BI in 1 Day—A Case Study

Getting started with business intelligence does not have to be daunting. The chances are you can get going in a day with tools that you already own. In this article, we will describe how a company can build a simple BI solution in one day. IT managers can use this “One Day Marvel” to show line managers the value of BI and to build some momentum behind a larger initiative.

We will use a hypothetical example of a small specialty manufacturer. The example is based on a KMA project. The manufacturer sells several hundred products through three channels to consumers in the US. They have an Enterprise Resource Planning (ERP) system that stores its information in a SQL Server database. The VP of Sales wanted to better understand sales by product and by channel to better allocate marketing spending. Here is how we provided the VP with key insights in a One-day engagement:

• The One Day Marvel involved a two-person KMA team—a senior business analyst with expertise in the business domain (e.g. marketing) and in creating and manipulating Excel PivotTables and a senior technical analyst with expertise in SQL Server Analysis Services.

• Upon arriving at the company, the team met with the VP of Sales to understand the key problems she faced. See the companion article (Helping Managers Tell Stories—Deriving ROI from BI).

• The team, led by the business analyst, identified the business problem and the data to be analyzed. The data was specified as a single measure—sales in dollars—and several dimensions—Products, Geography, Time and Customers. These dimensions were hierarchical. For example, several hundred customers rolled up into three channels. Approximate time: 2 hours.

• The team, led by the technical analyst, then worked with the company’s internal IT staff to obtain the needed data. The data was extracted from the SQL Server database using SQL Server Analysis Services and placed in a data cube. The technical analyst used tools provided by SQL server to identify the measures, dimensions, and the hierarchies of each dimension. The cube was then exported into Excel to be analyzed using PivotTables. Approximate time: 3 hours.

• The team, led by the business analyst, created a story about sales. They manipulated the cube, using PivotTables, to examine sales trends over the past three years. They then drilled down to sales by region, channel, product and customer. They created a PowerPoint presentation, using Knowledge Management Associates’ Mekko Graphics software, to share the insights with the VP Sales. Approximate time: 2 hours.

• The team then met with the VP Sales, shared the PowerPoint presentation and carried out some real-time analyses using the Excel PivotTables to drill down deeper into issues identified by the VP. Approximate time: 1 hour.

In one day, the VP Sales developed some new insights into factors that drive sales. More importantly, she had a story, backed up by facts, that she could share with the senior management team. Click here to view some sample slides from the presentation.

A One Day Marvel, like the hypothetical one described above, can be a valuable starting point for a BI initiative. At the end of the day, line managers understand what they can expect from BI, what data they need, and what resources they need to analyze the data and develop key insights (stories). IT managers learn how the data are being used, how to organize the data to deliver value, what tools are needed to analyze the data, and the level of effort required to build a system to provide BI data on a regular basis.