Think Big Be Smart Scale Fast
1. Most Organizations have high hopes for using Big Data Analytics in their Supply Chain but many have had challenges in deploying it. What are your thoughts on this?
Big data analytics—or data analytics in general—is driven by the envisaged outcomes. The clearer the outcomes are, the more likely a big data analytics initiative will be successful. Lack of clarity is the core challenge faced by many organizations. Organizations need to consider the following key points while planning a big data analytics program:
a. Why is it needed and how can the benefits be derived? Are the benefits quantitative or qualitative?
b. Is data available across the sources and if yes, what is the quality of this data?
c. Data democratization should be the larger goal to ensure that the right data is made available to the right audience so that the adoption rate is higher resulting in the success of this program.
With reasonable direction, organizations can start collecting data that’s relevant, and then build out initial models and subsequently refine them. A key point to keep in mind is that big data analytics is an ongoing journey, not a single stop.
We analyzed demand disaggregation at various levels of hierarchy in order to identify best level of hierarchy for demand disaggregation
The top three use case for big data in supply chain are traceability, relationship management, and forecasting. A good example of the right usage of big data analytics is the work Mindtree did on demand forecasting for a global consumer goods manufacturer. The need was to identify and automate the process of APO demand disaggregation with control checks and defined logics at various cuts of categories and customers or clusters. Our solution considered identifying data challenges involved in demand disaggregation. We analyzed demand disaggregation at various levels of hierarchy in order to identify the best level of hierarchy for demand disaggregation. The benefits included better Mean Absolute Percent Errors, meaning more accurate forecasting compared to the ‘As-Is’ method in the Quarterly growth plan Cycle 1 (120 days) and better value compliance in the Quarterly growth plan Cycle 2 (90 days). We automated the entire forecasting technique with lesser rework and more accurate results, reducing the time required in disaggregation and review.
2. With your Rich Experience of managing IT Organization and Steering Technology for your Enterprise, can you please share some of the unique lessons learned and your advice for fellow tech decision makers?
Mindtree has been servicing customers on IT services for the past 18 years and has been part of many exciting journeys. We are a company that was born when the “Digital Era” was initiated and have engaged with global customers to help them define their digital transformation programs and change the way their businesses run. Given this rich experience, some of our key learnings have been:
• Governance is key: The new trends in technology like IoT, Cloud, AI/ML, Big data analytics not only mean a big change from a technology implementation standpoint but also from an organizational change management perspective. Hence, the planning for such an implementation needs strong governance to come into play.
• Start with understanding what’s already in play: In a lot of instances, the existing systems and processes are not technically ready to integrate seamlessly or provide information in the as-is state which clearly implies that sufficient time and energy needs to be spent in understanding the existing technology landscape and the implications of the future roadmap.
• Think big, Be smart, Scale fast: This has worked best for most of our customers when they are in the path of adopting new technologies. For example, for data analytics, while the need for data analytics and big data has been well-articulated, the key to the success is user adoption and effective usage. Considering this, what has worked well is identifying a key pilot or BU or department where the right data is available, and the ROI is well defined. This is then used as a use case to spread the platform across the organization.
An example of the learnings depicted above is our work with a leading consumer goods manufacturing company that wanted to build an assortment planning mechanism for its salesmen to automate the outlet sales execution that allows more lines per call to be achieved. We engaged with the customer to ideate and build “Decision Support Group” (DSG) to serve as the analytics backbone. To that we added “Advanced Predictive Analytics” for generating cross-sell and out-of-stock recommendations at an outlet level. This was rolled out as a pilot to limited markets and then scaled across 25 plus countries. DSG currently publishes 20M plus width pack recommendations that translates to approximately $40 Million additional sales revenue per year.