$49.5 billion

value of lab supplies market by 2027

3,461

lab supply manufacturers

7%

anticipated industry growth by 2027

clients_ThermoFisherLogo

EXECUTIVE SUMMARY

TensorIoT and AWS helped leading developer of research equipment calibrate manufacturing schedules by creating more accurate product demand forecasts that leverage sales, price, and promotional data.

GOAL

How can we use existing demand data to more accurately predict production needs, while keeping costs low?

RESULTS

TensorIoT and AWS built a cost-effective solution to accurately predict demand forecasting using Amazon Forecast.

Background

A leading developer of research equipment was looking for a way to better sync manufacturing schedules with production needs. The ability to accurately predict future demand has the potential to save the client tens of millions of dollars in manufacturing costs. As a growing company, the client needed a solution that was scalable, highly accurate, and cost efficient. The TensorIoT and AWS machine learning (ML) teams were perfectly positioned to meet this need.

The Challenge

As with other ML projects, some of the big hurdles that the team faced were due to challenges in engineering the data. The client had large quantities of data that needed to be processed - millions of records and dozens of features. Additionally, our solution had to compete with the state-of-the-art industry benchmarks.

The Solution

Amazon Forecast took in site-specific manufacturing data, as well as a host of metrics surrounding historical sales and product demand, and then used machine learning to predict optimal production schedules by manufacturing site. Our Amazon Forecast model allowed the client to drastically improve their elasticity to demand—reducing over and under-production, and optimizing stock and inventory space, while cutting down on waste. We were able to achieve higher accuracy than many of the more established demand forecasting solutions on the market, at a lower price thanks to Amazon’s pay-as-you-go services.