7 steps to mastering big data

“We’re convinced about big data, but it’s almost like analysis paralysis,” a banker said recently. “If we could just find a starting place!”

Big data appears to be a confusing challenge right now. But if you still feel heavy pressure to solve it, you are in good company. Gartner says that until 2016 confusion about big data will constrain spending on business intelligence and analytics software to single-digit growth. But by 2017, more than 50 % of analytics implementations will make use of event data streams capable of producing autonomous insights. 

This means a long-ish learning curve followed by rapid gains in capabilities.

Will your bank keep up?

How do you begin?

Everybody knows banks are awash in data and struggling to discern what’s useful without being hampered by the rest. Sometimes they are urged to embark on pie-in-the-sky schemes for harnessing big data that tend to be short on process and proof. 

But experience has produced some proven steps to make your investment pay off.

1. Start with your business heads.

Ask, “What’s your problem?” Help them be specific.

Maybe they have assembled a great data set but are hobbled by primitive tools that take too long to populate and too long to answer. Regardless of the problem, data can help solve it. Despite a reputation of bits and bytes to be mined, stored, and processed, it’s data that drives your bank’s business decisions. Start talking about data the way you talk about value.

2. Innovate by creating new data-based products and services.

That is, package what the data teaches so that data users can reap ongoing value from it.

It’s one thing to analyze the data until you can confidently inform your mortgage unit why they end up turning down so many applicants. How much better to provide them with a data-driven tool for better targeting mortgage customers? Or to embed a data-driven alert in your email system that alerts employees when they venture close to a compliance issue?

3. Don’t skimp on the science.

You need trained data scientists to make sense of all the information that will be coming at them. Yesterday’s data analysts are accustomed to analyzing single-source data, or answering predetermined questions. The data scientist you need has to be equally comfortable exploring multiple and disparate streams of data—including unstructured data—and leaving their minds open to unanticipated clues and conclusions.

4. Integrate your new analytics engine into your existing architecture.

If, like many banks, you have spent two decades and a great deal of capital building up your business intelligence and data warehousing infrastructure, you are ready to reap the rewards.

That takes a team approach to bring together your data scientists with your experts on your business intelligence and data warehousing environment. Combining them will result in the business insights and predictive solutions you have long envisioned coming out of Big Data.

Until integration occurs, all those benefits are just “potential,” and as a famous coach once said, “Potential can get you fired!”

5. Employ new ways of capturing information.

With the internet of things, reams of new information are becoming available that you have never captured before. Your mobile-using employees and customers are potential sources. What can you use and how?

If you’re using biometrics for customers to authenticate themselves before transaction business, you can immediately bring together all information about that customer in a few seconds and be able to offer them the right set of new products based on his past history. That’s just one example among many that will emerge when you establish this mindset.

6. Identify and leverage previously unused internal and external data.

Now that you have integrated big data into your strategies, it’s time to go back and review it.

For example, you’ve been running social media campaigns for some time, but perhaps you haven’t been set up to incorporate it into product decisions. Your recruiting process has thrown off fascinating information top referral sources, but it hasn’t made it into your best practices. Now you are in a position to link this information to current data sets and improve your business decisions accordingly.

7. Automate to accelerate.

Information is power, and time is money. Big data, deployed properly, should virtually eliminate any latency between the data and your ability to know it, understand it, and apply it in real time.

For example, banks use manual processes to check for compliance and evaluate risk management. They tend to hire hundreds of staff to manually analyze emails and documents to check for potential fraud, insider trading and many other compliance issues. Using text-mining capabilities from a vendor a bank can automate the analysis and review of these documents.

It has probably been a long journey from the early days of your bank’s data warehouse to today’s ready-to-harvest potential. Now a proven roadmap can take you that last, valuable mile.

About the authors

Bob Olson is vice-president, Global Financial Services, and Rod Fontecilla is vice- president, Big Data Analytics Center of Excellence, Unisys Corp., Blue Bell, Penn. They can be reached at [email protected] and [email protected] .