We asked more than 1,000 Salesforce customers, “What keeps you up at night?”

Unsurprisingly, the overwhelming response was — data . This encouraged me to start a blog series on data and explore ways to maximise its value. First things first: let’s start by making sure that your data is clean and accurate.

Step 1: Analyse your existing data.

Determine what data you are going to look at, where it came from, and whether it has any special characteristics. You may already be familiar with the data, but consolidating everything we are going to work on in one place will help us document the details of this project as a resource for future data efforts.

• What data are we going to look at?
  
  Answering this question requires knowing the desired outcome of the project. For example, Sales wants its Salesforce integrated click to call dialer to work on every lead no matter the source, so we’ll focus on the phone field in Salesforce and any data sources that populate that field.
• Where did the data come from?
  
  We need to determine all the different ways our telephone field can be populated. Can salespeople edit the field? Do we collect phone numbers from web forms or during live events? Knowing this will help us control how this field is modified in the future, and help prevent new bad data being added. 
• Does this data have any special characteristics?
  
  Our click to call dialer works with several phone number formats, but getting our current data into those formats will take work. Knowing both the accepted and current formats of the data will help us determine how much work we have ahead of us, and if we need to use any specific tools.

Step 2: Create a cleanse plan.

Once we are clear about the state of our data and what we are trying to achieve, it is time to create the project plan. Look for ways to break it down into different stages. 

• What data needs to be cleansed, removed, or enriched?
  
  Because we’ve identified all of the data we are going to work with and where it came from, we can now decide how to address its deficiencies. 
• Which business processes are causing data quality issues?
  
  Each must be fixed before bringing clean data into production — otherwise, all that data cleaning and enriching will have to be repeated. It could be as simple as placing a validation rule on the phone field that salespeople can edit.  
• What tools do we need to execute the cleanse?
  
  There are many different kinds of tools: data loaders, cleaners, phone and address validators. Most importantly, as most of the tools cost money, include the costs in the project budget.

Step 3: Define a Governance Strategy.

Before starting the cleanse, a strategy for keeping the data clean must be in place.

• Have all the business processes been redesigned?
  
  ...And are there checks in place to keep future changes to those processes from adding bad data? Most importantly, as changes are made to processes, determine if we able to see their potential impact on data. 
• Has data validation been added at system entry points?
  
  Validation should be built in at the user interface, bulk upload, and integration points. Data that enters clean is likely to stay that way.
• Does the data have an owner, and are they incentivized to keep it clean?
  
  If an individual understands how bad data makes their job harder, they are more likely to be a champion of data quality. Find a data steward in every department, and make them part of your governance strategy.

Now that your data is clean, what are you going to do with it? How will you keep it clean, and use it to gain more value from your CRM investment? Tune in for subsequent installments in this blog series as we go deeper into data.