Opinion
A strategist from a large bank shared their focus areas related to data analytics with us. In that were areas such as Robotics, Block Chains, Bots, Artificial Intelligence, etc. While this long term view is commendable, this bank was ignoring huge opportunities for applying machine learning that can bring great efficiencies to their business, today!
Looking for Big Sexy
Today when companies look for Analytics projects, they have in mind large, complex projects where software applications look at data and talk back to business managers with recommendations and prescriptions for all their business problems — a Star-Trek like fantasy if you will. There is lot of talk about how Watson interface will help people interact with software using natural language, and give sagely advice on addressing business problems. This future is certainly desirable, but some ways off.
Getting Business Value Today
My strong belief is that for most companies, Analytics will be an evolution. In his HBR article, Dr. Andrew Ng gets right to the heart of where value will be created — Data and Talent. Both these will take time to acquire. Companies can today undertake many smaller analytics projects that will deliver high value — there is lot of low hanging fruit to be harvested. The benefits in organizational learning will eclipse the business benefits from the low hanging fruit themselves.
Some Examples of What To Pursue
Real-Time Dashboards and Live Scorecards
Many companies do not have the habit of staring at performance numbers. In fact, many are uncomfortable and make detailed excuses on why the politics around it will cause more harm than good. Real-time dashboards and live scorecards help companies get used to staring at reality, good or bad. In the short term, and especially when business goes through rough patches, these boards turn out to be painful reminders. Once organizations develop the maturity to truly run business by the numbers, there will be greater focus on problem solving and less on finger pointing and hiding.
Process Mining
Companies where process excellence is an important element of competitive positioning need to leverage Process Mining. Instead of going through one-time exercises of mapping processes by hand, they should let data describe the processes. Business teams should be spending most of their time on analysing and improving processes. Internal process teams would ideal candidates to leverage such tools, and one of the key responsibilities that they should have is to bring these process analyses online by creating live process mining workflows that are always connected to the underlying data.
Predictive Analytics
Applications for predictive analytics are ripe in many areas of business. Businesses should target areas where lot of data is available and impact can be seen rapidly. For example, a bank can use credit modelling to develop objective credit scores for applicants. Marketing can score customer leads and filter them before passing them on to sales. Manufacturing engineers can use process data to predict when to do predictive maintenance, etc. The list goes on and on.
What Needs to Change?
Companies must embark on deploying large number of smaller solutions developed by internal teams themselves, and not plan solely for large analytics efforts. Instead of waiting for large budgets, long approval cycles, and expensive company and external resources, this approach recommends that companies go in for small yet rich applications that can be built on-demand.
Our 2 Cents
The first set of data analytics applications built will probably be simple ones, but also open the organization up to the huge potential of data analytics. The important thing is that companies should take their first steps in learning to work with enterprise data and building organizational talent in analytics that will deliver value to them. We believe that under this approach, businesses can greatly benefit if they follow these three rules or guidelines
1. Any application should not take more than a few hours or days to build
2. Once built, applications should be capable of being rewired and improved over time
3. All applications built should run off enterprise data, and not create disconnected data sets
Going back to the bank discussed at the beginning of this blog….while companies should focus on long term, strategic projects in analytics, they should truly value the present and get going on smaller projects right away!
