MyPrivateBanking Blog
Daily Comments on the World of Wealth Management

Archive for the ‘Business intelligence’ Category

Why BigData projects often fail (and how to make sure yours isn’t one of them)

Thursday, June 1st, 2017

(By Onawa Lacewell, Analyst)

Big Data projects in the financial sector (and other sectors) often produce underwhelming results. It may be tempting to conclude that Big Data is just a passing fad and that wealth managers need not concern themselves with Big Data solutions. However, MyPrivateBanking’s new report on Big Data in Wealth Management thinks this is the wrong conclusion to draw from the current lackluster performance of Big Data projects. Instead, we argue that poor implementation planning that fails to take a needs-based approach is to blame for the inferior performance of Big Data projects to date. We suggest that by reversing the standard implementation process wealth managers can help ensure that Big Data projects succeed.

Why is Big Data more than just hype?

Big Data may seem like just a passing fad or a buzzword. Nevertheless, Big Data is here to stay and, from everything we’ve seen in the past five years, it seems highly likely that we are just now at the dawn of Big Data. “Big Data” simply refers to the increase volume, variety, velocity and veracity of data flooding into today’s businesses every day.

“It took from the dawn of civilization to the year 2003 for the world to generate 1.8 zettabytes (10 to the 12th gigabytes) of data. In 2011, it took two days on average to generate the same amount of data.”

(icrunchdata)

In today’s highly digital world, almost everything we do generates data: location data, social media activity, search engine metrics, and banking transactions. Additionally, as the Internet of Things brings more and more of our activity online this volume of data is only set to grow-from smart houses that provide insights into how you live to smart cars that, for example, allow insurance companies to closely monitor your driving habits. Businesses can now get nuanced, granular, and highly accurate data about their customers-both existing and potential. This is, of course, the lure of Big Data. For financial actors, the potential of Big Data is quite high. Banks can see when, where, and how their clients spend money. Wealth Managers can understand more about the behavioral profiles of their U/HNWIs. All financial actors can accurately and easily automate manual data entry processes both freeing up time for personnel to work on other tasks and reducing the possibility of data entry errors.

Why Big Data projects fail

The purpose of Big Data is clear: to help drive innovation and profits for businesses, to apply a fact-based business strategy, and to uncover insights to help increase organizational efficiency and improve client relations. Why, then, do Big Data projects have such a reputation for underperforming? MyPrivateBanking argues that much of the failure of Big Data is not because the data is somehow less useful than imagined but rather that the way many firms approach implementation sets these projects up for ultimate failure. The standard implementation process often starts with shopping around for a Big Data vendor-perhaps one that promises an all-encompassing Big Data solution that will use the firm’s internal and external data along with unstructured data (like social media commentary or search engine metrics) to modernize the entire digital ecosystem. Then, once the vendor’s solution is in place, the firm realizes that they actually don’t need all the data that they are collecting. Or, that there isn’t inhouse data science talent that can really get the most out of this new wealth of data analytically. Or, possibly, the firm realizes that the organizational siloing is standing in the way of using the new data.

A Needs-Based approach to Big Data

We argue that in order to get the most out of Big Data, and to ensure that Big Data projects are really successful, wealth managers and other financial providers should reverse the standard implementation process. Instead of focusing on the end solution, or trying to modernize the entire digital ecosystem with a general and comprehensive Big Data project, firms should instead take a needs-based approach to Big Data. The steps of this approach are rather simple, but this simple change in approach to the implementation process can make all the difference when it comes to whether your project will be successful or not.

1.     Identify the exact need (objective) of the project

This is a key step and should not be undertaken quickly-determine explicit needs, or objectives, where your firm needs a Big Data solution.

2.     Determine the type of data that best address this need

Do you need structured data? Unstructured data? A mix between the two? Determining which type of data addresses your need will help determine what type of data solution you require.

3.     Evaluate whether this data already exists within the organization

A lot of organizations think that Big Data means external data-using Facebook data, for example. However, Big Data can also mean internal data. Taking a deep look at the types of data your organization or firm already has may reveal that you don’t need external data at all-and this will determine what type of third party solution you need to shop for.

4.     Determine success metrics and expected ROI

Determining the success of a Big Data project can sometimes be difficult. Therefore, it is crucial that measures of profitability be part of the pre-planning discussion and strategy meetings.

5.     Shop around for a vendor who offers a solution that fits closely to the need

There are many different vendors offering everything from comprehensive Big Data solutions to narrowly targeted ones. Seeking a vendor that fits to your specific organizational need will help ensure that the resulting implementation plan will be a success.

6.     Determine the correct infrastructure and implementation plan to fulfill this need

Only after every other step in the needs-based chain is fulfilled should a firm or organization determine the type of infrastructure necessary for a Big Data project. The need should always drive the infrastructure-not the other way around.

By approaching Big Data from a needs-based plan, wealth managers and other financial providers stand a better chance that the resulting Big Data project will be successful. For practical information about how to take a needs-based approach to Big Data see our latest report. Here you will find practical tools that will help with determining success metrics, charting the implementation path, engaging in pre-planning and more.

 

In-memory analytics about to shatter the walls of bank’s traditional BI

Friday, October 24th, 2014

In the light of big data and real-time business intelligence discussions, one thing becomes increasingly clear: data latency is expensive, old-fashioned and not competitive. As things speed up, new technologies are needed that can cope with increasingly challenging demands. This is especially true for the financial industry where time really is money.

In-memory analytics has great potential to become the philosopher’s stone in this issue. The concept is simple: traditional BI queries data stored on physical disks whereas in-memory analytics uses data and queries located in the server’s RAM, making query results available near time. While this concept is not new, it is far from standard in the finance industry. Yet.
As pioneer banks are taking their first steps into IMC (in-memory computing) – such as Germany-based Dekabank or Swedish Avanza bank – we will be likely to see a fundamental technological turnover in bank’s BI in the near future, triggered by falling costs and increasing capacity of RAM as this article describes very well.

The advantages for banks are obvious: Dekabank’s use of Quartet FS is only one example how in-memory computing boosts performance through high-speed risk analysis combined with trading positions, which allows for faster reaction and near time alerting. Rapid fraud detection and credit card reporting are other benefits to name only a few.

As IMC gains ground, the heavyweights of the IT industry come up with their solutions. Quartet FS, Oracle TimesTen, SAS High-Performance Data Mining, SAP HANA or IBM DB2 with BLU Acceleration are some examples. We at MyPrivateBanking Research are looking forward to see how fast the finance industry will be able to adopt this promising technology.

 
Subscribe