/ Analytics

The Need For Speed

Making a case for quick access to insights derived from analytics.

In their book “Information Orientation: The link to business performance” (Oxford University Press;2002), Marchand, Kettinger & Rollins show that the degree to which a company implements and realizes the synergies across three information capabilities, predicts business performance. These three capabilities are the following:

  • Information behaviours and values (i.e. People);
  • Information management practices (i.e. Information);
  • And information technology practices (i.e. Technology)

This synergy is generally accepted in the business world and has been substantiated in a number of subsequent case studies.

However, it becomes increasingly clear that there is a trick to reaping the business performance fruits of this synergy… the speed at which this synergy operates.

Why Speed?

As data scientists, we see ourselves as translators of analysis outcomes into business insight and understanding. This means that we often find ourselves in the tough spot of having to inform decision makers of insights that drive business strategy; and very often those insights need to be derived quickly, using the latest data.

Why-Speed-Figure01

This poses a two-fold challenge:

How quickly can:

  1. Data turn into information?
  2. Information drive decision making?

As stated in our premise, many businesses understand that decisions should be driven by clear understanding of what is happening in business (either from an internal Voice of Business perspective or from an external Voice of Customer perspective). However, too often we find that the time it takes to collect, consolidate and clean the underlying data is a long-winded, time consuming and tedious task. In fact, a recent survey of Data Scientists found that almost 80% of a data scientist’s time is spent doing exactly this: finding and cleaning data. (Crowd Flower; 2016 Data Science Report).

How fast?

In his recent piece on “The Analytical Executive” (October 2015), Michael Lock states that Executives are becoming increasingly driven by Analytics and, furthermore, that 33% of these Analytical Executives need information to be delivered within an hour. Now, in the analytics real world, that is lightning fast, because few businesses can muster such impressive performance. The reason why many businesses are not yet in a position to perform at such speeds, is two-fold. Firstly, many businesses are hesitant to invest the time and effort it takes to develop an engine that runs that fast. Secondly, having the engine only brings you half-way: You need the right fuel – every time – to make this machine run the way it ought to.

How-Fast-Figure02

What to do?

1. Get your house in order.

You have data (CRM, Transactional data, Employee behaviour, etc.). Structure your data in a manner that allows you to pull specific information. Easily. A lot of this refers to the much bespoken (but actually hardly come by) commodity called “the single view of the client”. Considerations in this respect include basic elements like data cleaning, formats and storage (e.g. local vs. cloud).

2. Build a world-class engine

This engine consists of tools (hardware and software) as well as the right people.

These people should ideally consist of a team of good communicators, with past exposure to various aspects of business and who has interdisciplinary skills in the fields of programming and analysts. This team should also have a clear view on the direction the business is aiming to move towards, so that they can facilitate pro-active decision making, rather than post mortem type analyses. It is also recommended that independent thinkers included in such a team. Such team members help to keep ideas fresh and to create a degree of discomfort.

What not to do?

Unfortunately, the establishment of good quality data sets and streamlined analytics engines requires significant investment from business side in terms of time, effort and funds. This often deters companies from embarking on such a daunting initiative and eventually they just never start. Be careful not to fall into this trap.

Start. Even if you start small. In fact, starting small is highly recommended. (Mui, Chunka, Forbes 2016, Thinking Big About Big Data in Insurance).

As the Chinese proverb goes: The best time to plant a tree was 20 years ago. The second best time is now.