Information Lifecycle Management

Jake Freivald (Europe) - How Does your Data Perform for you?

Information Builders recently commissioned a survey of more than 700 senior IT decision-makers in the U.S., Australia, and Europe to see the impact of information management on everyday decision-making.

We already knew, anecdotally, that data quality becomes increasingly important as information architectures get larger, more diverse, and more complex. Timely, high-quality data streamlines business processes and helps decision-makers at all levels make choices quickly and with confidence. In other words, companies do better when they make sure their data improves their business performance.

Improving an asset's performance means improving the way it's managed, so we expected to see interest in improved information management.

The survey validates these beliefs and expectations, and shows what these companies think about information management in their enterprises. I'd like to share some of the findings from this timely research and then offer some insights about data quality improvements.

Global and Regional Anecdotes
There were distinct regional differences in the responses, but some universal beliefs emerged as well: High-quality data provides a significant commercial impact; spreadsheets rule the office reporting landscape; IT managers battle issues caused by disjointed reporting.

About half of the respondents indicated that poor quality data leads to higher costs, lower productivity, lower customer retention, or loss of revenue. As a result, 53% of the organizations surveyed have already implemented or are actively implementing a master data strategy and data governance strategy.

Most of the organizations in this study depend on multiple data sources for their business intelligence projects, although this varies by region. The UK averages eight sources, while Germany and Spain average only five. UK-based organizations are the least likely to have integrated these data sources.

The way companies organize decision-making data also varies widely. Just look at some of the differences in Europe: Spanish firms deploy double the number of data marts as Dutch ones (73% versus 31%). In addition, respondents from Spain reported a relatively high incidence of real-time access to data (57%).
Another clear signal emerging from the survey is that data quality issues will need to support the increasing need for real-time and mobile access, both of which are increasing rapidly.

Mitigating Data Chaos

Successful data governance depends on applying consistent data-quality procedures to every information source across your business, both internal and external. (Avoid the historically popular but untenable belief that data cleansing is something that happens on the way to the data warehouse.) Some of the needed data-quality management methods include the following:

Profiling is the process of gathering statistics about enterprise data. What are its primary characteristics and attributes? Which users access it most frequently? Does it have out-of-bounds data values? Profiling data helps you identify data quality problems and prioritize cleansing activities.

Cleansing alters existing data based on pre-defined business rules and criteria. The cleansing process appends missing entries, corrects data errors, and automatically standardizes data to specific formats.

Merging and matching automatically uncovers related entries within the same system - or across multiple systems - and makes them consistent.

Some organizations rely on scoring to more effectively evaluate data quality and prioritize problems. They assign a number to every data record, providing insight into its lineage and quality. Not all data should be treated equally; for example, customer information should be scored more strictly than data about office-supply inventory.

Automating the Process

The earlier you catch data quality problems, the easier and more cost-effective it will be to solve them. That means that your data quality process (including people, policies, and technology) should be a consistent feature of every application, feed, or system where relevant data is created or imported. Ideally, your process will help you:

• Cleanse and reconcile data across all relevant transactional and analytical     applications
• Improve and enrich customer information coming from all sales and customer support channels
• Validate and correct incomplete records within customer profiles
• Ensure quality throughout software integration projects
• Validate data input via online self-service applications
• Cleanse and unify data during system migrations.

This is clearly bigger than a one-time project or a simple process: It's a complete program. Treat it like one. Set goals based on how your data needs to perform for you, identify metrics on what you want to achieve, and then manage it for the long haul. If you want your data to perform for you, you need to be among the 53% of companies that are engaged in data quality and master data management initiatives, and ultimately you should measure your success on how well you compete against the other 47%.

By Jake Freivald, vice president of marketing, Information Builders




« Sean A. Hays (North America) - What is an Exponential Technology?


Sean Dolan (Europe) - Next Generation Security for the New Generation of Cybercrime »


Do you think your smartphone is making you a workaholic?