APICS is the premier professional association for supply chain management.

Big Problems, Small Technology

By Jonathan Thatcher, CSCP | May/June 2013 | 23 | 3

Jonathan Thatcher APICS Director of Research Ask APICSUnderstanding your customers with big data

Reader L.P. writes, “Big data is a hot topic at my company, but we are not a global enterprise and lack a massive information technology budget. Some are saying big data is not possible for us right now and that we can only aim for traditional business analytics. Is there anything we can do with our existing systems?”

Suppose you are a supply chain manager in a business environment characterized by 
  • high competition with even greater pressure expected in the future
  • relatively weak customer service and supply chain partner relationships 
  • management that wants to develop more new customers and demand rather than relying on the current customer base 
  • demand enhancement efforts that are not as effective as they could be.

These pressures can spark differences in opinion, less-than-objective dialogue, and uncertainty among stakeholders. Big data helps avoid some of these challenges by gathering and leveraging all available data into insights and objective-based decision making. 

The APICS research report “2012 Big Data Insights and Innovations” defines big data as “a collection of data and technology that accesses, integrates, and reports all available data by filtering, correlating, and reporting insights not attainable with past data technologies.” Most definitions similarly imply you can’t deliver true big data results without large-scale platforms, systems, data gathering, and analysis. 

However, big data can inspire organizations of any size. The result may not be big data in the strictest sense, but you can end up with a useful head start on business analytics that can serve as a foundation for eventual big data adoption. Start with a pilot project.

Whether formally or informally, most organizations score their products and services and product families. The 80-20 rule, or Pareto analysis, is a common way to do this. Products with higher scores get more attention and priority given limited time and resources. However, do you likewise segment and rank your customers according to shared importance factors? Scoring each customer enables development of trend analysis via “deep diving” into customer data. This can help answer such questions as: Is the company losing or keeping its most valuable customers? What about customers in the highest-value segments? Are you attracting more high-, medium-, or low-value customers? Which customers have the best relationships and customer service? For each trend, ask the five whys to get to the root causes. 

Customers might respond to offerings and promotions differently. One reason for this may be that the customers who are easiest to work with, no matter their value, receive more of your organization’s time and attention. High-value customers may expect better-than-average service, therefore seeming more difficult to work with and getting less attention. Driving to the root causes helps establish beneficial procedures and resources. Eventually, you can raise the bar and transition your medium-value customers into high-value customers.

Follow up your scoring with a survey or conversation. Find out how each customer regards its own importance to your company. There will be a spread of opinion, which might not match your own scores. Once again, ask the five whys and determine root causes. Figure out what the same survey should show in six months or a year.  

Each time a customer interacts with your organization via emails, conversations, new orders, and negotiations, consider this an event. Each event represents a data point, large or small, that can be captured by big data systems that correlate these clues with other data to help develop insight about the customer. These insights can include a customer’s sentiment, intentions, satisfaction, and the state of the business relationship. Over time, the clues paint an objective picture of the customer. The results will suggest strategies such as relationship-building or customized value-added promotions, with the ultimate goal of becoming more tightly aligned with the customer’s needs. 

For your pilot, begin with a handful of customers, having a team manually capture and interpret clues from customer events. You might keep track of such factors as the length of time customers take to respond to offers and communications, the speed of payments, the tone of recent communications, expressed trust levels, and many others. If, for example, a long-standing customer changes an order, is this part of a trend pointing to a changing relationship? A single order change might mean little on its own, but something larger may loom after combining all event-based clues from that customer. 

Vast insights might not present themselves at first, but over time you may spot new opportunities. Through the project, you may see future big data analysis that would integrate well with existing processes. 

Perhaps your account managers already perform this sort of analysis. But do they do it for every customer based on their score? A big data system has the potential to process clues from all customers continually and correlate findings with factors such as order shipments and payments. A recommendation based on these insights could be critical to winning a future order. 

Combining these clues and customer scores helps refine and optimize time and resources for both you and your customer. Perhaps less-than-efficient customer practices reveal weaknesses in some of your facilities or processes. The causes can seem insignificant—but what seems insignificant can add up to hard data showing clear trends. 

Without large-scale computing power and information technology resources, you may be restricted to a manual approach in your pilot effort. While a manual, human-scale pilot seems limited compared to the scope and promise of big data, it may reveal how to specify, configure, and develop future capabilities. 

The scenarios I’ve described more closely resemble business analytics than big data, but hopefully they have provided a glimpse of what is possible with the right technologies. No matter how big data develops at your organization, it will likely be a partnership between information systems and the human ability to connect insight to strategy, tactics, processes, relationships, and root causes. Don’t let the sometimes blurry line between business analytics and big data sway you from the real issue—making the best use of your data now and in the future.

Jonathan Thatcher, CSCP, is director of research for the APICS professional development division. He may be contacted at askapics@apics.org.

All comments will be published pending approval. Read the APICS Comment Policy.


  1. RadEditor - HTML WYSIWYG Editor. MS Word-like content editing experience thanks to a rich set of formatting tools, dropdowns, dialogs, system modules and built-in spell-check.
    RadEditor's components - toolbar, content area, modes and modules
    Toolbar's wrapper 
    Content area wrapper
    RadEditor's bottom area: Design, Html and Preview modes, Statistics module and resize handle.
    It contains RadEditor's Modes/views (HTML, Design and Preview), Statistics and Resizer
    Editor Mode buttonsStatistics moduleEditor resizer
    RadEditor's Modules - special tools used to provide extra information such as Tag Inspector, Real Time HTML Viewer, Tag Properties and other.