Signify Builds a Supply Chain Design Capability That's Ready For the Next Decision

Signify's supply chain design team used to validate decisions after leaders made them. It is now building a capability that tests the options first: reusable data and models, faster iteration, and custom analytics for questions that don't fit a standard design model.

"A study answers one question. A capability prepares us for the next one." — Sander van Schoonneveldt, supply chain network modeling specialist, Signify

At a glance:

‍

The challenge: every study started from scratch

Before 2020, Signify, formerly Philips Lighting, had the expertise for supply chain design, but it was scattered. People in distribution ran one-off studies on the distribution footprint. Industrial engineers in manufacturing ran footprint optimizations. The questions kept coming, because Signify's footprint had been changing for 15 years as traditional lighting gave way to LED.  

Everyone worked their own way, and the data took up most of the time. "We were spending 80% of the time in getting the data, transforming the data, reconciling the data, and talking to the stakeholders," said Sander van Schoonneveldt, supply chain network modeling specialist at Signify. That left little time for exploring scenarios. Stakeholders waited about three months for outputs. By then, leaders had often already made their choices, so studies confirmed decisions rather than shaping them.

Why Signify chose Optilogic  

When Signify's contract with its previous supply chain design software provider came up for renewal, the team had a choice: replace the tool, extend it, or look for something more. Faster project turnaround had already pulled the team into problems beyond traditional network design which required more capabilities than a standard modeling tool. The team was running its models on local desktop machines and wanted scalable computing in the cloud.

Optilogic gave the Signify team two things it needed:

• Answers to the traditional supply chain design questions: Footprint, greenfield analysis, sourcing, and the trade-offs between cost, cash, and service built on Optilogic’s Cosmic Frog application. ‍
• Room for custom analytics on the same platform: Scalable computing and parallel scenarios for problems that do not fit a standard design model, using Optilogic’s Atlas.  

"That flexibility turned into a much broader set of practical use cases," van Schoonneveldt said.

Less time on data, more time on decisions  

The first investment was a reusable foundation. The team gathered data on facilities, customers, sourcing relationships, and master data, along with the methods to clean them. Today the standard data, methods, models, and validated baselines now carry over from one study to the next.

That cut the time spent on data considerably, and it changed how the team works with stakeholders. “The iterations got much faster—from months to weeks or even days in some cases—and that helped the discussions with our stakeholders," van Schoonneveldt said.

Stakeholders do not always know exactly what to ask at the start. Now the team can show results early and refine the questions in the boardroom: running scenarios on the fly, moving between aggregated views and detail, from a regional or global scope.

One platform for answers across footprint, parcel, and packaging

The same team, data foundation, and platform answered two kinds of questions: redesigning the supply chain Signify needs and getting more out of the one it already runs. In North America, a Cosmic Frog redesign of the distribution footprint after a merger resulted in multi-million-dollar annual savings. Custom analytics on the same platform took on parcel selection and packaging, identifying millions of euros in potential savings. “With the compute power of Atlas in our hands, we could do this with our own created scripts,” van Schoonneveldt said.

What's next: continuous decision support

Signify's strategic studies and broader analytics continue. Three more steps are underway or planned:

• Continuous refresh. The team is developing continuous refresh with an always-on digital twin so its models and data stay current as assumptions change. ‍
• Selective self-service. With so many questions to answer, the team cannot be involved in every analysis. It is building an environment where regional organizations can run their own analyses on digital twin models to answer questions such as whether to keep delivering directly to a customer's outlets or send supply through a central DC.
• ‍AI-assisted workflows. Signify already uses AI for coding. Next it wants help with creating and evaluating scenarios and reconciling results, because more scenarios mean more output than anyone can read.  

According to van Schoonneveldt, "It's not only rerunning a network model. We want to learn every time and make our models better each time."