Revisiting enterprise Big Data Analytics trajectory
As Daniel Kahneman aptly puts in his book ‘Thinking Fast and Slow’, “companies would definitely like to improve the quality of judgments and decisions that are made on their behalf.” In an organisation, this is of paramount importance for the front-line manager layer where majority of day-to-day decisions are made to run the business. However, this is the same layer that suffers most from intuitive heuristics. There cannot be a better way to address this challenge across the organisation than cultivating the data analytics culture in it.
Albeit most of the large to midlevel enterprise already have an operational analytics capability and they also have SSOT (single source of truth) built through Enterprise Data Warehouse, however, majority of them are still at inception or struggling mid-way phase of their automated analytics-driven-business journey. What they are missing most in this is the fact that most of them never start with analytics as target picture in mind. Even though the value generation through analytics is happening in these organisations, most of these capabilities are not efficient. Over and above that, the hypothesis and experiment based analytics approach is missing from our work culture.
Reflecting upon my over a decade consulting experience in BI and Analytics domain and a recent enriching academic experience of Big Data Strategy and Implementation at UvA by lecturer and industry expert Mr Martin Heijnsbroek , I would suggest following aspects of tailored Big Data trajectory for organizations :
a) Revisiting the data strategy in the organisation
b) Consolidating the data platforms and up skilling of workforce
c) Analytics automation and making it process driven
To start with, by revisiting the data strategy, the transformation should begin with analytics as a big picture in corporate strategy. Organizations should start with the departments where analytics is already making or has the potential to make maximum impact through insight-to-execution process. For instance, departments such as corporate marketing, inventory management and e-commerce could be part of our pilot projects. They should redraw the expected trajectory of analytics driven marketing and inventory management departments. They should trace the supporting data lineage to build automated analytics capabilities that will allow a system of insights to drive these departments. Through data provisioning, they can ensure that the output of the system is ingested back and the delta correction is reflected through an analytics execution model. The learning from this pilot should be used to make the analytics system fully automated and further stimulate it to other departments.
Furthermore, organizations should form a strategy to consolidate the more than one SSOTs (if that exists) and encourage departments to build their analytics capabilities from single enterprise SSOT base. This would not only break the silos of data and encourage more collaboration across the departments but also bring down the total cost of operation. Furthermore, one of the vital aspects could be the retooling and up skilling of workforce across the company. This is essential not only for data engineers who handle plumbing aspect of data processing but also for analysts and front line managers who consume these insights through their reports and visualizations tools.
And lastly, all the transformation would be in vane if the data-analytics process is people dependent with manual interventions. At present, for most of the organizations, extracting the business insights from data is people dependent. As a consumer of the data-insights, an analyst creates the reports and shares it with executives for taking the strategic business decisions. In this data-to-insight-to-decision process, an analyst could be the weakest link. Therefore, it’s vital that through automation, process and application conversion, cycle of data-driven-decision-making is made fully automated.
