Digital Assets Report


Like this article?

Sign up to our free newsletter

Man & machine: Building a knowledge platform to develop predictive prowess in asset management

Related Topics

Big changes are afoot across a broad spectrum of the financial services industry. For example, in the wealth management and investment advisory space, thanks to huge technological developments in recent years, traditional wealth managers have a new competitor to face; the robo adviser, which sounds exciting but it is less the Terminator and more machine learning software, using algorithms to learn investor behaviour and automatically invest into portfolios based on clients' goals and risk threshold. 

Such is the rise of the robo adviser, which includes Betterment and Wealthfront, Nutmeg, Money on Toast (!), not to mention Vanguard and Fidelity, that projections made by consultant A.T. Kearney suggest that AUM in this sector could reach USD2.2 trillion in the next five years. 
The robo adviser trend is pertinent as it provides a snapshot into how leading-edge fund management companies might operate in years to come. Hedge fund managers are often considered to be the smartest guys on the street. But whereas investment prowess and AUM have long been key calling cards to attract investor dollars, those managers who are able to successfully harness Big Data and bring more of a quantitative, machine-learning element into their business model are more likely to be tomorrow's success story. 
As the New York Times pointed out in its "Don't Fear the Robots" article on 24th October, 2015, the more raw data (from a volume, depth and breadth perspective) that is ingested, the smarter the artificial intelligence becomes. This is being applied in fields such as medicine, where massive databases of information on clinical conditions are enabling consultants to accurately predict the outcome of a client's condition and, scarily, even predict when a patient is likely to die. 
Big Data 2.0 – harnessing data to drive insights

Asset management has always been an information-based business and as such, managers are not immune to innovative and disruptive forces making inroads in other industries. They now have the chance to incorporate and leverage the same predictive capabilities into their organisations whereby they can better understand investor behaviour and derive insights across all facets of their business. 
"We live in an age of data abundance," says Paul Schaeffer, a senior adviser at SEI's Investment Manager Services division. "To some extent, all businesses are becoming `Googlised'. The challenge for asset managers today is not how to gather the data but rather how to extract value from all the data they now have at their disposal."

Back in 1994, Stan Davis and Jim Botkin wrote a seminal paper for the Harvard Business Review. Entitled "The Coming of Knowledge-Based Business", the paper laid out a hierarchy of data to information to knowledge; with data representing the building blocks of a knowledge-based business. Information is data that has been arranged into meaningful patterns, said Davis and Botkin, with knowledge deriving from not only the composition of information but the skill that one is able to apply to that information.  
This prompted SEI to start thinking more seriously beyond the concept of big data about the components of a knowledge platform that an asset manager would need, going forward. They identified four main components, which include:

Sophistication – A knowledge platform needs to be highly sophisticated in order to handle and analyse all types of data, while ensuring data integrity. 
Support Key Business Functions – The platform must meet the data, information, and knowledge requirements for all the organisation's key functions. For example, what are the outcomes that your compliance function, or risk function, or your portfolio management team need? 
Flexibility – A logical progression of point 2 from the above, is how the platform must be flexible, adaptable and efficient. The information must be delivered in a clear and intuitive manner, but one size doesn't fit all. An online dashboard, for example, shouldn't look the same for the client relations' team, compared to the compliance team or operations team. 
Expertise – This is probably where the asset management industry is farthest behind the curve. Does the platform have that deep, specialised expertise and analytical capability? 
With respect to expertise, various industries (e.g. telecoms) are hiring data scientists and analysts to make that transition into becoming a knowledge-based business and embrace the "machine learning" paradigm that is fast underway. The biggest challenge for asset managers is putting such a strategy in place and making it a core tenet of their business. The technology components are, in that sense, less important. Once they've established a strategy, they will then be able to decide whether to build their own in-house expertise or partner with third party specialists, to process data, analyse it, and turn it into insight and knowledge. 
The Watsonisation process 
One of the revelations of Watson, the IBM supercomputer used in the US game show Jeopardy, was its ability to learn and understand human context, demonstrating just how far computers had come in a short space of time. 
What is exciting for asset managers is the ability to harness that same predictive, machine-learning capabilities in how they run all facets of their organisation, including portfolio management. 
"One of the earliest ways of bringing predictive analytics into the fund management industry was in investor servicing, knowing when someone was likely to redeem or were open to subscribing. Based on millions of transactions, analytics enable you to derive better insights into future investor behaviour," states Ross Ellis, Vice President and Managing Director of the Knowledge Partnership in the Investment Manager Services division at SEI. 
"Some firms are also starting to use predictive analytics in respect to cybersecurity strategies, to learn from hacking and security breaches to improve their cybersecurity framework and improve compliance controls," adds Schaeffer. 
This is particularly useful as managers develop a wider range of customised fund products, including regulated liquid alternative funds that require close monitoring of regulatory guidelines and daily liquidity provisions with the portfolio. 
The SEC is currently looking very closely at alternative fund managers who are moving into the '40 Act alternative mutual fund space with respect to liquidity management. As such, alternative fund managers are going to require a greater understanding of what their liquidity flows are and develop predictive models on how they are going to change based on market conditions. 
Predictive analytics could become an effective tool in meeting that challenge.

Decision-making platforms: the next battleground

Ray Dalio's USD165billion hedge fund manager, Bridgewater Associates, this year started a six-person artificial intelligence unit. The group is headed up by former IBM executive, David Ferrucci, who joined Bridgewater in 2012. As Bloomberg pointed out in an article on 27th February, 2015, the unit will be responsible for developing trading algorithms that make predictions based on historical data, and learn how to respond to changing markets. 
This is, in effect, incorporating a robo adviser component to the firm; but rather than advise investors, the algorithm suite uses predictive capabilities to improve Bridgewater's trading expertise. 
In that sense, they've laid down a marker. But as Ellis and Schaeffer are quick to point out, having what they call a knowledge platform is not the preserve of the biggest managers; firms of all shapes and sizes have an equal chance of success.

"The whole Big Data phenomenon can help smaller firms compete against bigger firms," states Ellis. "Just as Big Data has disrupted other industries, such as Amazon with books and Uber with taxis, you could well see not only a disruption within asset management as it exists today but also new players coming in from outside the industry," continues Schaeffer. 
This is already happening, particularly with respect to transaction payments. Apple now has Apple Pay, Google has Android Pay, and recently PayPal stepped up their game with the acquisition of Venmo. These three technology giants, in tandem with Amazon and Intuit, have created a financial services lobbying group called Financial Innovation Now. 
These firms have the capacity to harness mobile technology and use their knowledge-based platforms to start to compete with old-fashioned financial services firms, at the same time making financial services "more accessible, more affordable and more secure," said Brian Peters, executive director of Financial Innovation Now. 
"Asset management is an industry with great growth and great margins so to think that it's not going to find itself competing against disruptive non-traditional players, or indeed traditional players who are using analytical capabilities, is unrealisic. Asset managers who understand the merits of knowledge platforms versus those who don't, will determine competitiveness more than other factors, such as AUM, going forward," opines Schaeffer.

As Ellis adds, "it's not too far a leap for the likes of Apple or Google to take their analytics and quantitative capabilities and apply them to managing money funds, or index funds. That's not beyond the realm of possibility." 
This is ultimately more of a cultural orientation. Are asset managers willing to dedicate resources and attention to having a knowledge platform? 
Tapping into new data sources

Another advantage to having a knowledge platform is that it enables businesses to exploit myriad sources of new data to use to their advantage. 
Some of the big intermediary platforms, such as Merrill Lynch for example, are now offering fund managers access to information on what underlying intermediaries are doing, in terms of types of strategies and managers they are looking for on behalf of investors. Equity analysts can use drone photographs to determine the number of cars in Walmart parking lots on Black Friday instead of relying solely on reported daily sales figures to judge or anticipate consumer spending over the Christmas shopping season. 
"There is an abundance of data to harness. Tom Davenport, sometimes considered the father of Big Data, talks about the Age of Analytics 3.0 where the next challenge is for businesses to use data to create new products, figure out new services, and to really differentiate themselves in the marketplace. That is, moving from purely analysing their business internally to using data as a tool for developing and enhancing their business strategy. To do that, though, requires a knowledge platform," states Schaeffer. 
Get the right people and partners in place 
A firm's cultural orientation will ultimately determine the extent to which they build internal capabilities – be it through hiring data scientists, a Chief Information Security Officer etc. – and/or partner with external service providers. 
Asset managers should look at all the entities they work with – fund administrators, custodians, risk providers, brokers – and ask, `Do they have the same orientation around data and knowledge and, therefore, can they support my knowledge platform?' Are they going to deliver the data I need, do they have individual skill-sets that I can leverage? 
"That will apply both on your service provider side as well as your customer side. If there are multiple intermediary platforms that I could choose from, perhaps I should think about being on the platform that is able to give me the richest insights compared to the one that is less willing to provide as much transparency.

"You could see how this could really affect all aspects of the way asset managers run their business, beyond merely the functional aspects that people typically think of," says Ellis.

Establishing the knowledge platform 
When it comes to putting a knowledge platform in place, much of the infrastructure already exists. There's plenty of computing power in the cloud, plenty of analytics. Administrators such as SEI are building key components and online dashboards. 
The bigger issue is putting the strategy in place in the first instance to construct the knowledge platform. 
"One of the big stumbling blocks in the past was computer power. That's simply no longer the case. Machine learning and artificial intelligence is coming into its own, with IBM leading the charge. As mentioned earlier with respect to Google and Apple, you might start to see companies use machine learning at the core of their strategy and then wrap an asset management business solution around it," says Schaeffer. 
Ellis says that it is incumbent upon fund managers "to make a serious commitment to a knowledge strategy including hiring the right people to lead such a charge and determine who the best firms will be with whom to partner. You'll likely begin to see some new service providers emerge to add on the list of usual suspects."

Practical insights 
Over the last 10 years, SEI (and other fund administrators) has seen quite an evolution in terms of the breadth of services that it provides to clients.

It has gone from striking the NAV for pure-play funds and doing reconciliation, to supporting multiple asset classes across multiple product lines i.e. hedge funds and mutual funds or private equity funds and hedge funds or collective trusts. Recognising the efficiency, scalability and economic advantages of outsourcing, clients then started asking for services such as middle-office processing and global regulatory & compliance solutions. 
Indeed, it is a misnomer to call firms like SEI a "fund administrator", as back-office fund admin is just a small part of what they now offer, as Jim Warren, Head of Solutions Strategy & Development at SEI discusses: "In addition to being concerned about the standard end-of-day or end-of-period NAV, clients also want to know about portfolio attribution and drivers of performance as well as transparency into how the portfolio was processed operationally. No longer it is just about the end product – managers are looking for insight into our workflow and processing environment. Was the portfolio reconciled with the custodian or prime broker? Where are we in our processing throughout the day prior to releasing the NAV? Which investors have submitted all the necessary forms to transact and which ones have yet to complete the sub docs? 
"So it really changed the type of services we were providing, and the key to providing those services was one element: data."

This broader evolution led SEI to move away from processing data in silos to focus more on throughputs and intraday processes. 
"We set out to do data aggregation on portfolios, on the funds, on the clients, and on the investors and that allowed us to provide a greater level of service; we could more readily do middle-office processing with aggregated data, we could do cross-product reporting and analysis for clients so they could see exposure across their products. This enabled them to begin to do analysis on where they were more successful with investors, where they traditionally had struggled, and things of that nature," says Warren.

This was made possible because the data was normalised and aggregated in one area. With more data at their disposal, SEI was able to provide even more solutions including regulatory processing solutions, investor processing and reporting, and better front-end reporting tools. 
"Then we thought, `We've got a scalable model that allows us to aggregate data in an open architecture environment, so we should start collecting and aggregating data that we don't process.' It doesn't matter whether that information is generated by the investment manager or not, there is data in the wider marketplace that might help our clients make better, faster decisions. We decided that there should be no limit to the type of data that we aggregate for our clients," explains Warren.

Regulatory reporting insights

One of the first drivers of regulatory reform in the alternatives space was Form PF, the rule jointly adopted by the Securities and Exchange Commission (SEC) and the Commodity Futures Trading Commission (CFTC) back in 2011.

Initially, many hedge funds looked to accomplish this by pumping data into spread sheets, reviewing it manually, and then sending it out. 
As this would likely be unscalable and rife with errors, SEI built a tool to sit on top of its data aggregation layer. What this did was to make the Form PF process more efficient and repeatable but, more importantly, it allowed the firm to aggregate a wider breadth of data on the investments in its data warehouse. 
"Instead of just doing a compliance review process for the CCO and compliance team, now we could take the Form PF data and share it with the CEO of the fund manager, or the investor relations department, to give them more insight into the investments that are processed, to help with their decision planning and asset growth," says Warren.

The theory is, if service providers can give clients data more efficiently, they have more time to understand it, build insights, and make smarter decisions. 
"We've certainly seen that on the Form PF side, and have leveraged that learning to provide solutions for other regulatory filings including CPO-PQR and AIFMD," continues Warren. "While these dealt more with the investment portfolio, we were able to take our experience and offer solutions affecting the end investor. 
"One of our first investor-oriented solutions helped managers with FATCA tax processing. With FATCA, we've taken an investor record that resides in our data warehouse and through our open architecture platform, we've been able to enrich it due to the fact that there are additional data points on investors needed for this regulatory filing."

What this has done has normalised the investors. SEI can take in non-transactional data as well as data it doesn't process for NAV calculations, giving it a larger aggregation of data from which to report. Its online Investor Insight platform gives the manager the opportunity to use that larger, fuller data set to develop deeper analytical insights into how they are marketing their funds (i.e. is there a pattern among where investors live, their age, their net worth?) and hone their strategy. If the manager is losing investors through redemptions, is there anything consistent in the nature of those investors?

"This is where data starts to become a driver of your decision-making process. And that's the goal. That's the real value," says Warren.

For some time now, administrators have offered performance attribution tools: what happened in the portfolio, and why? More recently, this has been expanded to not only look at attribution in the investment process, but also with respect to the fund's investors, and other factors that drive the business. 
This has, until now, been purely a backward-looking solution, using historical data. Warren says that SEI is focusing on developing solutions that will give managers the opportunity to derive forward-looking insights by doing predictive analytics.

"What we're trying to find is the best, most flexible, way to deliver such a solution to our clients. We want it to be an open, not a canned, solution, to give our clients the opportunity to control the process and make the decisions themselves. That's a big push for us: taking some of the tools that we have and enhance them in a forward-looking way," confirms Warren.


Asset managers now operate in a world where complexity is pervasive, products and competitors proliferate, and profits are under pressure. Keeping up operationally is critical, not just a good-to-have. 
In this new operational frontier, infrastructure has become a source of competitive advantage. 
People have spent time getting their heads around how to organise Big Data and building data warehouses. Instead of being overwhelmed by the torrent of data streams and information demands, forward-thinking managers can utilise a well-constructed data management platform to gain economic leverage, meet the ever-evolving demands of investors and regulators, gain new insight into business dynamics and fuel product development. 
Now, managers must recognise the need to go to the next level and establish a knowledge platform; not just gathering the data, but harnessing it and utilising it into a decision-making, predictive tool that supports all key aspects of their business. 
The business of asset management is far different than it was at the turn of the millennium. Managers must not only be operationally adept in order to survive and prosper going forward. Through transforming their data platform into a knowledge platform, they will be better equipped to meet and predict client and intermediary needs, satisfy regulatory and compliance demands, reshape key aspects of a firm's business, and help differentiate themselves in an increasingly crowded and competitive marketplace. 

Like this article? Sign up to our free newsletter

Most Popular

Further Reading