When I first saw ‘Terminator’ I pictured the evolution of AI as a fairytale idea. But now it’s everywhere around us. And consumer goods industries are not that far behind! Today we delve into how companies are using machine learning & data to enhance efficiencies across operations, supply chain and marketing in consumer goods industries.
Various reports state that by 2035, AI-driven processes will have the power to increase the productivity of employees in consumer goods organizations by 40% and boost profitability by 38%. 92% of senior manufacturing executives also believe that artificial intelligence and robotics will help improve their degrees of productivity. Do you feel left behind in this race? Here are some examples & inspiration for you from other leading brands on what they are doing. Hope this kickstarts some of your thinking on how you can implement it in your own organization:
First: Predict the Unpredictable
FMCG is an olden industry and surely one of the late adopters when it comes to AI. The maximum most of the enterprises have gone is to digitize their operations. But capturing the data in digital format is the easier part. The more difficult part is gaining insights from the data. Marico, the Indian FMCG company, is drifting away from just data capture based thinking to an age of getting into analytics. A shift from the descriptive to diagnostics, prescriptive and predictive. Marico uses AI to predict which of their distributors could be facing financial trouble and drop out of business. In such cases, they take corrective action of either helping in financing these distributors or looking for alternatives in that region.
Another brilliant example of a company proactively using data is PepsiCo. When the company launched Quaker Overnight Oats, it was able to identify 24 million households that it felt were its ideal target audience. Then, PepsiCo identified the shopping venues that these households generally visit and created promotions to appeal to that audience. For the first year after launch, this idea could drive 80 per cent of the product’s sales growth.
Takeaway: Start being proactive with data, rather than reactive!
Two: Manufacturing + Machine Learning = Miracles
Now talking about the same company reminds me of another major point- on how the Frito-Lay (a subsidiary of PepsiCo) manufacturing plant is benefiting from machine learning. One of their AI projects uses lasers to hit chips and then listen to the sounds coming off the chip to determine texture. Algorithms process the sound and determine the chip texture to automate the quality check for Frito-Lay’s chip processing systems. In addition, PepsiCo also developed a machine learning model to be able to predict the weight of potatoes being processed.
Another work in progress AI project for Frito lays is to assess the “percent peel” of a potato. By recognizing this data, the Frito-Lay team can optimize the potato peeling system. All three dart in the bullseye! This project alone is estimated to save the company more than $1 million a year just in the United States.
Takeaway:machine learning is a goldmine for manufacturing processes!
Three: Arm your field reps with superpowers!
A couple of years back — one of our India’s largest FMCG companies had reached out to us at Fieldproxy to help automate their field operations. Over the course of understanding the needs & helping them deploy the product, we realized the nascent stage in which AI was currently being used even in such large companies. For one — when the company meant automation — they primarily meant data collection. Since that is what they had been doing for years.
However, using Fieldproxy’s product we helped optimize various aspects of their field force. One of the key aspects was with respect to order recommendation. Prior to Fieldproxy’s implementation — a lot of the suggestions by the field agents with respect to what SKU to pitch to the store or the quantity of product to pitch would be left to their intelligence. Which may not always yield the desired output. However, using Fieldproxy’s order recommendation engine — which takes into account the historical purchase data, competitor behaviour, similar store purchase behaviour — to propose the sales order, the client witnessed a 12–15% same-store growth while also decreasing store visit times by 20%
The above are just a couple of ways in which we help our clients. Our product has helped various other clients in helping them in various aspects including merchandising, training, sales forecasting, route planning, operational efficiency & target setting.
Takeaway: The insights generated from data should flow right across the value chain down your last field rep.
Try this using Fieldproxy!
Four: If you know, you know!
The skincare world is oversaturated. Especially if you include all the affordable convenience store brands. If you’re a shopper with a budget, you are likely to mix different products. Blindly guessing which combinations will work is a herculean task — like a chemist without a periodic table. To solve this problem, Olay, a well-known drugstore brand turned to machine learning. They created an AI to analyze your skin from your selfies.
So that it can tell you which beauty products to purchase — from just a selfie! With the new selfie predictor about a third of the women now walk out of the store without having found that right product for her.
Takeaway: You need to ease the process of consumer decision making. And the suggestions you give each customer needs to feel unique to her
Five: There is so much more to data!
North Face uses IBM Watson’s cognitive computing technology to help consumers determine the best jacket for them — based on variables like location and gender preference. Depending on whether you hike in Iceland in October or commuting in Toronto in January — the kind of jacket you wear is going to be vastly different, right?
Post-implementation of the above solution, North Face noticed an increase of 75% in total sales conversions. And that is not the end of it. Once North Face was able to millions of such consumers enquiries through their above recommendation engine -they were able to glean important insights and patterns on what customers are requesting and feed that into other aspects including product development, demand planning & so much more.
Takeaway: Look at the data you have in a more comprehensive manner — rather than looking at it through a single lens