Unlocking the Value of Advanced Analytics
At Informatica World, JP Morgan Chase shared how to unlock the value of Advanced Analytics
“I wouldn’t be surprised if we started to see the number of physical [bank] branches diminish over the coming years,” said David Gleason, Firmware Head of Data Strategy and Governance for JP Morgan Chase, pointing to the rise of banking from mobile as an example of the sort of major disruption we take for granted, in this case within the banking industry. He’s also seeing rapid growth in people testing robo-advisers, software-driven investment advice which he described as the investment equivalent of autonomous driving, and consideration of crypto-currencies like Bitcoin.
All of these changes are powered by data and analytics, of course, and the challenge for folks in his position, says Gleason, is to adapt to all the changes and take advantage of opportunities. But it isn’t easy and it isn’t automatic—Gleason offered the statistic that more than half of big data or advanced analytics use cases fail to flourish or be sustainable after initial promising results. They become unsustainable, or the cost is too high relative to the value, or the data quality isn’t right.
He defines six metrics for data that can help guide the hand of thoughtful data architects. They are:
- Meaning (are all internal stakeholders speaking the same language?)
- Sources (where did the data come from—there can be disagreements)
- Provenance (what’s the chain of custody and what was done to it along the way?)
- Quality (is this good enough for me to trust? For the legal team?)
- Permissible use: (am I entitled to use this data? Is it legal and ethical; will it keep me out of trouble; does it hold to our brand standards?)
- Obligations: (does it include personally identifiable data; do I have to encrypt, etc.)
He described how data responsibility now belongs to employees across the company. Today, many IT departments are demanding limits on their responsibility, taking the position that their job is to protect the perimeter, and to apply specific technical protection to data as directed. “We as biz owners need to tell them where to apply effort and how much.”
The prize for getting it right, said Gleason, is confidence in analytics, from multiple standpoints including Validity, Accuracy, Efficiency, and Compliance. It can be a moving target, but when a company has a communicating “data community,” when the data officers have the authority and budget and supporting tech to get the job done, and when there’s a sustained commitment to data over time, that’s a pretty solid formula for success.