Canadian Manufacturing

Machine learning helping food manufacturers battle pandemic and look to the future

by CM Staff   

Manufacturing Operations Research & Development Sales & Marketing Technology / IIoT Food & Beverage AI artificial intelligence COVID-19 Food Manufacturing In Focus machine learning Manufacturing sales Technology

350 million potatoes were left in storage because of bars and restaurants being closed due to COVID-19.

PHOTO: Machine learning and AI helping food manufacturers/Mike MacKenzie via Flickr

McCain recently announced an investment in Fiddlehead Technology after six years of working together, and the ringing endorsement came in a year that’s required food manufacturers to be resilient to the challenges the pandemic has brought on.

The data science company uses advanced analytics to help manufacturers avoid investments and predict trends in food purchasing, helping them forecast sales and invest in their products and technologies with predictive modelling data sets.

The New Brunswick based company plans to hire new data scientists and engineers at their Moncton office, and already has a number of other food manufacturers signed up to use their predictive modelling solutions.

Shawn Carver, CEO of Fiddlehead Technology, was excited to talk about the news when he sat down with Canadian Manufacturing.


“Traditional tools that have been used to forecast sales were no longer applicable in 2020, because they relied too heavily on historical patterns and sales orders. Once the pandemic closed down bars and restaurants, there were no sales orders food manufacturers could use to forecast sales. Companies started contacting us to apply causal modelling that looked at point of sale transactions that were truer to commercial demands,” he said.

Fiddlehead’s machine learning tools pull data sets from a wide variety of sources to measure food trends and purchases, including things like foot traffic patterns, google searches and payment card purchase size in reaction to COVID-19 case spikes and government restrictions. The findings were eye opening not just for food manufacturers but consumers as well.

“Consumer behaviour has changed dramatically. Dayparts (segmenting the day into a block of time meant for certain product sales) have been affected. Morning breakfast sales are down significantly, people aren’t commuting anymore, or grabbing breakfast before they head into the office. Working from home, not going out has really impacted that business. Conversely, afternoon snack time is way up. We’re seeing a significant rise in the number of people who go out after lunch for a walk, a quick snack, or a late meal.”

Shawn Carver noted that it would be interesting to see how many of these trends will stick post-pandemic, and felt that it was too early to conclude if these were permanent shifts in consumer behaviour. In either case, Fiddlehead’s machine learning technology was helping food manufacturers stay resilient and adapt to changing consumer behaviour.

The data found that when government restrictions were announced, frozen food sales spiked, with people buying longer lasting items because of market uncertainty, enabling food manufacturers like McCain to invest in their agricultural crops and products that would benefit from these trends. If average cheque sizes at restaurants diminish and disappear, food manufacturers would now know that consumers are investing less in fresh meals and crops that don’t keep as long as opposed to longer lasting pantry staples (and the crops used in their manufacturing).

Fiddlehead’s data continues to model scenarios to prepare manufacturers to localized trends as they happen and there is no reason to believe it can only be limited to food manufacturing. Manufacturing products across industries can take advantage of predictive modelling to forecast sales and offer products at the right place and the right time, bouncing back from the pandemic that much faster.


Stories continue below