Q&A: An Introduction to Self-Service Data Prep (Part 2 of 2)
Self-service analytics continues to grow across industries, but many organizations still struggle with implementation. Here's how to get started.
- By James E. Powell
- June 12, 2018
In Part 1 of our conversation with Paxata's Piet Loubser, we talked about the need for self-service data prep and the role it plays in the business. The need for data preparation is increasingly important because the data landscape is undergoing significant change. Organizations are struggling to find insight in the information generated from big data residing in on-premises and cloud data lakes.
Business users have long felt constrained by the lengthy turnaround time it takes for IT to complete BI projects, driving them to "take matters into their own hands" with a self-service approach. Here we look at how data prep supports the next wave of self-service, namely self-service analytics, and how enterprises can get started.
TDWI: Has self-service analytics been accepted as the status quo for businesses looking to manage big data and third-party assets or are we still working toward this?
Piet Loubser: I believe self-service analytics has become the accepted mode, but many organizations struggle with where and how to get there. Self-service analytics is not just the act of analyzing the data, but includes finding and preparing the data as well. I still see many one-off products and/or siloed implementations out there where self-service is managed in one tool at a time and where security and governance are not consistent across these tools. This creates major gaps in the overall trust and efficiency of these initiatives and quite often leads to compliance concerns.
What are the business and technology drivers of self-service analytics?
Businesses -- and every function within them -- are becoming data-driven. For instance, marketers are analyzing data from websites, email campaigns, and search optimization daily to better understand which leads to focus on. Supply chain teams are constantly vetting their purchasing behaviors to ensure the lowest possible inventory is supported by an optimal set of vendors.
Traditional approaches for analytics led to long cycles of analyzing requirements and building data sets to support those questions. Today's analysts want to explore all the data, not just a narrow subset. The reality is that there simply aren't enough available IT resources to support the growing demand. This leads to frustrated business users and long delays in getting the right information in the hands of decision makers. At its heart, the problem is a case of supply and demand that can no longer be balanced using traditional approaches.
Who or what is the biggest beneficiary of self-service analytics? What results are actually enjoyed?
It is difficult to pick one winner. Business users obviously win because they get to the data faster and they can use their unique business knowledge and context to find and shape the data to serve their needs. IT wins, too, because they no longer have to spend multiple iterations on every request. They can now focus on curating and optimizing the architecture and letting the business users handle the last mile of data prep.
Productivity gains for end users and IT is a major win-win. Furthermore, self-service encourages a mindset of experimentation and rapid testing of ideas. The business potential of that is major for organizations looking to adopt a true data-driven culture.
What do organizations need to know before adopting self-service analytics?
Self-service analytics is a major win for any organization, but it should become pervasive and broad because it is not helpful to have self-service analytics tied to one specific BI tool. A big part of self-service analytics is getting access to the data needed to answer your questions. That data needs to be provisioned to work in the BI tool of your choice, whether that is Tableau, Excel, Microstrategy, or Microsoft Power BI.
The other critical element is to plan for -- and implement -- proper governance and security to help manage your analytics initiatives. Data is a crucial asset that needs to be managed, especially with growing regulatory requirements such as the Global Data Protection Regulation (GDPR). Today, the ability to properly secure and track usage and lineage of data activities has never been more crucial.
What are the main roadblocks to self-service analytics? What best practices can you recommend to avoid these problems?
Success with self-service analytics requires you to cover the usual people, processes, and technologies. Too often, we pick a technology that does not scale to our data or does not support enterprise-grade security. We may fail because we invest in a technology for self-service analytics, but we keep running the business on traditional gut-feel decisions.
In terms of overcoming some of the roadblocks, here are a few considerations:
- Technology: Make sure your technology for self-service analytics is not limited to a single BI tool but is part of all tools you use today (and will be incorporated in tools you use in the future). Your solution needs to scale to enterprise-size data volumes and variety. You also need to ensure the product has enterprise-grade security and governance capabilities.
- People: Encourage data-driven decision making and an experimentation mindset at all levels. Failure is part of the journey. Learn how to fail fast, and be sure to celebrate each data-driven success.
- Processes: Establish internal centers of excellence and user communities. Encourage collaboration and sharing of data sets and experiences. Make sure end users and IT professionals regularly exchange ideas and best practices as well as demonstrate results and use cases for their colleagues.
James E. Powell is the editorial director of TDWI, including the Business Intelligence Journal and Upside newsletter.