When is real-time analytics the right analytics?
(This blog was written by Akhtar Saeed, Suketu Gandhi and Vidisha Suman)
In today’s dynamic business environment, most firms are focusing on agility, responsiveness and customer centricity in order to remain competitive.
A leading industry survey suggests that 70 percent of executives regard data analytics as a core component to achieve these objectives. However, only 15 percent of businesses report deploying their data analytics project to production. This is primarily due to the difficulty in quantifying the ROI and to budget constraints.
How do you understand your organization’s true requirement for speed? Is real-time analytics an essential component of speed across your organization?
There are three key considerations for improving the speed of the organization to meet the speed required for business decisions:
Technology decisions need to match the speed of business:
The increasingly faster pace of business has increased the demand for real-time, automated decisions and actions. Business leaders today expect results in vastly shortened response times, to support faster (and more dynamic) business decisions.
For example, a typical online retailer has 40,000 price changes in an hour compared to 15,000 price changes per hour for a typical physical retailer. In the digitally connected world, the speed of business decisions needs to match the speed of information. The expectations from technology by business leaders, in turn, are also evolving. The emphasis is now on greater speed and flexibility.
Speed of business varies for functions within an organization
Customer-facing departments such as sales and marketing need real-time analytics and decisions to provide the best customer service and engagement. For example, retailers such as Walmart and Carrefour use geo-fencing/beacon technology and corresponding analytics to enhance in-store customer engagement through proximity based discounts and product information. These applications require real-time analytics and automated actions based on the data. Customer engagement using real time analytics is known to have increased as much as 400 percent.
Backend operations such as supply chain can benefit extremely by high-frequency batch processing of information. They do not require real-time analytics and the scale and agility of technology infrastructure associated with such systems.
Consider key analytics used by Inditex (Zara), the fashion apparel giant known for managing a supply chain that is 12 times faster than nearest competition despite having 10 times more unique products.
Key innovations at Zara using end of day sales information include:
- End-of-day data gathering to identify customer preferences in cloth, color and style based not only on merchandise purchased but also items that were tried but not bought
- Inventory optimization models to help the firm determine how many of which items in which sizes should be delivered to each specific store during twice-weekly shipments, ensuring that each store is stocked with just what it needs
In the wine and spirits industry, in more than 50 vineyards in California's Napa Valley, solar powered sensors are continually gathering data about the health of the vines giving wine makers insight that can lead to higher quality wine. Careful analytics on this data can help optimize the overall cost of manufacturing wine as well as improve the quality of wine.
Operational reporting requirements and real time analytic decisions are distinct
Real time decision making requires strong IT infrastructure support such as scalable platforms and in-memory databases that often results in implementation budgets spiraling out of control. While all analytics and reporting requirements rely on efficient data architecture and storage systems, organizations should monitor what areas truly require real-time or near real-time analytics and what areas can be batch processed at pre-defined frequencies.
Consider point-of-sale analytics: An example for point-of-sale real-time decision making is in fraud detection and prevention. Online retailers such as Amazon are known to use sophisticated analytics to prevent fraudulent returns and claims. Offline retailers are also turning to real time analytics to prevent fraud.
Hudson’s Bay Corp. in Canada broke up a $26 million fraud ring with one analytical application that fights returns fraud when a customer brings back a shoplifted item or reuses a sales receipt to return an item multiple times. In both cases, the analytics needs to be performed in real time to prevent the fraudulent transactions.
On the other hand, operational point-of-sale reporting such as collection of statistical data such as daily, weekly or monthly sales from each store need not be conducted in real time using expensive in-memory technology.
So, before you finalize your investment plans in analytics, understand and prioritize your organization’s need for speed.
(About the authors: Akhtar Saeed is a vice president at Southern Glazer's Wine & Spirits.Suketu Gandhi is a partner with global management consulting firm A.T. Kearney in the Digital Transformation Practice, and can be reached at Suketu.Gandhi@atkearney.com. Vidisha Suman is a manager in A.T. Kearney’s Digital Transformation Practice.)