Today, organizations face a variety of threats to their business models due to the tremendous pace of technical innovation. New advances are coming at a rapid speed, paving the way for start-ups to emerge and challenge established companies across industries. For example, it’s highly unlikely that executives at FedEx and UPS are not talking about SideCar, or that those in top budget or mid-range hotel chains are not taking note of AirBnb. Trends such as the Internet of Things, big data and ubiquitous mobility are increasingly blurring the boundaries between physical and virtual reality, and forcing organizations to become digital enterprises in order to survive.
To do so, companies must create an architecture that helps them become nimble and resilient in the onslaught of big changes. They need to define a systematic approach to proactively identify disruptive technologies, evaluating their capabilities, testing out specific use cases and then deploying them widely. One of the critical pillars of such an approach is embedded analytics.
Given that enterprises are drowning in data from numerous sources, analytics must move beyond dashboards and predictive models, and bleed into business processes and applications. Many operational decisions can be automated by analytics on a much larger scale than what is being done today, which will lead to a more intelligent and responsive organization. Even some of the “non-routine” and rigid business processes, such as pricing and promotions, can become more flexible when injected with analytics.
For example, rather than a fixed price, a series of test and learn experiments can help determine optimal pricing, and this can be used to win back market share or react nimbly to competition, without sacrificing profitability. Sometimes this can even lead to discovery of new business models. In order todo this, the traditional barriers between operational processes and analytics must be broken, and analytic tools and systems need to become closed loop.
It’s also critical that business applications become more analytically driven. For instance, configuring acustomer facing application to learn from user behavioral patterns so that the tool then provides a better overall individual user experience, based on clicks, searches and usage of application features.
I witnessed an analytics-driven transformation recently with a traditional service provider that providesappliance installation, maintenance and repair services to a distributed customer base. The original business model required multiple visits by technicians—for example, one to diagnose the issue and atleast a subsequent session to resolve it—resulting in numerous inefficiencies and increased time to resolution. To address this, the company invested in a new solution driven by advanced analytics, textmining and supply chain optimization algorithms, operating on real-time customer feedback and individual appliance details. The supply chain teams can now ensure that the right parts are available to the technician on the first visit, eliminating the need for multiple trips and enabling the company to deliver better service.
This is just one example to illustrate how embedded analytics can engender a more nimble organization. Analytics is a complex endeavor, however, since it requires combining big data, algorithms, processing and presentation in a single place. In this environment, a single department can no longer be responsible for analytics and insight. Companies must collaborate with teams acrossbusiness owners, IT process owners, data owners, and sometimes even suppliers.
Moreover, there cannot be a big-bang upfront design for these types of intelligent platforms. They need to be developed through a process of experimentation, discarding what does not work and pushing ahead with what does. These iterative, collaborative and evolving processes are also the principles of agile methodologies, which has led to significant improvements in innovation and productivity over the last decade.
Authored by Ramesh Hariharan, Head of IdeaLabs