With more data to play with than ever before, asset managers need a solid strategy if they want their sums to make any sense...
The data revolution is real. And while algorithms generate custom-targeted ads and online shopping services remember our favourite foods, asset managers are also finding that this ‘internet of things’ means there is more information at their fingertips than ever before.
Like everyone else, asset managers are scrabbling to figure out how to make the best use of this new digital world.
And like everyone else, they’re well aware that when properly managed and analysed, data can bring about business opportunities, potentially giving them that much-sought after competitive advantage.
Francis Jackson, head of global client coverage at RBC Investor & Treasury Services, says: “Good data analysis can assist asset managers in revenue generation, product development and with distribution strategies in order to add value.”
“To that end, it can be used to provide useful intelligence to help, for example, identify sales patterns, distribution preferences and investor behaviour.”
Jackson also draws attention to emerging intelligence tools that can analyse asset managers’ data, allowing them to access the information in real time—or at least something close to it—such as RBC’s own Fund Sales Intelligence Tool.
Jackson says: “They can identify the countries and distributors where their funds are selling well, as well as search consolidated data extracted from all our data flows to gain a macro view of what is happening in each country, for each type of distribution network.”
Samir Pandiri, CEO of asset servicing for BNY Mellon, gets a little more into specifics, suggesting that good data analysis practices can help inform not only asset managers’ investment decision-making, but also their product design and distribution priorities.
BNY Mellon is pairing with Heckyl Technologies, a fintech company that scans social media and 120,000 web-based information sources, in a bid to help asset managers better assess public sentiment around the companies in their portfolios.
Pandiri also points to Albridge, an affiliate of BNY Mellon subsidiary Pershing, which allows asset managers to analyse the products that investors are buying, and the geographical differences that these results can uncover.
“Asset managers can use this data and then talk to their teams in each city or state to determine which funds are selling successfully and why. This in turn will influence how asset managers market their products and help to ensure that they are not missing a selling opportunity.”
In theory, Pandiri says, using technology like this, one management team could use the data to replicate another’s successful technique, or at least gain a better view of what to market in its own area. While agreeing that good data analysis is “frankly, invaluable”, Chris Ellis, senior vice president of business development at financial data service provider FactSet suggests that in order to glean any real insight from the data, asset managers have to know what they’re looking for.
Ellis says: “If an asset manager doesn’t have an approach or a strategy, if it doesn’t really know what it’s looking for in the analysis—what the question is—then big data can overwhelm them and they can get completely wiped out. It’s just too much.”
In fact, the sheer volume and complexity of the data at asset managers’ fingertips can actually add to the confusion. For example, managers could stress test for any number of different scenarios, without what Ellis calls “a clear thesis” of what they’re testing for—looking for the outcomes of a stress that they actually believe could happen.
He advises: “Don’t stress test for things that have little to no chance of happening just because you have access to that data. Complexity without clear purpose and vision leads to confusion.”
In fact, such issues of complexity come into play as asset managers choose which data streams and sources to prioritise and investigate in the first place. With so much data available in the ether, there has to be a strategy in place.
Jackson explains that asset managers should identify exactly how they see data as a commercial driver; a driver that “identifies the data they possess themselves, any additional information required from third parties—such as demographic or macro-economic—and how that data is to be extracted and presented”.
This data should then be made quickly available, easy to access and easy to interpret. And such things are in the pipeline.
“Within big data, enhanced and predictive data management applications are being developed to provide meaningful insights and aid in prioritisation,” Jackson says.
“Advances such as ‘data lakes’ where data is stored in its raw, unstructured form make data streams potentially more agile than when stored in a traditional ‘warehouse’—where the data by definition is pre-structured and categorised—as data can be configured, changed and extracted as needed and according to differing criteria or applications.”
However, no matter how well managed the data may be, asset managers must also consider the reliability of what they’re presented with. Big data, Ellis says, implies more data, and it is a “horrible assumption” that this means clean data. In fact, it is quite the opposite.
“Bad data creates a house of cards on your analysis and undermines it as the bad data tends to stand out, obscuring the real result.”
In order to prevent these kind of errors, first asset managers have to fully trust, understand, and scrutinise their sources—and, crucially, Ellis says, bad data is not always necessarily bad simply because it is wrong.
When using multiple data sources, it is equally important that the “symbology fits together”.
Identifiers must be the same for securities, companies and funds, so that when married together, like figures blend, rather than creating two separate data points.
For example, one security could be titled differently in two different data sets, and so it will show in the asset manager’s final figures as two different securities. While not inherently wrong, the data produced is bad nonetheless.
“First you have to ask whether it is high quality data unto itself. Then, when you marry it to another data set, do they fit together in a tight weave, or do they just bump into each other because they’re inconsistent? It’s not always obvious.”
It is a well-worn adage that the institutional finance industry isn’t known for its swift uptake of technology, or for its ability to change at any speed.
This could mean, Jackson suggests, that actually asset managers have simply been unable to get the best out of the data currently at their disposal.
However, he also points to two developments that have brought big data to the top of asset managers’ priority lists.
“Firstly, the increased requirements for greater collection and management of data to enable more effective regulatory reporting has meant that asset managers and their service providers now hold more information than ever before that can potentially be used to add value.”
“Secondly, the rise of fintech has meant new services and applications are in development to allow asset managers to use their data.”
However, while they may have made progress, there is still work to do. Pandiri stresses: “There isn’t a single asset manager that couldn’t enhance its business by doing more with data.”
He says: “Some asset managers are quite advanced and have sophisticated distribution analytics and performance analytics that really help them. Others are behind.”
Those that are successful here, Pandiri says, are those that adopt the technologies available to them early, “ensuring best practice; and maintaining a high rate of investment in technology either internally or with a third party”.
Ellis agrees with the sentiment, suggesting that the asset managers that are succeeding are those that are harnessing unique data—geographic exposures and supplier chains, which can offer a different perspective on companies that are about to thrive, or those that carry subtle, but important, risks.
“These asset managers are harnessing big data to get the results they need. It means they’re able to use much more complex, much less ‘vanilla’ data to get there, and that’s pretty exciting.”
