Business Intelligence

Angels and Devils: 6 Data Quality Problems That Drive Your Business Down

Oct 30, 2018

Data quality problems are the bane of today's business leaders. Learn about these 6 data quality issues that prevent businesses from getting value from business intelligence (BI) and data analytics.

Thomas LaMonteContent Analyst
Angels and Devils: 6 Data Quality Problems That Drive Your Business Down

What we'll cover

Has your business gone to a fiery inferno? Not yet? Let’s keep it that way. I’m about to conjure up advice on one business pain point you might be agonizing over: data quality problems.

Today’s companies are vying to become data-driven using business intelligence (BI) software to analyze data insights to make better decisions. According to Forbes Insights, 84 percent of CEOs are concerned about the quality of the data they’re basing their decisions on. When the lights dim on your data integrity, who knows what records or variables have been tampered with, duplicated, deleted, or forgotten about.

Bad data leads to bad decisions.

Small business leaders need to ensure the data they are basing their decisions on is accurate, legitimate, and pure, or risk some $15 million per year in damages caused by poor data quality (as measured by Gartner). To avoid anguish caused by data quality problems, two spirits have volunteered to sit atop your shoulders to give some much needed guidance. A devil on one side; an angel on the other:

6 data quality problems:

  • Incomplete or inconsistent data

  • Regulatory noncompliance

  • Lack of access

  • Duplicate and obsolete

  • Security breaches and exposure

  • Systems upgrade

1. Incomplete or inconsistent data

Data entry is often ground zero for mistakes that diminish data quality. When data is not submitted to systems correctly (e.g., misfiled customer forms, unclear fields, human error) the information record leftover may have absent variables.

Imagine the difficulty of locating a customer’s email address if the domain (e.g., @gmail) were incomplete. In addition, inconsistent formatting of data entries (e.g., mm/yyyy vs yyyy/mm) may cause data to be misinterpreted by key systems, processes, and cross-functional teams.

Recommendation: Marketing leaders should work closely with the IT department to align business priorities when creating customer forms, selecting fields or formats, and establishing data capture standards.

2. Regulatory noncompliance

Who has permission to access a particular data asset, and how should you store it? What are the guidelines for managing this data through its life cycle? These are key questions that, depending on jurisdiction or industry, require answers in accordance with data privacy laws such as HIPAA or the GDPR.

Noncompliant data is emblematic of poor quality. Related problems can boil over into bigger problems, such as levied fines and irrevocably damaged customer trust caused by a haphazard data governance strategy.

Recommendation: To shore up data governance in your small business, begin by revisiting-or drafting fresh-data governance policies with all key stakeholders. Ensure there are clear guidelines and procedures to determine if collected data is permitted for the intended use under relevant regulations. Also, ensure that clear methods exist outlining how to notify customers (or other data subjects) in the event of changes or security events.

3. Lack of access

Sharing data access beyond the IT department or data scientist is a problem for many small businesses. Distributing information often devolves into a game of telephone: each new recipient degrades the quality of the message, or in the case of shared information, a bit more data is shaved off at each juncture. Often this negatively affects data quality, as teams are interacting with outdated or varying versions of the same data-never able to view the full picture.

Recommendation: Self-service BI tools may be an opportunity to equip even non-technical users with functional data analytics tools and keep data assets in view of all teams.

4. Duplicate and obsolete

Data is almost never at rest. It moves from place to place, passing hands a dozen times. Duplicated or outdated data is inevitable in the shuffle. Erroneous copies of the same data asset can muddle up search and distort data analysis. Moreover, since it very easily evades cursory detection, outdated data badly misleads unsuspecting data scientists. Together, these data quality problems hollow out an otherwise viable source of data for BI and data analytics.

Recommendation: Even if your data-management processes are mostly automated, scripts or algorithms sometimes get it wrong; for best results, undergo a dual effort of human and computer-driven deduplication.

5. Security breaches and exposure

An obvious threat to data quality is a security breach via an overt cyberattack by a hacker. In 2016, Verizon found that 61 percent of small businesses suffered a security breach. Expectedly, data that is exposed could be corrupted, altered, or compromised with malware or viruses without your knowledge.

But what often surprises business leaders is that internal hires are often an immediate risk: More than 90 percent of all cyberattacks are committed with information stolen from employees who unwittingly gave away sensitive data.

Recommendation: Many organizations unintentionally hire their own hacker, but often it’s ignorance, not disgruntlement, that leads to a data breach. Work to train all staff in the basics of security common sense, password best practices, and data backup procedures.

6. Systems upgrade

An overlooked threat to data quality is a systems upgrade. A key system and source of data may be uprooted and replaced, causing an integration crisis for legacy content. Or perhaps your business’s primary data management system, file structure, or format requires a switch. Migrating or restoring data needs to be carried out with a reliably repeatable strategy and from verified sources.

Recommendation: Be methodical: place major effort into the planning of data integration before a systems upgrade. In addition, create and rely on emergency plans in the event of a catastrophe, security breach, or other unexpected event mandating data restoration.

Recommended resources:

About the author

Thomas LaMonte

Content Analyst