In the world of big data, analytical sandboxes have been there since the beginning, letting analysts pour in data and perform various modeling and other analytical tasks. These sandboxes have been installed at many organizations, including financial services institutions, where they are frequently used as an accelerator to perform data research and development frequently containing sensitive information and strategies.

As these sandboxes have grown in popularity, the opportunity they present needs to be balanced with some responsible boundaries. It’s a situation that, in many ways mirrors the popularity and decline of businesses using Microsoft Access in the 1990s.

Free & Easy – And Dangerous

Back then, many data analysts, operations executives and accountants jumped on Microsoft Access because it was easily accessible, easy to learn and a powerful tool allowing them to record, analyze and play with data. It provided a significant data processing capability, unrestricted by their organization’s IT SDLC and Governance. They liked it so much that they used it every way possible to work around IT.

On one level, it was understandable -- they could sit down at their desk and start making sense of data according to their role in the business. They could get a good enough view to drive decisions and activity in a fairly interactive manner. Except that they shouldn’t have, as the process often involves skipping testing and requisite checks and balances to ensure the data is accurate.

Since users had been dodging IT at every step, when the time came to hand off their plan to make it happen, everything froze. From the start of trying to transition their ungoverned analytics to IT for production, it could take many months to see the light of day. This produced a major headache for IT in the 2000s as they tried to translate workable information from this ungoverned desktop environment and convert it into robust managed solutions.

At this point, many businesses had been burnt by this process or feared a catastrophic financial outcome stemming from untrusted data. The friendly, easy-to-use desktop application of the people was becoming a nightmare for businesses. So they stepped in, hard.

Today, you’d be hard-pressed to find analysts or accountants using Microsoft Access at any major financial institution, and very few at most other organizations. In fact, many enterprises would not concede to install Microsoft Access on users’ desktops as a matter of policy. Lesson learned. Right?

Playing it Safe in the Big Data Sandbox

Now, we live in a big data world where users are jumping into big data sandboxes with both feet, excited by the possibility of data insights and actions once thought impossible. Careers have been made on less. When you’re on that kind of track, you don’t need IT clamping down on the fun. Sound familiar?

The good news is that big data really can propel data use to unreached heights for business, and it’s on a managed, scalable platform, not a desktop, so it’s not – necessarily – a screaming security risk. But the handoff, as it stands now for most organizations, is still not very smooth.

For many businesses, the trip from big data sandbox to IT to production implementation still takes many months. But the stakes are too high to just step on the process this time. No smart business wants to stop the imagination, creativity and planning emerging from those big data sandboxes, and they don’t have to.

Here’s the answer -- create a bridge between the sandbox users and IT by instituting some rules that work on both sides.

1. Establish a Data-First lifecycle management and data governance framework, established at the starting point of the sandbox. This approach has been proven to move data plans from sandbox to production smoothly and quickly – often in less than thirty days.

2. Commit IT involvement early in the development process. By making IT part of the solution, the transition process from business R&D to Production Readiness would be streamlined and become more efficient.

3. Let the business show IT what it is they are interested in and IT can guide them back on the constructs of how to develop their models in an industrialized fashion using methods requiring less conversion time.

Catching it early in the data development lifecycle, with a set of controls that the organization trusts, is the way businesses can capture both the innovation and application to succeed. The tools and technology are there for accountants, analysts, IT and management to embrace, adopt and deploy data development that can make the promise of big data come true for their organization.

(About the author: Avi Kalderon is the Big Data and Analytics Practice Leader at NewVantage Partners)