Baxter Healthcare is a $5.5 billion manufacturer of medical technologies related to blood and circulatory systems. Baxter IV Systems is focused on intravenous fluids, nutrition medications, technologies for administering IV therapy and anesthesia, as well as clinician management systems. With revenues of $2 billion and over 17,000 employees, IV Systems is the single largest division of Baxter Healthcare.

Baxter IV Systems products have been primarily distributed through direct channels, smaller dealers and large distributors. However, as managed care has started moving health care away from the hospitals, there has been an increased focus on our alternate site channels such as home care, surgery centers, outpatient clinics and doctors' offices. It has become imperative that we have a better understanding of this aspect of the business. Our salespeople and managers of our alternate site business are in dire need of more accurate and timely information.

Baxter had already developed an enterprise data warehouse using Teradata and Oracle databases running on a Digital Alpha machine. This warehouse contains detailed data across multiple subject areas including financials, sales data (down to the invoice level) and inventory management. This was valuable data, but truly a complex analysis environment that was not for the faint of heart. For the analysts who required this amount of detail to answer their complex queries, the warehouse was ideal. Unfortunately, we discovered that this solution did not address the large number of users that needed to conduct quick and timely analysis of their business on a more regular basis. This is when we realized that we needed another solution: a system that could provide the users with timely query and reporting on the state of the business, but also the ability to drill back down into the warehouse when it was required. Broadbase was the answer.

In the decision support arena, query response time needs to match "think time" (the time it takes a user to visualize the previous answer) and hypothesize the next. While detail data in the data warehouse is extremely important, since it is the cleansed view of corporate transactions, pre-aggregated data in a dependent data mart is a requirement to provide the responsiveness needed by end users. Broadbase was able provide answers in seconds that used to take 40 minutes.

The Implementation

Broadbase Information Systems, headquartered in Menlo Park, California, developed a complete data mart product that is focused on reducing the complexity that is typically associated with data warehousing without compromising functionality. Because of the fact that Broadbase has integrated all of the required functionality of a data mart into a single product, we were able to have the prototype of our channel management application up and running in less than three days.

After a brief introduction to the database, links were established to the two data warehouse systems and source views were created to establish the scope and view of data from the data warehouse. Then the local Broadbase fact table was defined and loaded. This table represented the fact table of the star schema design and used special Broadbase functionality such as virtual calculated columns and supported a multidimensional cube on selected dimension groups (to enhance speed while saving disk space). Performance was enhanced by applying indexes to tables, as well as indexes to a relational cube to speed multidimensional analysis. Because of Broadbase's ability to view external sources as relational tables, installing the software initially and using it to help design the solution significantly reduced the overall time to implement. The amazing thing was that we had previously spent nearly six months attempting to link the Oracle and Teradata systems to provide cross-functional analysis. Broadbase accomplished this in hours.

Broadbase provides complete data mart functionality including Brio for Broadbase as the analysis tool. Two additional tools are being used to access data from Broadbase: Tango (a Web front-end development tool) and Andyne GQL. Broadbase is running on a Compaq NT server. At this point we were able to perform product sales analysis, slicing and dicing the data by channel, customer, product and time (rolled up to a monthly level). Two views were created to perform the initial analysis, which included YTD97 versus YTD96, [month] 97 versus [month] 96, and [week] 97 versus [week] 96.

These types of comparisons over time were not previously possible and gave us remarkable insight into our business, as well as validation of activity that we previously believed to be true. The data warehouse has given us the ability to have a small number of power users conduct complex analyses. The Broadbase data mart solution gives a greater number of our knowledge workers direct insight to the business when they need it--now.

There were a lot of advantages that Broadbase brought to the table, but when we look at what we were able to do in such a short period of time, three stand out.

Immediate Value: By being able to connect to our existing warehouse systems, view the source data and then load the required tables all in a matter of hours, Broadbase proved that their product really did support rapid deployment and would help to demonstrate the true value of the solution. The ability to access our sources quickly and easily demonstrated what a compelling solution this was.

Simplified Management: In a division such as IV Systems, there are a limited number of IS resources available for supporting new technologies. Broadbase realized that this is often the case and designed a product that is easy to administer and manage. My existing relational databases require highly skilled DBAs for administration, while a technical business user, such as a data administrator, has the ability to administer Broadbase.

Extensible Analysis: Broadbase was able to quickly have a channel management application up and running that was able to answer our top queries. But I am also aware that our business will change again and that we need a solution that will answer tomorrow's problems, not just today's. Fortunately, Broadbase has provided this as well. The componentized architecture lends itself to quickly developing new applications, whether in IV Systems or another division. And, the available types of analysis do not end with simple query and reporting. Broadbase's integrated cubing technology allows multidimensional analysis to be optimized, and advanced data mining functionality will provide us with some of the predictive capabilities that we will require as this industry becomes even more competitive. We were able to connect multiple existing tools to Broadbase quickly and easily, even developing a Web-based application utilizing a standard Web development tool, which allowed us to administer a Broadbase multidimensional cube via the Web--all in one hour.

Focus on the Business

It is important to make sure that you don't lose focus on the users and their requirements. Just because data is available, doesn't mean that you can get that data out to the users when and how they need it. It is important that the solution architecture allows you to have your data in the optimal location. The Broadbase architecture allowed for loading the data that was required for our application, while leaving the other data on the enterprise warehouse. Broadbase's drill-through capabilities allow us to access that data when it's required, but limits the load process to exactly the data that is required for everyday analysis.

The focus for a decision support solution needs to be the end user and the analysis that they require. When you are forced to spend the majority of your time having to integrate multiple technologies together, and then spend even more time making sure they continue to work together effectively, that is all time and effort that could have been used to provide additional functionality to the users. When your business is moving and changing as rapidly as the health care industry, it is imperative that you are able to provide value in a timely manner.

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