Data warehousing is getting pretty long in the tooth. Not many people see it as a driver of renewal or transformation in an enterprise. There are lots of reasons for this, the most notable being that its original formulation is basically flawed ­– a largely passive repository of data capable of very little except being a middleman between unruly source systems and a swarm of downstream data marts, cubes, reports and plain-old extracts. That's the original formulation. Luckily, many chose to deviate from that formula and create data warehouse architectures that were less magnificent in their (promised) scope but far more useful as a tool. Way back in 1995, one case in particular decided to take a critical look at this landscape and came up with something that was truly transformative.

Can a data warehouse be transformative? A data warehouse alone cannot be transformative; but it can be a part of, even the center of, an initiative to change how an organization does business. It can be part of a program to move people out of their comfort zones and adopt some new ways of doing things. Here is one example of how a company I worked with did it.

On the surface, they appeared to be a cosmetics, fragrances and skin care products company. Instead, they were a fashion business. Their ethos was not selling products; it was projecting an image of beauty and sophistication and fiercely protecting and nurturing their brand image. That's very different from just selling lipstick. That fact is key to understanding why they made the decisions they did and, in some cases, why it took so long to arrive at them. For example, it wasn't just the levels of inventory at a counter that mattered to them; it was how the inventory looked, how it was arranged and the pleasing display of colors, shapes and sizes that mattered to them. When a customer stood at the counter, the inventory was part of the experience. The decisions were more complex (and so were the models) than just fast-movers and slow-movers.

They knew what they produced, what was in warehouses, what was shipped and returned. What they did not know was what exactly happened to their products after that. Even though they had a very close relationship with their retailers, the stores were not theirs. Tracking movements in shipments that occur every week or two is not useful in understanding the gestalt of merchandising; and that is what they needed to understand.

After spending some time studying the problem and gaining an understanding of the upside potential of having that information, a plan was agreed upon that was, for the time, audacious. We would collect this information from the stores and allow up to 2,000 employees, many of whom were remote and on dialup, to have direct access to the actual data warehouse –­ not a summary –­ with the ability to use a ROLAP tool through a thin-client interface (a Netscape browser). Allowing users access to the complete data warehouse for ad hoc analysis was the first part of the conventional data warehouse rule book that we violated.

The second part of the data warehouse rule book that we ripped up was the accepted gospel that the blessing of senior management is essential to success, although they were very supportive after the fact. In fact, we didn't even have the support of the IT organization. For those initiatives that are meant to be transformative, gaining senior management support upfront may actually be unrealistic. There was a charismatic and powerful CIO who, short of supporting the project vocally, at least provided some resources and cover. Luckily for us, that turned out to be enough.

When the project was finished, the entire organization had access, for the first time, to a complete set of sell-through data ­– roughly 3,000 counters in the U.S. This had an immediate and stunning impact on the organization. Imagine living your entire life with cataracts and suddenly being able to see clearly for the first time. This is precisely how people felt. The response was overwhelming. They could, at any time, construct their analysis against the entire database and actually understand, with clarity, which products were selling and when, which products sold in combination with others (market basket), promotional effectiveness, local media response and even the effect of local weather. These analyses were never more than conjecture before.

Maybe moving from shipment data (a relatively good indicator of demand) to sell-through data is just an incremental improvement; but for this organization, it was far more. This capability changed the way they did business. First, by being more fact-based in their decisions, they were able to make smaller adjustments in shorter time frames. As one brand manager told me, "Reaching our targets is not just a function of making a few big, strategic decisions once a year; it's making hundreds of small ones every day." Second, because the semi-annual promotions in each store represented a disproportionate level of sales for the year, their timing and execution were critical. The availability of this data for planning purposes, as well as monitoring purposes during the promotions, increased the effectiveness of those limited opportunities.

Ultimately, this data warehouse became the bedrock on which the continuous replenishment system was built, a win/win situation for the company and the retailers. Today, the in-house group is still making improvements and enhancements to the system, but the fundamental design remains unchanged. No one in the company can imagine what it would be like without it.

Why did it work? What was so different about this effort? It worked because this data warehouse was conceived to do something that was never before possible and was important, even central, to the organization's operation. In simple ways, it just changed the flow and timing of information, changing roles of publishers and subscribers of reports. Everyone became a publisher and subscriber. However, more fundamentally, it changed the way the organization looked at its business and provided the mechanism for making faster, more accurate decisions and implementing them, closing the loop between events, measurement and action.

The opposite of the transformative data warehouse is an existing data warehouse environment of average capability and disappointing results that can be revivified into a transformative effort. That will be the subject of my next column.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access