Accelerating innovation with virtual twins

PARIS — Staying relevant in today’s fast-moving marketplace is difficult for many companies, but product development could become easier using virtual twins.

Consumer packaged goods (CPG) companies wanting to reduce formulation time, test products and revise packaging are using predictive modeling to develop prototypes that might otherwise take years.

“If you set these models up correctly, they can prove out thousands of different opportunities within this portfolio window that I want to shoot for and boil down to half a dozen to take to a prototype,” said Ray Wodar, global director of business consulting for CPG and retail with Dassault Systèmes. “If you try to do that with a fifth engineering team, you’d have to staff up to where it wouldn’t be efficient.”

Wodar said Dassault Systèmes’ Virtual Twin as a Service (VTaaS) model places the company’s software in a cloud environment and deploys it immediately for client use.

“Instead of the typical 24-month project where you buy the software, enable the people, and deploy it on system, we can get time to value with a project in eight weeks to several months,” he said.

Defining the terms

Virtual twins are different than digital twins or AI-generated personas. Virtual twins use data, models and simulated variables to predict how a new product might be formulated or how plastic packaging might be made lighter.

Virtual twins can mirror reality and model what hasn’t yet been built, plus they can predict future behavior, test design alternatives, simulate manufacturing processes and explore operational scenarios completely in the virtual world.

In contrast, digital twins are real-time digital replicas of a physical object, process or system. They reflect the state of a product as it currently exists.

AI-generated personas are built from curated data and market research to test market reactions for a specific purpose. For example, a CPG company could build a generic Gen Z snack consumer for product development and marketing purposes.

Modeling bio-based packaging

Wodar said virtual twin technology may help to enhance biodiversity by modeling a shift away from single-use plastics to more bio-based materials.

“Certainly in food and beverage, those that have been plastic-based are moving to fiber and plant-based, (including) mycelium and wood pulp that’s been processed and can have the same performance as plastic but can be used in fill,” he said.

Dassault Systèmes has worked with Nestle and Bel Group on formulation technologies.

| Photo: Bel Brands USA

Since packaging is one of the major costs for CPG companies, Wodar said they’re all looking to “lightweight” existing packaging and reduce that expense.

“We’ve seen 5% to 15% documented savings in a typical lightweighting-type project,” he said. “Most recently, people are looking to totally replace most plastic and are moving to fiber-based overwrap.”

Putting better packaging on the shelf may lead to enhanced consumer acceptance, Wodar said. Consumers may consequently “feel better about buying a brand” and CPG companies might see an increase in market share.

Moving ideas faster to market

Dassault Systèmes has worked with Nestle and Bel Group (maker of Babybel and The Laughing Cow cheese brands, among others) on formulation technologies, Wodar said. The work has included how to shift to virtual insights “where they can use platforms and move more quickly to market,” he said.

Using virtual twins may help manufacturers take a product concept to research and development and from there to prototyping much faster, Wodar said.

“Companies are using prototypes for virtualizing packaging and using digital models instead of markups they may have put in front of a consumer panel in the past,” he said. “Those take months to create and (there’s) not much really good data you get from a consumer panel anyways.”

To further develop the approach, Dassault Systèmes has partnered its Biovia software tools with Foodpairing.com, a Belgium-based data and flavor intelligence platform that helps chefs and food manufacturers create formulations.

Wodar said the partnership connects with consumer preference sites that Foodpairing.com tracks “and basically do formulation design and lab workflow and scale that up.”

“Data science is really the new gold … when it comes to innovation,” he said. ”Data, and good data, is what drives predictable outcomes. If you don’t have good data, you’re not going to have a good model, and if you don’t have a good model, you’re not going to have good results.”

Wodar called using virtual tools transparently “crucial” for any brand.

“I don’t think people would be shocked to hear that people are investing in this technology because it’s not only good for the innovation side of things, but, at the end of the day, it’s making a safer product, too,” he said.

Some challenges lie ahead

While virtual twins can reduce time to market in several areas in the production cycle, Wodar said sensory models “are probably the most difficult to create.”

While progress has been made around stability models and virtual shelf-life testing, thereby reducing required time frames, there’s still work to be done in the sensory area.

“… I would agree that mouthfeel and taste are really hard to create in digital models,” he said. “We are getting down to a molecular model, so we are getting closer to these things.”

It’s also very difficult to get a full virtual twin in production at a factory level.

“Scientifically, we could probably do it, but it would be difficult to make the numbers work,” he said.

In the meantime, Wodar said CPG companies need to leverage virtual technologies to be able to gather insights, develop innovative products and move quickly.

“I think we get to a point where everyone’s portfolio is going to have to move as fast as social media does,” he said, adding, “It’s going to be imperative from a competitive standpoint to stay relevant in the marketplace.”