How to Conduct a Data Audit
Posted September 7, 2022
How to Conduct a Data Audit
Posted September 7, 2022
Data. The backbone for strategies and campaigns. For your marketing to be data-driven, your data must be validated. Your single source of truth identified.
Ben Grossman, Chief Strategy Officer at Doner Partners Network, recently spoke about the importance of starting a project with a data audit. This is when you ask yourself what data you already have versus what data you need.
What is a Data Audit?
A data audit is a series of processes surrounding the evaluation of data quality, data security, and data compliance compared to the use-case of that data set and what components the auditor is looking to solve for. Audits can either be 1st Party audits or 3rd Party audits.
A 1st Party Audit is an internal, do-it-yourself solution to your data problem. You can launch all processes with your team internally.
A 3rd Party Audit is an external audit by, you guessed it, a 3rd Party. This is the direction you should take if the amount of data you have is overwhelming and would overtake your workload.
Formal vs. Informal Audits
Audits can be formal or informal.
The formal ones are planned and should be done at consistent intervals.
Informal audits are more likely to come up when your data doesn’t line up, something is not right with the results of your activities, and you need to validate your information. This is common in demand gen activities.
How to Complete an Audit
To complete a data quality audit, there are a few steps to take. You’ll start at the site level (information that spans your company), then move to the contact level (information in individual records) and then on to the detail level (information inside individual records). Think about from this perspective, your data is being organized by the largest definable points, then moving to the most specific.
In general, your data is taken into the data processing/analysis team, and it is standardized via an address standardization software. Different software modules exist for each country to resolve each country’s standardization protocols, so your software could be different based on your data’s country origin.
Next, your data is checked against the National Change of Address database (NCOA) to flag companies and/or people that have moved from a particular location (these records are removed and bucketed). The database is also Delivery Point Validated (DPV) to code the record to the USPS delivery-point level.
Then, if necessary, the clean and accurate database is put through subsequent validation processes which can include SuiteLink to determine/add missing suite #s, email verification, and phone validation. Confidence scores can also be added to various firmographics and demographics. That was a giant overview, now let’s dive into the specifics.
Step 1: Standardize your data
This is basic hygiene of your data at the site level. Work with your internal team(s) to look at all the places your data lives. Then, compare how the data is stored in those locations. Your ultimate goal is to get all of your data in the same format. This applies to both layout (e.g. the order the fields are listed) and inputs (e.g. using NY vs. New York).
Standardization is the necessary first step in any audit. You need your data to be standardized before you check and/or augment data sets.
Step 2: Verify company addresses
This applies to both physical and website addresses.
For physical addresses, use a national change of address software to verify the information in your database. Note that you’ll want your addresses in the standard USPS format before you do this.
Even if you don’t send physical mail, this step is important for location targeting. If a company moves from California to Florida, you’ll want that company to be included when you pull a list of Florida prospects to invite to your event.
For web addresses, you have three main options. You can either do a manual check (type the url into your browser) or you can write a program (such as SQL) or hire a vendor to do this for you.
Though most companies don’t change their web addresses, it does happen. Companies rebrand (we did!), and companies are acquired. A quick check at the start of your audit can save you time and money down the road.
Step 3: Verify delivery point
Now that you have your addresses verified, use delivery point validation (DPV) software, such as CASS, to validate postal information down to the postal delivery code and the walk sequence code. This is important as some of your customers might have different physical locations.
Think about when physical mail is delivered to your house. The mailman has the mail in order inside their vehicle before they arrive at your street, and they are not organizing the mail inside the truck when they pull up. Delivery point validation software is the same when dealing with high volumes of data.
The more avenues that you have to corroborate accuracy the better.
Step 4: Verify the email
Now you have reached the contact level. You want to know the data you have on each individual person is accurate, and the email address is the best place to start.
You’ll probably want to work with an external vendor (like Anteriad) to verify your email addresses. Note which ones are deliverable, undeliverable, or spam. Be sure to also verify that each address is associated with the individual assigned to that email address.
Step 5: Verify the phone
This is another place where it’s beneficial to work with a vendor.
This step is important to check not only the validity of phone numbers but to also check the relation of the phone number to the person, business name and address. You could be trying to reach an individual who left the company.
As our market has changed over the past few years, verifying phone numbers is another way to validate the contact record you have is reliable.
Step 6: Assign confidence scores
This is one of the final steps in your data audit. These are some of the specific details that can make or break your marketing activities. Assign confidence scores on your firmographic and technographic information. In other words, how confident are you that this data is correct?
This applies to title, function, annual revenue, number of employees, and industry codes, as a few examples.
Other things to note
When conducting your audit, you will start to see pieces of info you need to action on later. Examples can be:
- When did you last engage with these contacts? When did they last buy from you?
- It’s time to de-dupe your records. Maybe you have unknown overlap between virtual and in-person events. Maybe you have merged with another company and are now handling two databases of customers and prospects. Be sure to clean your duplicate records.
When to audit
Now that you know how to conduct a data quality audit, how often should this take place? For the site level, this should be done monthly, but many teams start quarterly. Contact level data should also be audited monthly, and your detail level audit should be done every quarter.
If you need help conducting your data audit, we are here to help! Anteriad has built a technology that makes sure the URLs open to the correct website and the business name, address, and phone numbers corroborate with the correct information.
To learn more about why up-to-date data matters, check out:
1. How Data Can Be A Marketers Secret Weapon
2. The Marketer’s Guide to Customer Data Platforms (CDPs)
3. How to Find Creativity in Your Data