You know that you have dirty data. You know that there are valid reasons for cleansing it, such as a pending Salesforce implementation. Now all you need is to obtain approval for the project. A well-crafted business case can help you receive the authorizations you need to proceed. However, the first step is to understand thoroughly just what a business case really is.
Definition of a Business Case
A business case justifies a specific course of action. Many people interpret this in terms of “dollars and cents.” The financial benefit or return on investment is certainly one aspect of a solid business case, but other aspects should be included. Make sure that your business case also covers:
- The impact that dirty data is having on business users
- The potential risks that are posed by dirty data
- The intangible costs, such as loss of customer goodwill, that dirty data can cause
- The pros and cons of your recommended solution or vendor
- How the project will be approached
Including these details can help garner support for data cleansing. Often, management is unaware of how dirty data can affect employees, customers, and profitability. Your business plan may be the first that management has heard of the issue or the first view that includes specifics about the problems that are being created.
Tips for Crafting a Preliminary Business Case
Companies have a variety of reasons for needing to cleanse their data. Some perform the task as part of an extensive IT modernization effort. Others need clean data to personalize marketing campaigns, while still other enterprises fear they would soon find themselves out of business if they provide inaccurate data. Regardless of motive, the basic steps for crafting a business case justifying the project are the same.
- Search for missing data. Perform an analysis to determine the number of records missing one or more fields. Missing data can make it impossible to offer a personalized experience to your customers. For example, if a customer’s birthdate is missing, you cannot send her a coupon for a free item on her birthday or include her in a marketing campaign targeting a specific age group.
- Randomly select as many records as you or your staff can validate manually. Keep track of the dirty data and separate it from the newly cleaned data. You can use the information to estimate just how inaccurate your data is, but you can also run a selection procedure on both sets of data to demonstrate the inaccuracy of attempts at personalization.
- Audit data to locate records that are actually duplicates but contain slight variations. For example, you might have two customer records that are identical in all aspects except for the spelling of the first name, or the customer might have a different address in your master file than in your order file. If you regularly sell to businesses, you might discover the same individual listed as the purchasing manager on multiple customer records.
- Identify data that lacks uniformity. Perhaps your standard shipping weight is in pounds, but your inventory master contains records with weights recorded in kilograms. Perhaps you are planning to require shipping to use ZIP+4 for addresses, but you discover records in your customer database that lack this information
Quantify Your Findings
Now that you have an idea of just how dirty your data is, it is time to quantify the results. The first thing you must determine is how much of an effect the dirty data actually has. If the impact is minimal, you might prefer to invest your time and budget in other endeavors. However, if you feel you need to proceed with data cleansing, you will need to support your decision. Use all facts at your disposal to build a strong business case, such as:
- Unnecessary labor costs resulting from dirty data
- Inaccuracy of targeted or personalized marketing campaigns
- Lost sales revenue
This is the section where you should detail the financial impact that dirty data has on the enterprise. Intangibles should be listed elsewhere — reserve this section strictly for the monetary impact. This is also where you demonstrate the projected return on investment for the data cleansing.