What to Do When Your Data Sucks

Let me be honest. For many companies, their first reaction after setting up their first data mastering run is a let-down.

It’s great you got so much work done in one afternoon! What’s not so great? The data. It’s bad. 

You knew it was going to be bad… but not this bad. Huge swaths of attributes are missing. Someone appended “do not use” to the name field in 50% of records. Some random third-party source created hundreds of thousands of contacts with no email addresses, and hundreds of thousands of records have no names at all. 

It’s especially painful because you had a vision: You’d bring your data together. You’d start powering amazing dashboards and reports. And you’d find millions of dollars of upsell opportunities. You’d have the best contact details available in your systems as soon as they were updated, and you’d start onboarding customers in a fraction of the time. 

But sometimes reality hurts.

You can’t show this stuff to the business. It’s not usable…at least not yet. They say garbage in, garbage out. But you probably didn’t expect that meant running a waste management operation. 

It’s not good news. But the almost-good news is that you are not alone. No matter how bad you think your data is when you consolidate it in one place, trust me, we’ve seen worse. We understand that for many, this part of the journey seems impossible to overcome. Stopping now seems like a viable option. But doing so is actually the worst possible option.

The True Costs of Bad Data

That messy data is costing your business: missed revenue opportunities, incorrect regulatory and investor reporting, and bad customer experiences. Unless someone takes responsibility for cleaning up the mess, everyone will continue along in a steady state, pretending the organizational friction of bad data doesn’t exist.

The first step, and the most relieving one, is to take a deep breath and reframe your thinking: This is not just an overwhelming mess, it’s the beginning of a structured data cleaning process. 

You are not responsible for the state of the data. And you are not responsible for your colleagues’ reactions to years or decades of bad data management coming to light for the first time. As the saying goes, you can’t manage what you can’t measure. And even if the measurements are cringeworthy, it’s the first step toward actually getting value from your data (even if that process is going to take longer than you thought). 

Take a Deep Breath—and Take Action

Start by assessing your data. Think about it as a corpus, a corporate asset that can be managed, measured, and improved. Data quality improvements can be defined as KPIs. 

Without an assessment phase, a conversation for third-party enrichment might look like this:

Sales Team: “We should buy Dun and Bradstreet. It will make our data not suck.”
Budget owner: “Sounds great. How will it make it better?”
Sales Team: “Our data is really bad, so it will make it less bad.” 
Budget owner: “How much is it?”
Sales Team: “A whole lot. But our data sucks a whole lot.”
Budget owner: “Ok… what products will you need?”
Sales Team: “Probably all of them.”

But if you have a way to assess your data as a corpus, the conversation may go differently:

Sales Team: “We should buy Dun and Bradstreet. It will make our data not suck.”
Budget Owner: “Sounds great. How will it make it better?”
Sales Data Team: “Third-party enrichment should improve our URL attribute completeness and corporate address field completeness rates by 30%. It should also improve nearly all of our SIC codes. We’d expect customer matching to improve a bit less than you might think—perhaps 5-10% increase in compression rates. But we’d probably add about 200,000 parent-child relationships across our 1.2 million accounts, which means we’d have 200,000 or so new accounts in Salesforce.” 
Budget Owner: “Sounds great. What does any of that mean?”
Sales Data Team: “It should help us develop sales territories in about half of the time since we’d be able to quickly verify parent-child accounts even if some of the subsidiaries have addresses in different geographies. We should also expect to see about 15% more of our inbound prospects being matched to accounts that would roll up to existing customer accounts.”
Sales Team: “Sales territory development takes us all of Q4 to figure out—and it diverts our attention from closing out opportunities for the year. Our sales reps spend way too much time calling prospects who turn out to be customers.”
Budget Owner: “I’m in! What products do you need?”
Sales Team: “Probably all of them.”
Sales Data Team: “I think if we start with the Level 1 and Level 2 data blocks with hierarchies we’ll be able to get those fields we want on the accounts and get hierarchy information. We don’t need the credit risk product which is where it starts to get really expensive.”
Budget Owner: “Do that.” 

The Master Data Management Journey

These kinds of conversations are where the lightbulb goes off: Messy data isn’t just a headache, it's a solvable business problem. And to solve it, you need a systematic way to measure, prioritize, and improve. That's where master data management (MDM) comes in.

MDM includes the processes, technology, and governance practices that enable organizations to deliver clean, consistent, trustworthy data that everyone can use to make better decisions. Following the MDM journey helps you to assess your data’s quality, measure and manage it, create a framework to evaluate and execute initiatives to improve it, and review it with the people who use it most. That way, you can confidently answer questions like:

• What’s the value of a good email address for our marketing campaigns? 
• What’s the incremental benefit of a 10% improvement in prospect-customer matching process? 
• How will updating 60% of industry fields with third-party data improve our go-to-market motions?

For most organizations, the first goal isn’t to frame their initiatives in these clear ways but, rather, to get to a place where they can see what gaps they have—so they can start thinking about their data in a strategic way. 

In that context, the painful, yet necessary, first step of the journey is to confront your data as a corpus and consider the steps needed to improve it. Do this, and you’re on your way to having data that not only doesn’t suck, it actually powers important, quantifiable business benefits.