A Two-Part Strategy to Weed Out Bad Data
Weeding out bad data can be one of the most practical ways for companies to improve their bottom line. As one striking example, a 2009 Gartner survey of 140 companies revealed that missed opportunities and inefficiencies due to “bad data” was costing each business more than $8 million per year – 4 percent of respondents estimated the annual damage to be as much as $100 million while 25 percent reported bad data costs of $20 million or more annually.
Does your company have bad data? In this article, DataEntryOutsourced provides an overview of what typically causes the problem and what you can do about it.
Categories of Bad Data
Bad data comes in many forms – for example, non-connected data sets for relational data, corrupted files, inaccurate data fields and duplicate files. Despite the multiple variations of bad data, here are the five primary categories:
- Non-compliant Data – Data doesn’t follow your company’s naming rules.
- Missing Data – Empty data fields where there should be data.
- Irrelevant Data – Data entered in the wrong field.
- Inaccurate Data – Incorrect data (including data not updated properly).
- Duplicate Data – One contact appears in multiple database records.
The High Cost of Bad Data
Whatever form bad data takes, it is probably costing your company some serious damage in the form of inefficiencies, damage to reputation, missed opportunities, reduced workforce morale, shoddy customer service, inaccurate forecasting and bad business decisions – as noted in the first paragraph of this article, one study reported annual costs ranging from $8 million to $100 million.
In more recent research by Experian Data Quality, 88 percent of companies reported a direct impact on the bottom line – on average, each company lost 12 percent of revenues due to bad data. Revenue losses were typically due to excessive staff time, wasted resources and unnecessary marketing expenditures.
Based on the 1-10-100 quality principle, costs of fixing problems can increase exponentially over time. For example, if $1 is the initial cost to prevent bad data from impacting your CRM system, then $10 is the cost to correct an existing data problem and $100 is the cost to fix a bad data problem after it causes a failure with either the customer or the company.
Adopting a Two-Part Strategy to Cure Bad Data
Dealing with bad data effectively requires a two-front operation for data entry accuracy:
- Prevention – Controlling and minimizing bad data “before” it gets into your system.
- Remediation – Monitoring and cleansing data “after” it enters the system, with a goal of maintaining prescribed quality standards.
To prevent bad data, start with user training that addresses issues such as searching for duplicate entries before entering data and completing all data fields. To remediate existing data, start by defining a data quality process and then monitor the databases based on those standards.
A few words of caution – data cleansing can be a monotonous process. To avoid overburdening your employees, think about practical alternatives such as outsourcing to a data cleansing expert such as DataEntryOutsourced.
Find a Data Cleansing Expert
In an era of “Big Data” and increasing data-sensitive processes, the potential damage due to bad data is virtually unlimited. While no organization is immune to bad data, you can detect and correct bad data problems.
Data cleansing is an effective strategy that can dramatically improve your bottom line by transforming ambiguous data into consistent data sets. Equally important is data validation at the “point of entry” – think of it as checking data passports before the data is allowed to enter your system.
DataEntryOutsourced is a recognized global expert that can help with your two-front operation to weed out bad data – data cleansing and data validation are two of DEO’s primary daily duties on a 24/6 basis.
What are your experiences with bad data? Please share them below with a comment and then pass along your thoughts by using the social media icons.
– DataEntryOutsourced
