The big trends in data reshaping financial industry

Ed Piolet
Share on facebook
Share on twitter
Share on linkedin
Share on email

For years the domain of sophisticated hedge funds, so-called ‘alternative data’ is gaining popularity among traditional fund managers and asset owners. Coupled with other advances, alternative data can give institutional investors an edge in both alpha generation and risk management.

According to FactSet, a global data and technology provider, it is becoming almost impossible for the average institutional investor to keep up with the demand for capacity, power and speed with in-house hardware and software and the costs and time constraints that come with it.

Ed Piolet, FactSet’s head of sales for their Content & Technology Solutions for Asia says it’s still early days but things are moving very fast.

He said on a recent visit to Australia from his base in Hong Kong: that there were four major trends in data which were reshaping the financial industry:

  • the increasing use of “alternative data”
  • cloud computing
  • machine learning, and
  • data science.

A recent survey by FactSet found that 64 per cent of fund managers believe the use of alternative data can help them beat a benchmark and 59 per cent believe they can improve a portfolio’s risk/return profile with alternative data.

Piolet says that the data sets available now either did not exist or were too difficult or costly to obtain 10 years ago. The use of satellite images for instance, as an indicator of retailer traffic in America’s big strip shopping malls, is an example.

Harnessing the value of Alternative Data is not just about innovative coding and programming. It’s also about getting hold of quality data that can be consumed by these programs and also about being able to quickly process and combine huge volumes of disparate datasets,” he says. “The whole process from data discovery through to its evaluation has become much more complex and lengthy.”

The research ranked managers’ “wish lists” of the types of data they would like to access, with ESG at or near the top of most lists. “we see a huge interest in ESG data the moment,” Piolet says. “This is a result of the combination of new types of mandates, regulation, ethical pressure and also increasing recognition it can contribute to alpha generation. FactSet wants to make the access and the consumption of this data as quick and easy as possible for clients. This means moving to an open and flexible model.”

Alternative data tends to be more unstructured and of generally poorer quality, with gaps in the information than the more traditional datasets used until now such as Financials, Estimates, Pricing, etc. It requires various machine-powered techniques to clean it up and make it usable by mapping it to financial instruments, with the help of the power and capacity of cloud computing and the increasing use of machine learning systems. It’s what Piolet describes as “connecting the dots”.

The two key challenges for data scientists are the discoverability of the required data and its evaluation. You need to be able to get your hands on it as quickly as possible to test your investment ideas or risk scenarios and then be able to fine-tune them.

FactSet’s approach has been to launch an “Open Marketplace” (, which involves an online catalogue of nearly 70 and growing ready-to-use datasets and access to a cloud-based environment, Data Exploration, for evaluation and testing. This can cut down the time involved in such a process from several months to just a few weeks, with huge savings in observable costs and opportunity costs, according to Piolet.

An example of its value to front-office fund managers is the collection of sentiment data on company earnings calls, which typically take place with stock brokers and managers after results announcements. Earning calls is voice data, it is highly unstructured and difficult to use as is in any systematic investment process. The data is then transcribed into a text format and then using a natural language processing toolkit sentiment data can be derived from it. It can then be easily combined with Estimates data for example and new insight can be generated from these calls to anticipate the impact on the stock price before it happens. “When you apply this technology to 10,000 earnings calls a day, which can’t be done by any human but only by machines, it really starts to give you an edge over your competitors” he says.

Stephen Bappert, FactSet’s vice president and senior sales specialist for the Pacific region, says that because of the increasing use of alternative data, managers will need to continually search for new datasets to exploit and maintain a competitive edge. By nature a dataset is only alternative for so long meaning that investors have to be able to quickly extract value from them while already being on the hunt for the next one in order to stay ahead of the pack.

– G.B.

Share on facebook
Share on twitter
Share on linkedin
Share on email