Intelligent Enterprise

Better Insight for Business Decisions

RSS
Webcasts
Whitepapers
Subscribe
Home

November 18, 2003

Embracing Excel

Microsoft Excel is a much-beloved desktop tool for business intelligence, yet many people want to contain its use or replace it

by Philip Russom

Millions of business users rely daily on Microsoft Excel for business intelligence (BI) functions such as data manipulation, data visualization, presentation graphics, forecasting, and so on. For many users, Excel spreadsheets are a form of reporting, and Excel pivot tables a type of data analysis. In fact, "spreadsheet reporting" — where a client-side spreadsheet application such as Excel serves as a report viewer — has been with us for decades and continues to slowly grow in popularity. Microsoft claims 150-million business users on Excel. Even if a mere 1 percent of them use it for BI functions, then that makes Excel a very popular BI tool.

Assuming all that's true, why are so many people eager to limit the use of Excel for BI or replace it with a dedicated BI tool?

Spawning Spreadsheet Silos

Much of the resistance to Excel as a BI tool stems from a common "worst practice." Users with access to a BI platform sometimes export data from it into Excel. Once the data is in Excel, the user manipulates it in various ways to create new calculated values, reorder the data, and add visual interest through formatting and charts. On the one hand, the business user is able to add value to the data and do so productively, thanks to Excel. On the other hand, the results may alter the data and its presentation so it bears little or no resemblance to data in the BI platform that other users rely on.

When multiple users bring Excel-formatted sets of data drawn from a BI platform to a meeting, the result is "dueling spreadsheets," where the content and presentation of spreadsheets differ strongly, although their data was drawn from the same source. Avoiding dueling spreadsheets is one reason corporations build data warehouses as the "single version of the truth," and why they create reports for a repeatable presentation of that truth. In fact, managers in a few companies have tried to forbid Excel, exhorting business users to stick to reports run on the BI platform, which are preferable to the dueling spreadsheets that derail meetings and stymie decision-making.

Furthermore, when users export data to Excel, that data set becomes a "spreadsheet silo." In other words, it's spinning on a user's local hard drive, disconnected from enterprise data, which is a moving target. As enterprise data changes, the user must re-export data constantly to keep up to date — a manual process that is time-consuming and loaded with opportunities for error. For this reason, PC-based spreadsheet silos typically suffer from data quality and data integrity problems.

Then there's the problem of collaboration. Exporting to Excel is largely a one-way trip, in that the user can't close the loop by returning the value added in Excel to the BI platform where all colleagues can access it. So, users often send an email with a spreadsheet silo attached to colleagues who also manipulate the data, further exacerbating the problem of proliferated spreadsheets as islands of information that are impossible to contain or maintain.

A common attitude is that a BI platform and its reports are good, if they halt the problems of dueling spreadsheets, proliferating silos, and uncontrolled distribution of sensitive corporate data. Replacing Excel with a dedicated BI tool has compelling benefits — except in one area. It reduces the productivity and creativity of business users who know Excel well and work in departments where spreadsheet reporting makes sense, such as finance, sales, and marketing. For these users, the ideal solution is to keep Excel as the client tool for certain BI functions, while tying it into a networked, collaborative BI platform. The catch is that this has been near impossible, until recently.