Why Build a Data Warehouse?

Everyone seems to have built one and those who haven’t want one. A rational person might question whether or not this is simply a fad – a buzz word to throw around in conversation – or if there truly is a business justification for the effort and expense involved in data warehousing. For small and medium-sized companies, a data warehouse and business intelligence solution can seem to be a daunting expense. While those concerns are not unwarranted, a successful data warehouse implementation can yield significant value.

1) Time Savings

A data warehouse can save your organization a significant amount of time. Transactional databases are not designed to handle the type of analysis-oriented queries that a data warehouse is capable of. Data warehouses are primarily driven by read operations (querying against the data) as opposed to being heavily write operation driven (inserting data in the database). A data warehouse, designed for read operations, will return data much faster than your existing transactional databases. The turn-around time to produce useful information is greatly lessened allowing your organization to make more decisions faster.

2) Centralization and Data Source Integration

Integrating various sources of data can yield incredible cross-functional insights into your business operations. A retail company might integrate their CRM system with their marketing and sales databases. A financial institution might look to integrate their loan origination system with their core processing system. Both instances represent examples where organizations can integrate two disparate data sources into one centralized location. Data warehouse design allows you to analyze several measures (housed in fact tables) like counts, amounts, and averages against several dimensions (like product, location, branch, etc.).

Think of this scenario and see if your organization has a similar task. Let us assume you house your customer sales data in a Microsoft Dynamics CRM system. Your marketing survey information might be housed in another database or some other third-party application. A data warehouse would enable your organization to query your sales data AND your marketing survey results at the same time for side-by-side comparison. Perhaps you wish to compare customer satisfaction ratings of your sales force side-by-side with their sales metrics for the current month. Through the centralization and data source integration a data warehouse provides, you are able to identify unique cross-functional insights into the health of your organization.

3) Historical Analysis

A data warehouse can be developed to enable historical trending and pattern recognition. Most CRM systems, for example, contain only the current location of a customer. If your organization is interested in the migration of customers over time, your data warehouse can be designed to handle that type of request. If product categories change over time, most often they are simply overwritten in the source system. If you would like to track what product category a particular product was in this year, versus five years ago, your warehouse can be developed to handle such queries.

With more businesses understanding the value of trending and modeling, the ability to leverage historical analysis can yield valuable insights into past and future outcomes.

4) Automate Data Extraction and Reporting Tasks

Manually extracted data sets and manually run reports are a common practice in source systems. Having those basic tasks automated can yield significant FTE savings with minimal upfront effort. Leveraging SQL Server Integration Services (SSIS) and SQL Server Reporting Services (SSRS) – both included in a standard Microsoft SQL Server installation – allows you to automate data extracts and report creation. As a complement to the aforementioned benefits of a data warehouse, automating these tasks can give a low-effort ROI boost.

5) Unburden I/IS Departments

Many organizations have an IT/IS group responsible for creating data extracts from the source system. A data warehouse takes that responsibility off of the shoulders of the this group. This allows the IT/IS group to focus on maintaining and improving the source system without handling tasks that the end users could perform themselves with a data warehouse.

These are only a few of the benefits a data warehouse can provide your organization. However, a data warehouse is not a “plug-and-play” system. It requires in-depth business analysis and business process research to understand how best to design the warehouse. It takes time to develop the ETL (extract, transform, and load) and validate the data in the warehouse. If the proper planning occurs, however, your organization will have a data warehouse that is built to last and provide a significant return on investment. Your successful data warehouse will become the platform for your organization’s business intelligence strategy.

For questions about data warehousing and business intelligence solutions, contact The Knowlton Group here.

(Photo Credit: tec_estromberg via Flickr)