we prototype an app where E-Commerce Entrepreneurs can run their business from anywhere in the world with via SMS. Underneath the hood, state of the art machine learning algorithms are building SKU level forecasts, calculating product margins, and optimizing future cash flow based on real time data.

What it does

Our project produces SKU level 12 week ahead forecasts for 3K+ SKUs. The analytics and forecasts can be consumed in a traditional way via dashboards or via SMS (prototype interface) where end user can query real-time data and receive real-time optimizations. Business owners can literally make hundreds of thousands sitting on beach and interacting with their ML powered Chief of Staff.

How we built it

We used Python and Pytorch for ML training. We deployed our forecasts using Plotly/Dash on Heroku. We prototyped the interface in Voiceflow.

Challenges we ran into

• Building semantics into chatbots is very difficult but can be done with more time and especially sitting down with business owners (for this use case) and figuring out their top queried items. ## Accomplishments that we're proud of
• We matched and trained state of the algorithm in Forecasting and built 12 week ahead supply chain forecasts for 3K+ daily time series with 5 years worth of training data.
• We experimented with and prototyped a new interface for users to interact with ML Algorithms via SMS/Text.
• We built SKU and order level metrics for calculating profit.
• We deployed and shipped a clunky MVP!

What's next for Move Product

• Build real-time integrations to get this data.
• Add more semantics and complexity into the chatbot.

Built With

• dash
• heroku
• plotly
• python
• voiceflow