This article is co-written with Arild Nebb Ervik.
Analytics and statistics have been around for a long time; it is almost a hundred years since an assembly line improvement made the T-Ford affordable for a regular worker. Since then analytics have been an integrated part, and a key enabler, of many successful business strategies.
To use real time data to gain strategic advantage or to identify and control risk is not a new concept, what can be said to be new is the shift of the usage of term “data”. Where we traditionally were talking about structured data – as they developed from our ERP, customer, billing systems etc. – we now tend to use a broader perspective, including unstructured data such as documents and records, emails, social media and other text based sources.
There are lessons to be learned watching the frontrunners in usage of analytics. We observe that Business Analytics on big data are especially gaining momentum within two types of business. High volume, high speed B2C business like credit cards, telecom and retail. And highly complex industrial operators within Railroads, Plants and Oil wells.
What brings those different businesses into significant investment in better analytics is a bit different. The credit card players strive to increase top line growth through faster and more agile customer and marketing analysis, while at the same time protecting bottom line through extensive risk and fraud analysis. The oil companies primary target, however, might be to optimize costly maintenance and reduce the risk of unscheduled shut downs.
The bottom line for both examples: It’s about using information as a strategic differentiator.
I will try to headline some areas here where Big Data & Business Analytics usage is enhancing quite advanced business processes:
Condition Based/Predictive Asset Maintenance
Oil Platforms, Telecom network infrastructure, rail infrastructure and even vending machines, are producing a wealth of data, often as a continuous stream, that can be analyzed and used to enhance risk management or maintenance processes for almost any asset with some sort of strategic importance. Today, the trend is to install sensors to measure e.g. frequency, temperature, humidity, etc – to detect trends and patterns that may predict a failure or breakdown on a mechanical part, before it actually happens. Hence a condition based maintenance strategy become more feasible; expanding the lifetime of assets covered by an interval or time based maintenance strategy – and at the same time being able to avoid critical and unexpected downtime. The same data, models and techniques can be applied to detect root-cause effects; not only predict when and where an unexpected event happens, but also the reason why this is happening. We know from our projects, and this is confirmed by US Department of Energy; that condition based maintenance projects are on average able to reduce the number of unexpected breakdowns by approximately 75% and at the same time reduce the maintenance cost for the same asset. As a result the ROI could reach 10 times the investment.
Advanced Optimization, Planning and Scheduling
The new information available, from a multitude of sources – not only the ERP systems, makes it possible to combine several sources of information that until now have not been looked at together. Companies want to maximize the usage of critical resources – and control the cost. Hence we need asset data, human resources data, and financial figures. Using railway as an example, imagine this chain of questions:
- What alerts should be answered first?
- How can I optimize our maintenance strategy?
- How do we determine the optimum between cost/funding, stock/inventory/spare parts, quality/risk/downtime and man-hours/competency to maximize the customer utility/experience based on the resources available?
Truly understanding your customers – what is your share of wallet, what will they buy next – when will they churn – is about combining historical customer behavior with real time (or near real time) information to profile your customers and tailor your communication and how to make interaction with them more effective and mutually beneficial. It is not only about knowing who your customer is, but to try to anticipate their needs and next moves. It is about being a proactive partner – not only a reactive provider.
Social Media Analytics
Broadly it is two different approaches. Listen and learn: understanding the market perception of your company, brand, products and services. And Identify and react: rapidly detecting and responding to negative sentiment or proactively position your brand trough e.g. using geo location data to customize own or third party partners direct marketing.
Risk & Fraud Analytics
For the credit card company eager to gain market share or the tax authority trying to optimize tax payments from citizens. Risk and Fraud management is critical to all businesses. Understanding and quantifying the risks facing your business and modeling risk scenarios give business managers the tools to effectively set the right business strategy.
The common theme for all these cases: They require real or right time insights into increasingly complex questions; and these questions seem to be increasingly data demanding.
The value generated from business analytics will come from starting to systematically learn from analyses, by advocating fact-based-decision-making within the organization and, as I concluded in my former post about key challenges for Business Analytics:
Regardless of how “big” the data are, success in analytics relies at least as much on organizational alignment and process as on the chosen analytical tool.