Organizations that use data effectively focus on the business problems they want to solve using analytics and the decisions that need to be taken with the help of such data.
The words “data”, “analytics” and “algorithms” are being heard across organizations. Boards are prompting companies to go digital and use analytics for decision making. Even the government is planning to use analytics in its functioning. For instance, the chief minister of Andhra Pradesh tracks the progress of key initiatives in his state through descriptive analytics on a visual dashboard.
So how exactly does one kick-start an analytics journey in an organization?
Organizations that use data effectively focus on the business problems they want to solve using analytics and the decisions that need to be taken with the help of such data. Thus, the two key operating words are “data” and “decisions”.
The Gartner Framework on Analytics is a good reference to begin the analytics journey. It shows how data is at the root of all decision making and, based on the quantum of human input involved in decisions, one can classify the different stages of analytics as descriptive, diagnostic, predictive and prescriptive. Further, once the decision is acted upon, more data is generated which is fed back into the database and used in the model.
For any organization, there are four important steps for setting this up.
Set up the right data flows for the right business need
The first step is to set up the right data flows within those departments which impact growth and cost the most. For some companies this is in the sales function where the interaction with the customers is the highest. In some industrial companies, it may be in the factories with high quantum of operational data, while for yet others, it is in the vendor interactions involving large amounts of purchase and cost information.
Examining the reliability and availability of this data at the right time with the right quality, generates the initial agenda for any analytics journey.
Enable KPIs and descriptive analytics
The next step is to convert this data into useful information which can be used for decision making. The focus here is on Key Performance Indicators (KPIs) which help describe what is happening in the organization and why.
While such information was available in the traditional business intelligence (BI) systems, it is now possible to get the same on a real-time basis, at a granular level, with visual dashboards and have it delivered across the hierarchy of the organization. Such analytics is called descriptive and diagnostic analytics.
A graphical means of looking at information brings alive the causals in an effective manner and shows the outliers which can be actioned right away.
For example, in a fast-moving consumer goods (FMCG) company, one can get daily SKU-wise, bill-wise secondary sales to retailers instead of the earlier brand-level totals, owing to better bandwidth and lower storage costs (SKU stands for stock keeping units). These can then be structured to trigger alerts on an exception basis to enable better performance reviews. Such dashboards can also be rendered on mobile devices and can be deployed to field supervisors to enable their daily reviews.
Kick off predictive analytics projects
Lower bandwidth costs coupled with cloud computing enables enormous amounts of past data to be now accessible for better predictive analytics. In fact, predictive analytics can act as a base for further scale-up towards algorithmic and artificial intelligence (AI)-led business. Hence it is necessary to set off a few projects in this space to build analytics capability.
Typically the easiest and high-impact projects tend to be in the realm of forecasting. This can be in finished goods forecasting for consumer goods companies, raw material price forecasting for commodity-led companies, or predicting market growth, customer loyalty or risk profiles for banking customers, etc.
Predictive models are also used at an operational level—predicting breakdowns or downtime in factories and for attrition and retention forecasting in human resources. Setting up action standards, allotting the right resources for the project, and then collecting the relevant data and working along with the right partner helps generate success in such projects.
Develop a talent and process culture
In order to build the analytics capability, it is also necessary to put in place the right governance processes as well as develop the right talent and culture.
Letting algorithmic models run on clean data alone will not drive analytics. It is essential to have talent with the right skills like data management, statistical and data processing skills, and business acumen.
And finally, such initiatives need to be sponsored right from the top with correct, timely governance and review processes.
It is imperative that the end objectives be always clear to all the team members so that one does not get lost during the journey of any new project.