Back to Basics

Data warehouse consultants need to bone up on the soft side (and hard facts) of business

By Michael Venerable

Irrational exuberance is over. On that we can all agree. The technology shakeout was thorough and far-reaching. Well-known companies presumed to have bright futures have lost more than 90 percent of their market capitalization. The B2C and B2B e-commerce sectors have vanished, and companies too dependent on these next-wave industries have suffered accordingly. (Some suggest that B2B now stands for "back to basics.") Many well-known business intelligence vendors are in tenuous circumstances. Even category leaders face a more difficult sales environment than they have faced in their entire history.

In austere times it is best to return to proven practices that lead to successful projects and products. These practices may seem mundane or even obvious, but it may be time for everyone to just focus on putting one foot in front of the other.

TECHNIQUES FOR PRACTITIONERS

I spent the first 10 years of my industry life as a practitioner, working for customers or managing consultants for my own company. When I started, there was no Internet. There were no ETL tools, no graphical query tools, no data mining tools, and the most popular PC was an IBM XT. In the late 1980s data warehousing was not a big thing. The first project I worked on was akin to building a dam with buckets and shovels. Brute force labor was required to build the system, and the concept was brand new. Trendiness was not yet trendy, so we could not rely on the technical press and analysts to preload customers with need. We had to discover needs by listening and then propose realistic, achievable solutions. Much more listening and much less selling is the first practice I would recommend.

BE AN EXPERT LISTENER

Yes, I am advocating listening to the customer. Big deal. Well, actually, it is a big deal. For those of you who have joined the ranks since 1995 - when the data warehouse wave really came in - the last five years have been a cake walk. New ETL products, graphical query tools, analytic applications, and relentless hype made the data warehouse project an easy sell from 1995 until the end of 2000. ROI white papers from vendors and consulting firms provided a friendly backdrop for most business intelligence projects. There was a tornado effect in our space and the projects were often pre-justified. This was true even in normally austere organizations.

Those days are clearly behind us now. IT budgets may actually shrink this year for the first time in nearly a decade. When budgets shrink, necessities always come first. Decision makers consider a new data mart only after salaries, software maintenance, hardware leases, critical consultants, and other operating needs are satisfied. To understand where you and your project currently stand you must carefully listen to project sponsors.

Once-hot projects may now seem like luxuries. If the Web site upgrade or e-commerce project is postponed or canceled, then clickstream analysis might not be such a high priority. Projects that survive are those that clearly lead to top-line growth, increased earnings, or improved cash flow. In the event that you do hear warning bells, focus on the second guiding principle, which is:

BE AN EXPERT IN THE CUSTOMER'S BUSINESS

You must know the business of your customer. You must know more than what they sell or manufacture or distribute. You should clearly understand the customer's business model and how your project fits in. Do you know how the business is valued - by cash flow, earnings, EBITDA (earnings before interest, taxes, depreciation, and amortization), or something else - what makes up cost of goods sold, how the income statement is structured, or what the value chain looks like? If not, you might want to do some more research. Remaining relevant in difficult times requires that you have a far better understanding of the soft science of business than your colleagues who manage the storage area network or the new Web portal.

In recent years, many practitioners simply ignored basic elements of business knowledge. I vividly recall attending a 1999 presentation by an e-business consultant to a group of managers at a multibillion dollar industrial company. This gentleman derided Wal-Mart because the company's revenue per employee was considerably lower than eBay's revenue per employee. With this apples-to-rutabagas comparison he displayed a complete misapprehension of the two companies' business models and value chains. Understanding your customer's business will make your listening more productive. It might also save you some embarrassment.

The easiest way to test your understanding of the customer's business model is to be sure that you can map your project to the income statement or balance sheet. This task might appear boring at first glance. But in a tighter operating environment, displaying an intimate knowledge of the business might just keep your project alive. Relate your project to a reduction in cost of goods sold, or an improvement in EBITDA. Hook into a line item on the income statement or balance sheet, find the people concerned about that line item, and bring them along as sponsors. These details represent the concerns of operating managers and executives, especially when they are under intense scrutiny by shareholders and financial analysts.

BE AN EXPERT IN YOUR OWN FIELD

When I worked on my first data warehouse project in the late '80s, there was only one book that I could find on the topic. Conferences and seminars were nonexistent. In preparing this column, I looked at that book again to see how it read in hindsight. Not surprisingly, many of the techniques and ideas seemed dated. We have made remarkable progress in the last 10 or so years as practitioners. The original idea remains novel, but the practices and technologies in place today were unanticipated back then.

To pay proper tribute to the idea and the pioneers (such as this column's editor), I ask that you become an expert in your field. If you are a regular reader of this column and others like it, then you are on your way. Reading a sampling of nonproduct-specific books each year is mandatory. And by reading, I mean reading to understand, internalize, and criticize. Attend conferences and seminars, keep up with vendor developments, and work around the edges of your current comfort zone. Understand emerging technologies such as multidimensional expressions, storage area networks, and XML and their importance to data warehousing. If you are really bold, add to that body of work by writing an article or a book. The act of organizing and documenting your thoughts into a published work will add to your expertise immeasurably.

Another excellent way to hone your knowledge and test it at the same time is to teach or train others. It is especially valuable to teach core concepts in such a nascent field. Inevitably, you will encounter a question from a student that challenges you in an unexpected way. Find opportunities to teach others, and you will emerge better prepared for your next assignment.

RULES FOR VENDORS

For vendors that thought they would be able to skip the first part of the column, I have some bad news. My recommended practices for vendors are exactly the same. In addition to working as a consultant, I have spent the last four years working with or for vendors in the business intelligence market. In that time I have noticed that vendors that listened carefully to practitioners and customers have weathered difficult times and navigated changes in the marketplace quite well. I have also noticed vendors that follow business needs and market changes most closely are those that gain market share quickly.

Finally, it is imperative that vendors employ or consult expert practitioners in the data warehouse or business intelligence space. The differences between solutions in our space and those in the mainstream transaction market are growing. Simply showing up with a data warehouse version of some core technology or development tool will not satisfy the increasingly specialized needs of the data warehouse world.

With these principles in mind, I would expect any practitioner or vendor to survive in this new old economy just fine. It's time for "B2B."

Guest columnist Michael Venerable [mvenerable@oaktonsoftware.com] cowrote Data Warehouse Design Solutions (John Wiley & Sons, 1998) and is an expert in applying dimensional modeling techniques to real-world business problems. He is an executive with Required Technologies, a New York-based research company.