Most IT and business executives believe that big data has the potential to create a significant business advantage, but few are using it. We discuss some of the barriers to big data analytics and how they can be overcome.
A recent survey by pwc has shown that one in five IT and business executives has a big data strategy for their company, and one in four is very or somewhat likely to create one in the future. Given the power of new data analytics technologies, these low numbers may point to some limiting misconceptions. Most companies are aware of the benefits of big data analytics: analyzing aggregated consumer data, such as web search and navigation patterns, provides companies with much faster and more accurate information on consumer needs and interests than previous methods such as forecasting software. New techniques for analyzing big data rapidly signal companies to new market opportunities, changes in consumer demands, and consumer responses to marketing campaigns. Companies that utilize big data are able to adapt to emerging and changing needs and invest in the most promising areas well ahead of their competitors.
Why, then, are executives reluctant to adopt big data analytics? A couple of forces may be in play here. One major barrier to adoption is the fear of privacy invasion and discomfort with the idea of collecting consumers’ personal data. The other main issue is adoption itself: data analytics technologies are complex and their implementation requires cooperation between different departments with different priorities. Realizing the benefits of big data analytics requires companies to overcome these barriers and get big data strategy right. Four key elements of an effective big data strategy are:
Big Data Bootstrapping
Big data increases a company’s data collection and analytics capacity in three ways: volume, velocity, and variety. Trying to increase all of these at once will result in disorganization. Companies seeking to utilize big data effectively should allow staff time to adapt to new technologies and processes before adding new data sources and analytics features.
Executives are rightly concerned about the implications of big data for privacy. The key here is de-identification – personal identifiers should be removed from data as soon as possible in the analytics process. Privacy should be built into the design of technologies and business processes, and de-identification practices should be above average for your industry sector.
Information security practices remain essentially the same as before the adoption of big data analytics. However, organizations need to develop effective methods for evaluating the level of residual privacy risk posed by de-identified data.
People who lead the implementation of data analytics should be creative leaders who can think outside of the box.
As mentioned, the implementation of big data analytics is a complex project requiring cooperation between different departments. This can become a major obstacle when competing priorities surface. IT professionals tend to see data analytics as a tool: from their perspective, their job is simply to install and maintain the software. Business and marketing will be most interested in working out how to integrate data analytics into the business process, and will usually want to utilize as many new features as possible. Finance will also want to track and calculate numerous variables. Trying to meet all of these demands can result in companies attempting to implement new data analytics technologies faster than the business can adapt. It simply is not possible to fully utilize big data all at once. We recommend a staged implementation process designed to introduce people to a new data analytics environment at a realistic pace:
Implement the tool without making any other changes. Install the new software and give staff time to learn to use it for the same tasks performed by previous technologies (e.g., forecasting software): tracking sales, complaints, number of users, etc.
Implement a new business process adapted to the capabilities of the new software. New data analytics software is more user-friendly than earlier technologies, and this is the time to transition to self-serve analytics for all staff. Data analysts will transition to a training and advisory role, teaching other staff to interpret basic trends while they themselves perform more advanced trend interpretation and comparison.
The company is now in a position to take full advantage of the capabilities of new data analytics technology. First increase the volume of data by adding new data sources (e.g., more web analytics). Then increase the velocity of data by collecting data more frequently (e.g., monthly instead of quarterly). Then increase the variety of data by collecting new types of information (e.g., demographic factors to be linked to consumer behaviour).
Utilizing big data is a complex project, but one well worthwhile for companies looking for faster and fuller consumer feedback. Collecting and analyzing big data enables companies to pick up on new trends more quickly and accurately than ever before, see further ahead into the future, and invest in the most promising areas.