This week’s PMA virtual town hall focused on the rapidly growing field of predictive analytics, which helps companies draw from past data trends to make informed decisions about the future. Led by Doug Bohr, Executive Director of PMA’s Center for Growing Talent, panelists Diane Wetherington, CEO of iFood Decision Sciences, Garland Perkins, Senior Manager of Insights and Innovation with the Oppenheimer Group, Kathleen DeBrota, Senior Manager of Research and Development with AeroFarms and Gina Jones, Vice President of Commercialization, Insights and Analytics at PMA, spoke on how predictive analytics can support decision making, risk management and drive business.
Top left to right: Garland Perkins, Diane Wetherington. Bottom left to right: Gina Jones, Kathleen DeBrota.
Uses for predictive analytics
Predictive analytics can be useful in a variety of avenues. Oppy uses predictive analysis in three main ways, Garland Perkins shares. “We look at category trends to understand what is happening within a category, what the volumes and pricing are like, and look at the ‘why’ behind the numbers. We also collect market intel, such as import and export trends and how they influence the domestic production, which is impactful from a retail perspective. Finally, we also look at potential expansion opportunities – we’ll see if we need to adjust the volumes of an item already in production, or if we want to introduce a new product into the market,” Perkins explains.
Predictive analysis can be particularly useful when companies are looking to introduce new products into the market. Perkins shares: “A few years ago we were considering introducing Brussels sprouts into our portfolio. The data helped us understand that the overall consumption has gone up and that it’s no longer limited to specific holidays. We were also able to find a time of year where there wasn’t much volume in the markets and with all of this information, we were able to introduce them at a very opportunistic time.”
For AeroFarms, the use of predictive analysis is slightly different. “We look at the continuity and predictability for our yields and quality, and how we can leverage the data to get the best outcomes possible. It’s partially about optimizing the plants themselves – how much nutrition is in a plant, the flavor balance, the organoleptic qualities, and more. Leveraging the data has allowed us to be more agile as an organization. One example of this is that we were able to use the data to go from a 16-day growth cycle to a 14-day growth cycle. Through this, we’ve been able have growth cycles throughout the year and produce more efficiently,” Kathleen DeBrota shares.
Finally, predictive analytics can be utilized by prioritizing risk, particularly in the area of food safety. Diane Wetherington shares: “Companies collect a lot of data about food safety throughout the day already, to comply with the standard guidelines. This data can be used to give them additional guidance, too. Companies often have multiple different risk areas, and they can use the data to see what the probability is that they’ll have a food safety incident.”
How to get started
The panelists shared a few tips for companies who want to incorporate predictive analysis into their business. They all agreed that ensuring you have people on your team that are skilled in data analysis is key. “Whether it’s a single person or a whole team, you need someone who can accurately interpret the data,” Perkins says, adding: “It’s also important that you know what questions you want to find answers to, because analyzing the data becomes a lot simpler if you know what you are looking for. Finally, understanding what is compelling to you audience is also important. What is important for the retailer, or the grower? Knowing this will guide your efforts to get to the information that people will care about, and you’ll be more effective and efficient.”
Wetherington warns of a common issue with predictive analytics that should be avoided: “It’s important not to go beyond what the data tells you and to understand the limitations of your data set. Don’t make assumptions that you can’t actually back up with the data. For example, the data can be used to assess the probability that a consumer will buy a specific product based on their previous purchases, but only to an extent. If a consumer bought a specific produce item, we can make predictions on other produce items they might buy, but we can’t make predictions on the probability that consumer might purchase deli meat, for example,” she explains.
PMA Insight Solution Platform
Recognizing the importance of leveraging data in the industry, PMA is developing a new data analytics tool: The Insight Solution Platform. Gina Jones explains: “It’s a platform to help members get started on predictive analysis. To help them get the information to get started, as well as the data they need. It is designed to close the intelligence gaps in the industry by providing a single environment for users to work with point-of-sales data, shopper sentiment information, as well as giving users the opportunity to add their own points of view to benchmark their position in the marketplace and ultimately increase their profitability.” The new Insight Solution Platform will be available in February 2021.
Next week’s town hall takes place the day after election day and will explore the outcome of the US presidential race, and how it will impact the industry, businesses, and the global supply chain.