Predicting Weather, Influencing Behavior

Last week I had the privilege of participating in two IBM events, each focused on opposite sides of the food supply chain.

The first event was a dinner discussion held in NY that focused on what IBM calls “precision agriculture.”

Precision agriculture leverages a new IBM weather-modeling service, called Deep Thunder, that provides local, high-resolution weather predictions customized to business applications with the goal of providing weather forecasts at a level of precision and fast enough to optimize specific processes and operations such as resource allocation, scheduling and routing, and others that may be constrained by weather events.

Sensors placed throughout the fields are used to measure temperature and humidity of the soil and surrounding air. In addition, pictures of fields are taken using satellite imagery and robotic drones. The images over time show crop maturity and when coupled with predictive weather modeling showing pinpoint conditions 48 hours in advance, IBM Research is able to build models and simulations that can predict future conditions and help farmers make proactive decisions.

So for example, a farmer knows he is going to harvest a crop of tomatoes in two days, he can allocate resources for picking, arrange transportation, and alert potential buyers, so that as soon as the tomatoes are harvested, they move seamlessly through the supply chain.

Other aspects of the producer’s business are also affected by weather, such as the decisions made around fertilizing and maintaining the crops. If farmers know they’ll have heavy rain the next day, they may decide not to put down fertilizer since it would get washed away.

Weather not only affects how crops grow, but also logistics around harvesting and transportation. When harvesting sugar cane, for example, the soil needs to be dry enough to support the weight of the harvesting equipment. If it’s humid and the soil is wet, the equipment can destroy the crop. By understanding what the weather will be over several days and what fields will be affected, better decisions can be made in advance about which fields workers should be deployed to.

Once the food has been harvested the logistics of harvesting and transporting food to the distribution centers is crucial. A lot of food waste happens during distribution, so it’s important to transport the food at the right temperature and not hold it for longer than needed.

The other IBM event, a Webcast which I moderated, focused on influencing certain types of events – shopper behavior at the shelf, to be specific. It was basically an extension of a Progressive Grocer roundtable IBM sponsored in September, and included as panelists three roundtable attendees: Karl Meinhardt, VP social and digital marketing, Albertsons LLC; Rick Brindle, VP of industry development, Mondelez; and Steve Brown, general manager, IBM Global Consumer Products Industry.

The webcast discussion focused on two key areas of shopper marketing: the challenges of retailer-supplier collaboration around shopper marketing programs, and mobile shopper marketing programs. This was followed by a brief presentation by IBM’s Brown in which he outlined some shopper marketing tools and how they can be leveraged to help sort through all the data gathered by retailers. The webcast concluded with a 15 minute live Q&A session during which Meinhardt and Brindle fielded questions from webcast attendees.

Those of you interested in learning a bit more about Deep Thunder can visit the Deep Thunder home page.

If you’re interested in viewing the archived webcast, visit the Progressive Grocer website.