Jabil Balancing Risk In A New Frontier

Krishnan Ramanathan, Sara C Wolski, Zain Nensey, Raniel Marc Barcenas, William Nowel
Muma Case Review  •  Volume 5  •  2020  •  pp. 001-022
Candy Mitchell, Information Technology Director at Jabil Inc., sank into her desk chair and looked out the window pensively. She had just finished reading a report from her team of data scientists about a predictive model they had created. She saw huge potential in this model as a potential solution to power Jabil’s customer growth strategy. Part of the strategy was to engage with technology startup companies to be their partner of choice from ideation to supply chain management.
Working with technology startups would be new for Jabil and could create huge returns (when one of those technology startups took off, so would Jabil). Typical customers for Jabil (a manufacturing solutions provider that delivers comprehensive design, manufacturing, supply chain and product management services) ranged from FORTUNE® 500s to governments.
Mitchell knew how risky startup customers could be, and Jabil only wanted to work with financially stable companies. To serve Jabil’s new strategy, Mitchell placed priority on finding a way to reliably vet the startups’ creditworthiness to avoid bad debt. The third-party credit rating agencies Jabil used could only provide a score for large companies with accessible financial information and history.
Mitchell’s computer pinged with a calendar reminder – it was almost time for her next update meeting with her data science team. So far, their model had achieved an accuracy rate of 86%. This was promising but not yet good enough. Mitchell knew they would have to address this – and soon. The faster a way to test their financial viability could be found, the sooner Jabil could partner with them and profit from them.
Mitchell wondered what impact adding more quantitative and qualitative data would have on the creditworthiness prediction model. She knew it could lead to higher accuracy, or it could lead to a redesign, which meant more time and resources. The model had to be accurate, explainable, and auditable. Was it possible to create this model internally at Jabil?
Risk, Credit Ratings, Model, Credit, Electronics
30 total downloads
Share this

Back to Top ↑