Manufacturing data comes in many shapes and sizes, from big data sources to unstructured documents. To be competitive and operate efficiently, manufacturers must use all the data at their disposal to improve their business processes. One-quarter of manufacturers report that they are working with more than 20 different data sources. These sources include everything from accounting and financial data to internet of things data. Bringing these varied data sources together from internal sources and even across the supply chain requires technology that can access and prepare the data for both operational and analytical processes.
Our research finds that nearly three-quarters (73%) of manufacturing organizations that use data preparation technologies say they have improved their operations. Data preparation technologies enable line of business personnel to access, combine and prepare data. Manufacturers’ most common data preparation requirements are to join data sources and to make the tasks available for reuse. Organizations report that the most important benefit of data preparation is improved quality of information, which is critical: Nearly half of manufacturers (48%) report spending significant amounts of time preparing internet of things event data for analysis. Also included in the top five benefits are meeting analytic needs more easily and eliminating manual processes.
Data preparation is often used in conjunction with predictive analytics. More than half (53%) of manufacturers report that they require data preparation capabilities to support data science and predictive analytics and nearly half (48%) report that machine learning is important to their internet of things initiatives. Many of these analyses require that data be prepared and formatted in particular ways. For instance, a continuous range of values may need to be normalized as values between zero and one. In other cases, data must be “binned” or “bucketed” into several groupings of values, such as equal quartiles of values. Data preparation tools allow data scientists to easily manipulate the data into the format needed by various algorithms and therefore try more alternatives to produce more accurate models.
Data preparation tools can also provide a collaborative framework in which knowledge of the data sources and its uses is shared across an organization, increasing its usefulness to all involved. In fact, more than four out of five (82%) manufacturers said they consider collaboration in these processes to be important. Collaborative capabilities allow participants to “like” or rate different data sources and comment on the ways they are using the data. This shared knowledge creates more trust in the data, results in higher data quality and makes the data more valuable to the organization.
While data preparation technology can provide these many benefits, organizations face barriers to improving their data preparation processes. The most basic issue is that as our research shows, many organizations lack awareness that these capabilities exist. In addition, more than one-fourth report inadequate skills in their organization and many manufacturers (38%) said they consider their current data preparation technology inadequate or somewhat inadequate. Reflecting this dissatisfaction, more than two-thirds (68%) of organizations said they are planning to change the way they assess and select data preparation technology within the next 12 to 18 months.
Analysis of manufacturing process data can be used to improve data quality, delivery timeliness, customer satisfaction and other aspects of the manufacturing process. Often manufacturers work with data from many independent systems including external systems. Our research shows that manufacturers can benefit from using data preparation technology to access and combine these various data sources. Organizations should consider how they can use or improve their data preparation technology and processes to achieve these benefits.