Using a dashboard to report on business status offers stakeholders a quick, high-level view—but a targeted causative analysis can yield valuable information that may enable bottom-line improvements.
More than any other entry point into big data and analytics, dashboards have delivered value to businesses because they display bottom line results on operations status, sales outcomes, and a host of other company events in an eyeshot to busy managers and executives.
Dashboards also offer drill-down capabilities into the data, so you can move to greater levels of detail if necessary.
But having dashboards and drill-downs into data doesn't always get a business where it wants to go—and companies are beginning to find this out.
SEE: Digital transformation in retail: How consumers are using tech to make shopping easierand more enjoyable (Tech Pro Research)
Several years ago, I learned about an online retailer in Britain that probed its analytics for clues to human behavior and found that the wives of soccer fans purchased more products online when their husbands were away at soccer games. This behavioral discovery improved sales. It was also an insight that never would have been found just by drilling down below a dashboard summary.
Probing data into the essence of behaviors and events is still a nascent discipline in most corporate analytics. As a consequence, managers get summaries and immediate drill-down data, but they don't get breakthrough insights that can help them reshape their businesses.
Here is an example use case from the retail grocery industry that illustrates the point.
Examining causative factors
Retail grocers routinely lose 30% of produce due to waste. In fact, the losses are so expected that grocery CFOs automatically bake 30% of produce losses into their annual budget numbers.
Internet of Things sensors and supply chain tracking attempt to reduce this waste, and to a degree, they succeed—but like the soccer example, unless you know the causative agents behind the problems you're trying to solve, you're going miss the solution.
Here's how most IoT sensors work when it comes to protecting food freshness.
They measure the humidly and temperature in enclosed environments like warehouses or truck trailers to ensure that the produce being shipped stays within specific cooling and humidity parameters. They also track produce from farm to table, which gives you a gauge as to how long the produce has been in transit or storage. Sensors can even alert you if a seal on a truck or a container has been broken.
But is this enough?
"It isn't, said Kevin Payne , VP of marketing at Zest Labs, an AgTech company that specializes in post-harvest shelf life and freshness management solutions.
The company provides supply chain software and sensors that track produce and that use proprietary algorithms known as ZIPR Codes to continuously recalculate the freshness of each individual produce pallet as it moves through the supply chain.
"By using this algorithm and providing freshness readings, the system enables growers, distributors, and retailers to immediately see which pallets are most perishable and need to be sourced to markets faster," Payne said. "They can also proactively make decisions as to where to ship these products, shipping pallets with higher freshness scores to more distant locations and sourcing products with lower freshness scores to closer retail locations."
SEE: 60 ways to get the most value from your big data initiatives (free TechRepublic PDF)
Payne said his firm performed an analysis of data collected on California strawberries harvested during the warm summer months of August and September, and that the analysis revealed that pallets experienced different cut-to-cool times. Some were also exposed to high temperatures for long periods of time before reaching the packing house.
Despite the fact that some of the strawberries in this batch had been exposed to higher temperatures that speed spoilage, the producer labeled all the pallets that were being shipped with the same "best by" date. "Fortunately, the sensors and the algorithms calculated actual freshness per pallet, enabling retail grocers to speed the berries most likely to spoil to market first," Payne said. "What we saw was that the analytics and the sensors were able to improve delivery freshness and reduce shrinkage by 50% or more."
The Zest Labs use case is just one example of how companies are beginning to examine behavioral and other causative factors to arrive at business savvy results that deliver bottom-line improvements.
What practices can CIOs and analytics leaders learn from this?
1. Focus on causative factors that might not be immediately obvious
The online British retailer found that wives took advantage of their husbands' being away at soccer games and went shopping. The retailer launched some of its most aggressive promotions during these times. In the Zest Labs use case, grocers took advantage of "smart" sensors and software with algorithmic abilities to really tell them the age of produce at the pallet level so they could reduce waste and improve profit margins. Neither use case was immediately obvious. Instead, analysts probed into business problems to identify causative factors for non-performance so they could improve results.
2. Test your theories
In the British retail case, sales grew with online promotions during soccer games. In the Zest Labs example, grocers made noticeable advances in reducing produce spoilage and improved profit margins. However, not all causative factor identification works. If yours doesn't, abandon it and look for something new.
3. Once you're successful with causative analysis, expand the practice to other areas of the business
Seeking out causes of non-performance instead of just reporting on status is a step up from dashboard reporting and drill-downs. Causative analysis is also a discipline that is not widely practiced in many companies. Once you have a successful causative analytics model that is generating results for the business, demonstrate it—especially to the CEO and other C-level executives. When the results are there, these decision makers and influencers are more likely to supportive expanding causative analytics.
- Business analytics: The essentials of data-driven decision-making (ZDNet)
- Why dashboard design is critical to analytics success (TechRepublic)
- Microsoft Power BI expansion aims to help analysts leverage business data more easily (TechRepublic)
- The 3 most important types of data for your business (TechRepublic)
Does your organization rely mainly on dashboards and drill-downs or have you adopted a causative analytics model? Share your experiences and advice with fellow TechRepublic members.