Building Toward Materials by Design in R&D

This content was originally discussed in the webinar, A Technical Framework for Materials by Design in Enterprise R&D.

Many materials R&D organizations are still running a highly sophisticated search engine powered by people. This process can be very effective because it draws on deep scientific expertise. But it is also slow, distributed, and hard to scale. Data is fragmented, key decisions are spread across functions, and some of the most important knowledge exists only in the heads of experienced scientists.

That is why Materials by Design matters.

Materials by Design is about changing the operating model of R&D itself, moving from a human-centered, trial-and-error search process toward a more target-driven design process.

Traditional R&D Is a Search Problem

In many organizations, materials development still works as an iterative search through a large formulation space. Teams make informed guesses, test them, learn from the results, and gradually narrow the field.

The challenge is that this search process is distributed across the entire organization. The result is a system that works, but one that depends heavily on human coordination, human memory, and repeated manual iteration.

Materials by Design Changes the Question

Materials by Design flips that logic because instead of starting with a candidate formulation and asking, “What properties will this produce?”, the process starts with the desired outcomes and asks, “What combination of inputs is most likely to get us there?”

That distinction changes experimentation from a broad search activity into a more directed design activity. It changes the role of data from passive record keeping into active guidance. And it changes the role of computation from a support tool into a core part of how the R&D system operates.

The Three Foundations of Materials by Design

For organizations moving toward Materials by Design, the foundation is not just better tooling. It is a clearer way of structuring the design problem itself. In the framework discussed by Enthought, that starts with three essential steps: specify the design space, shape the design space, and navigate the design space.

1. Specify the Design Space

The first step is defining the space you are actually working in. That means identifying the inputs you can control such as ingredients, formulations, ratios, process parameters, and other experimental variables as well as the outputs that matter. But the outputs are not just raw properties in isolation. The more important question is what “good” looks like for those properties. 
This is one of the most important mindset shifts in Materials by Design. The goal is to define the problem in a way that supports target-driven design.

2. Shape the Design Space

Once the design space is defined, the next step is making it more useful. A raw design space will contain many possibilities that are not actually viable. In a traditional workflow, those constraints often show up later, as different functions evaluate a candidate after it has already moved through part of the process.

Shaping the design space means bringing more of that knowledge forward. It means incorporating what the organization already knows about feasibility, manufacturability, safety, and other constraints directly into the design problem itself.

That matters for two reasons. It shrinks the search space, and it increases the density of viable solutions within it. In practical terms, that means teams spend less time exploring dead ends and more time working in regions that are actually worth pursuing.

3. Navigate the Design Space

Once the space is defined and shaped, this is where data, predictive models, and computational workflows become especially valuable. Navigation is the process of finding promising candidates without relying entirely on repeated physical trial-and-error. The better the space has been specified and shaped, the more efficient navigation becomes.

This is also where Materials by Design begins to create real leverage. Rather than treating every iteration as a fresh search, the organization can use computation to guide exploration more intelligently, evaluate more options, and focus experimental effort where it matters most.

Taken together, these three steps shift R&D from a process of repeated guess-and-check toward a more structured and target-driven design approach.