Kubernetes helps orchestrate and automate tasks associated with containers - an essential need as you scale. Here's a deep dive for IT leaders on what Kubernetes can do, key terms, best practices, trends for 2020, and more.
4 focal points for Monsanto digital transformation
Jim Swanson discusses how to pave the way for the benefits of near real-time data
Editor's note: The following interview was conducted prior to Bayer’s acquisition of Monsanto. James Swanson is now Senior Vice President / CIO and Head of Digital Transformation at Bayer Crop Science.
Over the past few years, the IT organization at Monsanto has undergone a tremendous transformation in which we redefined our structure and the role we wanted to play in the business. This transformation was key to enabling a digital ecosystem, which requires collaboration, automation, data, and data science.
Prior to our IT organization’s transformation, for example, it took the company five or six years to grow seven petabytes of data. In the year following our transformation, we grew by two petabytes alone.
[ See our related story Crucial CIO skills for digital transformation success. ]
To make sense of this high volume of data, we launched science@scale, our internal data science platform. Science@scale is helping us accelerate the development of analytics-driven decision models by turning data into actionable insights across the company.
Science@scale, along with automation and collaboration, is helping us develop our digital ecosystem. On the technology side, we’re focused on four layers:
- Modernization: This includes our move to the cloud, our move to high-end capacity networks so we can move around large amounts of data, and a workforce platform that’s scalable for employees.
- Data: This is about decoupling data from applications, creating data assets that can be consumed by the company through APIs, and democratizing that data for multiple use cases.
- Platforms: We have six major platforms that enable our business, so we want to build capabilities off those platforms.
- Decision science: We’re embedding models in every decision we make, as we have a scalable foundation and data that are rich and can be consumed.
These four technology-focused layers work together to enable the business and our customers with the data and insights necessary to make smart decisions. For example, there’s an average of 40 decisions growers need to make every season, from what to plant and how much to water to what fertilizer is best and which pesticides fight the right diseases.
We want to positively impact every decision through a model because if we can reduce the input cost — using less fertilizer or a prescription seeding rate that gives a better yield — we can increase the output. We apply this concept internally as well; right now, we’re mapping and tracking more than 100 business decisions, and are using models to inform those decisions.
One internal example of how this works is in the amount of safety stock we produce. We grow our products in our supply chain, so we can’t just turn a factory on and off — we need to forecast what will be the best-growing hybrid in order to produce enough of it.
On average, we’ll produce a 5 percent safety stock. But, if we apply models through data that say we can reduce our safety stock from 5 percent to 2 percent and still meet the demand, that’s a huge 3-percent savings in resources that we can apply to other products we’re trying to grow.
While these models are important, it’s also important to consider the economics of applying them to your decisions. If it takes days to run a model and four months to pull the data together, you’re too late — your information will always be in the rearview mirror.
Instead, you need to move to more predictive and prescriptive models that help you decide where you need to go. That’s what science@scale does for us; now, we can run these models in almost-real-time, and we can ingest a large amount of data, which provides the inputs we need to make informed decisions.
Near real-time data benefits
This near-real-time data has benefited us in more ways than one. For example, every seed we plant has a DNA profile. Our goal is to match the DNA profile to a certain phenotype — things like a good root structure, a broad leaf that can absorb more water, or its resistance to drought.
Before, we used to look at the data, then plant every seed to test whether the phenotype is what we wanted. Now, we’re doing more than a billion simulations on the computer before we put a single seed in the ground.
This has resulted in a huge reduction in cost of labor and has had a tremendous impact, because we’ve moved from regression-based models to machine-based learning models that self-adapt as we enter in more data.
These are just a few examples that showcase how we’ve applied data to our business decisions — there are many others that have driven equally substantial outcomes, too. Throughout this transformation, we’ve set our sights on transforming agriculture — one of the oldest industries on the planet — through our use of the latest, cutting-edge technologies and decision science. Seeing the impact we’re having now, and the potential that lies in the future, has been truly exciting.
[ Read our new report from HBR Analytic Services: Transformation Masters: The New Rules of CIO Leadership. ]