Q&A: Lack of Data Background, Quality Can Hold Back Business

Too many business executives lack not only good data, but an understanding of the data’s origination and history.

Drew Rockwell, CEO of MDS and an expert in business process management and business analytics, points out that a surprising number of executives lack any detailed knowledge of their own business data. In this interview, Rockwell discusses why executives have such poor access to good data, why he thinks BI adoption is stagnant, and the importance of breaking through information silos to get information distributed throughout the company.

Rockwell’s company, MDS, provides business performance analytics software to companies around the world and in a variety of market sectors; customers include Comcast, Ernst and Young, Intuit, and Ericsson.

BI This Week: Why do so many executives lack access to good data within their own companies?

Drew Rockwell: Good data equals accurate, timely, and useful information about what is happening within a company.

This is hard, as not only are data volumes growing exponentially, but every application, every spreadsheet, every field in a record, every device, and every transaction is a source of potentially valuable data in conducting a useful analytic. Moreover, valuable data often comes from third-party sources all along a supply or value chain. It is incredibly hard and expensive, and some would argue impossible, to corral all this data and position it to be analyzed across sources. One spends a lot of time and money pouring the foundation before even the first room can be framed.

Even if you have what you believe is the relevant data, it is hard to construct complex business analytics across multiple sources of data. First of all, programming languages are broadly used to manipulate data across hundreds and even thousands of discrete steps to get to the answer. This is a very opaque process for business or operational leaders who do not speak SQL or Java, but understand the business logic or rules they want to apply to the data, so they need to understand the analytic process in order to believe the output.

Business leaders are also becoming dissatisfied with merely asking questions or queries; rather, they want to explore the data to discover what they don’t know. They want and need to be much more closely joined to the data analytic process, much faster. There are times when they don’t just want the answer; they want to know how you got the answer.

What kinds of tools do executives use now for making financial- and BPM-type decisions?

I’m always surprised at how prevalent Excel still is. It is a testimony to how powerful a business tool it is, but it has some very real limitations.

For example, I was recently at a gathering of a few hundred CFOs (most from Fortune 1000 companies). When a speaker asked the attendees how many depended on Excel feeds to complete pieces of the monthly close, nearly everyone raised their hands -- not surprisingly. Next, however, the speaker asked whether audience members had methods to validate that the calculations being driven from those workbooks were accurate. Almost no one raised their hands. That’s actually quite startling when you stop and think about it.

Why aren't complex analytic processes traceable and understandable? Is it the tools or the way they are being used?

The industry has made fantastic advancements in the processing of huge volumes of data, as well as in the presentation of that data and the ability to conduct ad hoc queries, but there has not been enough investment in changing and optimizing the analytic process.

An executive might be shown a dashboard or set of charts each morning that was created by processing four billion records, but they have no idea if the analytic steps behind those reports are accurate. For mission-critical analytics, they want that transparency.

TDWI studies show that user adoption of BI tools is stagnant -- users who could clearly benefit from BI simply aren't adopting BI products into their daily routines or using BI for decision-making. Why?

I think this comes down to time. Executives need to make decisions based on data and calculations from a lot of different sources and do it quickly. Doing that means they need to be able to trust what they are being presented, and most BI tools don’t enable that.

Visualization is critical to helping solve this, but not visualization in the sense that executives are presented with a pie chart or set of graphics -- Excel can do that easily. Rather, the analytic process used to arrive at an analytic conclusion needs to be more visually accessible. In some sense, the “seeing is believing” adage really applies here.

What are some of the most challenging areas of the business for getting meaningful, current information to executives?

Near-real time information at a detail level, not an aggregate level, is a significant challenge.

I know of very large service providers that have analytics in place to monitor every single customer order and service delivery/installation or maintenance visit to validate whether it was performed correctly. Previously, they had to aggregate data or wait for a customer to complain.

With their new approach, they identify the specific representative technician or process that is causing errors in near-real time. Previously, when they saw aggregate issues, they needed to deal with them with aggregate management tools such as group training or massive process redesign. Today, their analytics are much more granular and pinpoint the exact cause of the error rather than broad systemic problems. That is actionable insight using current data that saves these companies real money and provides a much better customer experience.

One issue is that data is often in many different "silos" throughout the company, meaning different owners, different management, and different tools. How can companies address that?

A lot of companies immediately think that data warehousing is the place to start solving this. Whether it is the right solution in the long run, one thing is certain -- this approach costs time and money.

Some agile analytic tools can help companies get to value much more quickly and easily. It may be that the right long-term solution is then to automate that analytic in a warehouse. It may not, but there are ways to shorten the time to value with these new generation approaches.

Perhaps more important, application systems tend to breed a “silo” mentality. People who run CRM systems often do not have great visibility into what is happening to the data once it flows downstream.

Most of the problems in a business happen in the white space between systems and processes and organizations. A more collaborative, workshop, discovery-driven methodology to analytics is a huge opportunity to transform a business.

What does MDS Lavastorm Analytics bring to this discussion?

Our software and solutions simplify the design and execution of complex analytics, as well as the acquisition and transformation of the data required. We help businesses and professional users analyze and assure that business is performing as designed. Our brand promises to help business and professional users analyze, optimize, and control their operational environment through powerful discovery-driven analytics that deliver value quicker and easier than competitive alternatives and are far more extensible.

To use an analogy, our software is more like a MRI than an X-ray or blood test. We create visibility and transparency at a record or field level of data. We have found that there is huge value to be gained by creating analytics and controls at that level of detail.

comments powered by Disqus