Getting The Best From BI
By Matt Pillar, chief editor
March 2016 Innovative Retail Technologies
Modern business intelligence apps are extending a glut of data from server rooms to the sales floor. How should merchants manage the mega data that’s now at their fingertips?
Business owners, C-suite executives, and line-of-business and store managers have all fallen in love with their dashboards, which have become part-and-parcel of just about every retail software application on the market. Retail pros have so much access to data that it’s easy to get caught up in “analysis paralysis.” We sat down with execs from two BI- and analytics-intensive retail software providers to gain insight into best practices for BI in retail.
Analytics: What’s Imperative, What’s Extraneous?
Every tool out there seems to have a dashboard today, and that’s not lost on Derek Rodner, VP of product strategy and software development at Agilence. He says determining which of those dashboards are imperative is a role-specific decision. “Having a store manager review a dashboard that highlights promotion trends is a good idea, if that store manager has nothing left to do with their day,” explains Rodner. “The store manager’s time is better spent on the floor, working with employees, ensuring the shelves are stocked and customers are happy.” Thus, he says promotions dashboards are probably better for those who are responsible for marketing. “Likewise, a dashboard on supply chain and logistics should be left for those who are responsible for it. When it comes to the tools available, it’s important to make sure the relevant information makes it into the hands of the people who can use the data in a timely manner. Nothing is worse than too much data or data that is too late,” says Rodner.
The Danger In App-Specific Data Disparity
Siloed data sources are another factor in the dashboard delivery decision. Empowering personnel at many levels within the organization with access to information they never had before dashboards can be powerful, but it also introduces risk, says Net Payne, CMO at March Networks. “Each group within the organization might have its own application-specific dashboard that draws from the same or related data, but they all might draw different conclusions from that which is presented to them,” he says, pointing to an example of the danger in application-specific BI. “Not long ago, companies were on their first or second ERP (enterprise resource planning) deployments. ERP began as a financial control or supply chain application serving specific functional groups. Since then, it’s all grown up, added more functional capabilities, and now it’s being sold to virtually everyone,” he explains. At issue, he says, is that ERP solutions haven’t solved the issue that every group in the company is looking at the data in a different way, and for different things. “We’re not yet to a place where that’s cohesive. You’ll have the same data sources, or slightly different data sources, with a similar look and feel to reporting, but different groups will draw different conclusions from that data.”
"It’s important to make sure the relevant information makes it into the hands of the people who can use the data in a timely manner. Nothing is worse than too much data or data that is too late."
Derek Rodner. VP of product strategy and software development, Agilence
Overcoming that risk often starts well before applications are deployed, says Payne. “Retailers are increasingly making purchase decisions in buying committees,” he says. “We’re seeing loss prevention, retail operations, and marketing coming together from the outset to make application decisions.” Often, he says these committees discover that they can help each other achieve their business intelligence objectives by sharing resources that already exist.
Asked about the battle for BI mindshare being waged between app-specific dashboard and reporting packages and enterprise BI providers, Rodner ponders an age-old question: Is it better to have an all-in-one business analytics solution that is OK or to choose the best-of-breed solution for each line of business? He says the question is still a valid one to ask. “The answer depends on the customer. Some are OK with a solution that may not be the best in any one area because it is good enough in all areas. Part of their rationale is that if there is an issue, there’s a ‘single throat to choke,’” he says. “However, many companies subscribe to the idea that the burden of managing multiple applications is worth it because they get the best solution in each functional area and thus a competitive advantage.”
Like Payne, however, Rodner cautions that it’s important to recognize that there may be further inefficiencies in having multiple platforms, each with a single area of focus. “In those cases, the same data may have to exist in multiple spots, leading to a bloat in storage. Worse, if the data changes, are you sure that all of the platforms have a single version of the truth?” Rodner also warns that if the application is too specific, its dashboards and analytics may not off er the actual best answer because they fail to take external factors and other data points into account, which an all-in-one solution may be able to provide.
Extending BI/Analytics To The Enterprise
Why are business intelligence solutions suddenly so accessible to so many retailers? Because the technology they’re built on has matured to the point of commodity. “In the past, inhibitors to enterprise-wide BI/ analytics were mostly technical in nature,” says Rodner. “Databases and hard drives were too slow to process the data. But, with the introduction of low-cost solid-state drives and the emergence of new database technologies, these types of solutions are now becoming more viable,” he explains. Rodner says BI tools should do three specific things — get the right information to the right people at the right time — and that the last requirement has been lacking in traditional tools. “Getting the right information to the right people has always been possible, and new UI (user interface) tools have made that even easier,” he says, “but the timeliness of the information has always been a challenge. With better and faster hardware and software, this is now becoming a reality.”
"From our perspective, the single best indicator of success is how many customers performed a proof or concept or pilot, and performed it with full breadth of intended users."
Net Payne CMO, March Networks
Payne agrees. “Sheer processing power is key. Whether video or transactional, the ability to affordably process data, thanks in part to the cloud, is far more attainable than it used to be.” He says the processing power of video systems, for instance, has dwarfed what it used to be, and that merchants can harness that power at a comparable price point to yesteryear’s.
Mobility has also driven the value of BI to the far reaches of the enterprise. “The ability to easily centralize intelligence and extend it to virtually anywhere has fundamentally changed how it’s used. It’s much easier to inspect, analyze, audit, and mystery shop without a physical presence,” says Payne.
Ensuring Adoption Where It Matters Most
Whether you’re leveraging app-specific intelligence or running an enterprise-wide BI solution, the reports you’re generating are only as good as the hands they’re delivered to. How can merchants ensure their BI investment is being applied at the ground level? Rodner says adoption begins with culture. “Having a corporate initiative and buy-in at the highest level is key to ensuring that these tools gain necessary adoption and are a success,” he says. “Once that has been established, the tools themselves must be able to cater to the needs and experience level of the end user.” Some advanced users, says Rodner, prefer mining the data themselves, and for those power users, the tool should be accommodating. Others just want the system to spit out answers and provide actionable events. “If the tool can’t cater to both, then it will ultimately fail. It all comes back to the three goals: the right information, at the right time, to the right user.”
Payne says the application interface is key to accommodating both types of analytics users. “The user interface should be customizable to different groups within the company. You have to have a simple version and a version that allows expert users to go deep.” To ensure users are engaged, he strongly suggests hands-on usage in advance of deployment. “From our perspective, the single best indicator of success is how many customers performed a proof of concept or pilot, and performed it with full breadth of intended users,” he says. “I would venture a guess that among those who execute a proof of concept, some 90 percent will move forward with a full deployment. Among those who don’t, we see twice as many issues with expected functionality and adoption.”