Proof of Concept: Inventory Management Modeling

Industry Manufacturing

Specialization Or Business Function Supply Chain and Logistics (Supplier Analysis, Supply Chain Optimization, Sales Forecasting, Inventory Management), Actuarial Science (Predictive Analytics), Business Analysis (Requirements Analysis, Requirements Management, System Analysis)

Technical Function Analytics (Predictive Modeling, Machine Learning)

Technology & Tools

COMPLETED Jan 14, 2019

Project Description

About Us: We are a metal distribution company with an extensive inventory within our metal market niche. We satisfy customer needs by stocking both commodity and hard-to-find grades, and by providing extensive value-added processing and custom cutting based on customer specifications.

The Problem: We have an extensive inventory offering which includes 35 different grades of steel in 35 different thicknesses and are looking to build out our existing inventory management system to assist in demand forecasting to avoid overstocking while also ensuring we have product availability to meet customer demands. There is a lot of demand volatility and pricing volatility, as well as the mills we source from could have anywhere from 4 weeks to 6 months+ lead times depending on grade, mill location and current demand. Inventories are currently replenished based on historical 3-month, 6-month and 12-month usage data for a particular grade and thickness.  In addition, there may be minimum sizes that must be ordered for certain grades and thicknesses which can far exceed the weight that would be ideally ordered based on historical usage.  All of these factors may result in overstocking on our end to ensure we can meet customer demands, or under stocking for unforeseen surges in demand.

Deliverables: This first project consists of 2 goals. In your proposals please break out the cost and effort you expect each step to take.

1. Experts who understands inventory demand forecast will look at our systems and data available to better understand the roadmap and if we have everything needed to build the model.

2. Proof of concept: Build inventory and forecasting algorithm for a specific grade of material, taking into account current mill lead times, to ensure we have material in stock. Ensure the algorithm is not static, but develops over time to ensure that as market conditions and product demand changes, the algorithm develops as well. In addition, assistance with deploying the model within our infrastructure.

While these are the two goals for this initial proof of concept project, if the model is successful we envision the project expanding to other grades within our inventory.

Data sources: We have been collecting data for over a decade and can provide data in csv format. Data is not too much about 1000+ lines of data except quoting activity data. We have our own custom inventory management software that runs on SQL server. We can provide data like

• Historical inventory level

• Historical demand by month

• Quoting activity, quote conversion, and sales activity data

Metal prices are publicly available, and would help to look at supply demand correlation.

Please propose how you would tackle this project and an estimate. We would also like to know your past experience performing similar work. Lastly, we would require a non-disclosure agreement to be signed as part of the project. Applicants preferred from USA, UK, Canada. 


Project Overview

  • Posted
    April 25, 2018
  • Planned Start
    May 16, 2018
  • Preferred Location
    United States

Client Overview


Matching Providers