Six ways to go from Analytics to Outcomes

Analytics may seem like new phenomenon, but it really isn’t. A newspaper delivery boy has been using it for ages to optimise his route. Frederick Taylor used it in time and motion studies. For several years, carmakers have used it to lower inventory costs and predict usage of parts. And large retailers pioneered its use in reducing stock-outs and optimising inventory.

Why, then, is analytics all the rage today? Because of the volume of data generated, with the ability to collect and analyse it at much lower cost. But it’s easy to get lost in the technology mirage—how does one avoid the pitfalls of analytics becoming just another data set? Here are six ways to overcome the challenge and go from analytics to outcomes.

Over the years, data has become the fiefdom of the function—be it HR, finance, or marketing. Such data, residing in function-specific data servers, seldom travels to the rest of the enterprise. Consequently, the key performance indicators (KPIs) that the function head tries to improve also tend to be ones that reside in his silo. However, to improve a KPI, a process or a sub-process must be viewed with the customer in mind, as must the handoffs from one sub-process to another. Analytics becomes far more successful if KPIs or metrics are measured across business processes like procure-to-pay or order-to-cash, which cut across functional silos.

A case in point: To reduce the time taken for vendor payments, the procurement team needs to ensure that invoices are received on time and passed to the finance team well within the cycle time to process an invoice. To improve this KPI, one needs to ascertain the time to digitise an invoice, the terms and conditions in the master data to process an invoice, and the number of invoices withheld for dispute resolution. Each of these should be measured by sub-KPIs using dashboards and appropriate visualisations to arrive at the right reason for low performance of the KPI. If data silos are not broken, the effectiveness of analytics shrinks to a fraction. A strong and robust business intelligence (BI) framework will make life easier, with the right combination of data structures like warehousing, ETL and other technology solutions that speed up report generation and enable faster insights.

Domain experts will need to identify strategic and tactical KPIs that need improvement in a given scenario and context. It is possible that several KPIs have never been measured in a particular process. For example, while fulfilling a sales order, it is possible that an organisation has never gone beyond anecdotal reasons for improving on-time delivery or identifying reasons behind frequent delayed payments. These KPIs must be made measurable by having the right parameters and data points to measure.

There is no end to improving KPIs. Once the KPIs are identified and appropriate dashboards for measuring the feeder metrics are set up, it is then a continuous exercise to improve the metrics—some long-term and some short-term.

An analytics project becomes successful only when the business operations team adopts a data-led system for decision making. Several hypotheses will arise from anecdotal incidents, the importance of which must be judged by data.

For analytics to succeed, BI must be democratised across the organisation, so that data is brought together to work on KPIs. The usage of analytics will be a culture change and will need to be driven by the CXOs and their respective teams, who have clear vision on the KPI that needs to be improved. The CTO/CIO’s role should be that of an enabler.

Most enterprises were built on straight-from-the-gut decisions by their leaders, and not data-guided decisions. Consequently, most leaders believe they know what’s best for their domains, and enterprises lose out on well-timed decisions that could come from a vast employee base. But to promote data-driven decision-making, one must empower employees—encourage them to look at appropriate data and then make objective choices. This starts with training employees on using analytics and, most importantly, ensuring the data being collected is usable for decision-making.