Inside the Greek yogurt wars: Dannon taps predictive analytics

How one company was able to quickly capitalize on consumers' changing tastes using data analytics.

Yogurt manufacturers are straining for more market share -- literally. Thick, strained Greek yogurt, a single-digit portion of the U.S. yogurt market in 2007, now accounts for 28% of today's $7 billion in U.S. yogurt sales. And sales are growing, while demand for non-Greek yogurt is declining.

But while slick ad campaigns and novelty products like Greek yogurt smoothies may pique the interest of consumers, not all yogurt manufacturers are banking on advertising savvy alone to help them capture more market share. For Dannon, a subsidiary of France's Groupe Danone, science is the secret weapon in the yogurt wars.

That wasn't always the case for the White Plains, N.Y.-based dairy behemoth. For years, Dannon's sales force relied "on a very manual, Excel spreadsheet-driven process" to forecast sales and manage inventory, says Dannon CIO Timothy Weaver.

"It was a very cumbersome process and not particularly accurate," he says. What's worse, drawn-out forecasting was leaving Dannon's sales reps with less time for the actual execution of sales plans and customer engagement. And attempts at forecasting that were only about 70% accurate were prompting production teams to side-eye sales reps' reports.

That is, until Dannon turned to IBM's predictive analytics tools, including the vendor's Strategic Trade Planning and Customer Trade Planning systems, in April 2010. Dannon won't disclose hardware and system costs, other than to note that its technology investment "returns significant business value."

Forget about gut feelings and guesswork. Nowadays, Dannon sales reps first design a promotion, such as a 49-cent reduction on the cost of a four-pack of Oikos fat-free Greek yogurt. Then they enter factors such as pricing conditions, target customers, length of the promotion and product SKUs into a computational model.

Using sophisticated algorithms, the model crunches the figures, along with historical, regional and market data, and spits out "the expected base sales you will always see from that pack group, the incremental sales you can expect from that particular promotion, and its financial implications -- such as gross revenue increase and profitability," Weaver says. Dannon became a user of IBM's tools through a series of acquisitions: Its original predictive analytics software was from M-Factor, which was acquired by DemandTec, which in turn was acquired by IBM.

Since rolling out IBM's predictive analytics system in May 2011, Dannon has increased its accuracy in forecasting and inventory planning from about 70% to 98%. But predictive analytics is about more than simply arming sales representatives with a crystal ball. While data analysis has helped Dannon retain market share in the hypercompetitive yogurt industry, the company has found that success comes at a price. Like many data crunchers, Dannon is discovering that for predictive analytics to pay off, it's necessary to devote some time and resources to making the system work. Careful testing of models, senior-level buy-in, occasional IT hand-holding and a refreshing of talent are key requirements.

A Dog-Eat-Dog Market

In order to fully appreciate the power of predictive analytics, it's important to understand just how cutthroat the Greek yogurt industry really is. Capacity constraints, retailer squabbles, poor pricing strategies, misguided promotions and missed sales quotas are serious stumbling blocks that can lead to lost market share for Dannon and competitors like Chobani, General Mills and Liberte.

1 2 3 4 Page 1
Page 1 of 4
Bing’s AI chatbot came to work for me. I had to fire it.
Shop Tech Products at Amazon