Architecting for Multi-Manager Platforms
Architecting For Multi-Manager Platforms
It’s no surprise that some investment managers walk a fine line between fostering investment diversity and pursuing economies of scale. With few managers able to achieve consistent top quartile performance amidst a proliferation of lower-cost ETFs, crowded investment strategies, and burgeoning operating costs, investment product has become increasingly commoditized. Whereas passive management is a low margin, high volume business that necessitates significant automation to deliver positive operating leverage, active managers, in particular, have acutely felt the pinch of margin compression. This has driven consolidation throughout the industry.
I continue to believe that opportunities exist in private markets and in the delivery of objective-driven investment solutions, including for retail clients. Still, few if any managers have mastered the ability to deliver a heterogenous suite of strategies and more customized investor solutions at scale. Moreover, the challenges of doing so affect large and small managers, alike.
Amidst this backdrop, we find the so-called “platform” managers – those who operate a multi-manager investment platform – exploring, if not already embracing, a shared-services operating model. The primary objective, of course, is to deliver uncorrelated returns in a cash-efficient manner. But there are other potential benefits, including reduced operating risks, an improved client experience across products and strategies, brand recognition, and greater investment transparency.
Certainly, not all shared services are digital in nature – HR and legal, for example. Still, the vast majority of investment management functions are highly dependent on the data and systems that support them. Enterprise asset allocation models and risk management systems are critical to delivering strong risk-adjusted returns across the franchise, while data management and straight-through processing also ensure efficient operations.
Without these capabilities, the multi-manager platform is little more than a collection of bespoke, semi-autonomous units that sit under a common marque, but otherwise offer questionable value to the client and to the underlying managers. It’s therefore worthwhile to consider the digital enablers that can best be applied to such platform managers.
SOME DISCLAIMERS, FIRST
First, let’s be clear: there’s a genuine risk of trying to force diverse investment teams onto a shared platform that proves to be too homogenous. The old adage, “jack of all trades, master of none” applies. Firms that pursue this path rarely succeed. The byproduct is often a cacophony of bespoke workarounds and end-user applications (“EUAs”) that seek to fill the functional voids left by such an approach, but instead only increase risks.
Moreover, the culture of many multi-managers typically promotes autonomy of the underlying investment teams. At least some autonomy is justified, as manager independence helps to generate low correlations…within reason. Multi-affiliate platforms, in particular, tend to be more vertically integrated at the boutique level. This presents a range of both political and technical challenges.
Even as a wholly owned subsidiary, a boutique may be more guarded about their clients and portfolios. Bespoke applications and data may also be so tightly coupled that integration can prove especially difficult. Collectively, this makes finding the right balance between autonomy and shared services particularly important.
Empirically, if investment teams are forced to onboard shared services that are delivered poorly or ill-suited to their requirements, a level of dissatisfaction will permeate the franchise. I’ve personally seen more than a few former investment affiliates celebrate, if not demand a release from their parent organization for exactly these reasons.
BNY MELLON AS A CASE STUDY
When I first arrived at BNY Mellon as Chief Information Officer of Investment Management in 2012, the franchise housed 14 affiliates and the technology landscape was littered with bespoke boutique applications, a lack of any shared architectural standards, and policies that promoted autonomy above nearly all else.
The central technology organization I inherited was a direct reflection of these policies. Somewhat diminished, the team only provided application development for the bank’s central distribution and cash management platforms. It also offered a “center of excellence” for Charles River OMS installations, and acted as an intermediary for infrastructure and help-desk support. The business-as-usual (“BAU”) allocation accounted for nearly 85% of the total budget that year. Simply put, this was not built to be a “change” organization.
To be fair, the policies promoting boutique autonomy were authored during a time when top-line growth was robust. Active management still outweighed passive by a healthy margin, and cost containment wasn’t a significant concern. In fact, it wasn’t until 2024 that aggregate passive AUM finally overtook active AUM, but the trendline has been clear for years.
Moreover, bank executives – accustomed to leading an asset servicing institution – weren’t quite as comfortable navigating the sizable investment management business that was acquired as part of the 2007 merger between Bank of New York and Mellon Financial. One business, investment servicing, was steeped in low-margin processing scale; the other, investment management, favored intellectual capital.
Talent, culture, and metrics were very different between these two divisions. Notably, investment management leadership of the newly combined institution was also largely filled by executives from Mellon Capital Management, then the largest of the asset management affiliates…and, often amongst the most vocal in support of its continued autonomy. Bank leadership, conscientious of not wanting to run the investment management franchise like a processing business, embraced a more hands-off posture with the investment boutiques.
MARGIN COMPRESSION CHANGES THE GAME
In December 2009, Blackrock completed its acquisition of Barclays Plc’s investment arm, BGI. Originally a nascent fixed-income manager, the deal instantly propelled Blackrock to become the largest, most diversified manager on the Street. Blackrock was committed to growing its newly acquired iShares platform, joining the ranks of Vanguard and State Street Global Advisors as part of an ETF triumvirate that currently accounts for over 55% of the total ETF AUM. The race in passives was on! Aggregate global ETF AUM at the end of 2009 was $793 billion; by the end of 2012, it was $1.95 trillion; today, it’s over $11.7 trillion.
By 2012, margin compression was evident…not just at BNY Mellon Investment Management, but throughout the industry. While passives continued to ride a steep upward trajectory, the growth in active management assets had slowed.
With reduced top-line growth and increased bottom-line expense, some multi-manager platforms, like BNY Mellon, turned to the next tool in the shed: operational efficiency. This, of course, is simply a euphemism for cost cutting.
Armed with this new mission, boutique autonomy was now on the table, though still somewhat contained, especially for the largest of managers. With a new investment management leader from Blackrock – known for its embrace of centralization – at the helm, the investment affiliates were understandably concerned that executive management would pursue a one-size-fits-all strategy that would strip the boutiques of their control and dilute their in-house capabilities.
To assuage any such concerns, leadership embarked on a communication and lobbying campaign that sought to re-underwrite the value of boutique diversification and specialization. The message to the boutiques (as well as to bank executives, investors, and analysts) was clear: we will not force you to do anything that might hurt your business.
Still, it was obvious that the operational realities of the investment management sector now demanded greater bottom-line efficiencies and enhanced oversight. The days of unfettered autonomy were coming to an end.
PERFORMING AN ASSESSMENT
Economies of scale and positive operating leverage are like motherhood and apple pie! This is especially true for any set of commoditized products and services where differentiation is limited and margins are under pressure. While certainly not restricted to financial services, many investment management shops today fall into this category.
For multi-manager platforms, the challenge is in how to deliver such scale without diluting the specialization and efficacy of the underlying investment teams. It’s also critical to deliver investor solutions that intelligently leverage the full capabilities of the platform, rather than just offering a series of one-off strategies. This is hard stuff! Missing the mark could undermine the whole value proposition of operating such a platform.
In my experience, the first step is to assess the underlying managers. This is not purely a technical assessment, but also a qualitative one.
- Does a given manager fit the overall platform and client objectives?
- Are they additive to the platform, or redundant?
- Which products and strategies do they manage?
- Do they exhibit strong relative performance to benchmarks?
- Are they growing or hemorrhaging assets, or simply treading water?
- What talent and skillsets are resident in the manager?
- Do the leaders and portfolio managers embrace more contemporary, data-driven processes or are they wed to legacy practices and bespoke EUAs that inflate operational and key-man risks?
- Which functional disciplines are working well, and which ones need improvement?
- Beyond investment functions, how effective are their operations, distribution, client service, compliance, cybersecurity, and risk management functions?
- Which tools and services are utilized, including applications, databases, programming languages, infrastructure, data providers, prime brokers, administrators, etc.?
- Is there a prioritized roadmap and backlog for new, or enhanced capabilities?
- Are there any legal, compliance, or regulatory hotspots?
- What do their DDQ responses look like?
- How is the operating budget allocated?
There are, of course, many other considerations. These findings may ultimately lead to a culling of managers, or even an expansion into existing strategies or new ones. Through a strategy of consolidation and divestiture, BNY Mellon Investment Management, for example, now houses only 7 managers.
ASSEMBLE AN INVENTORY
Second, it’s important to assess and create an inventory of capabilities across the full platform, including those that are both centrally provided and at the edge. This is in part a gap analysis measured against best industry practices, but also a way in which to identify those functions and components that may be ineffective or ripe for consolidation, development, vendor solutions, or outsourcing.
This inventory will also identify EUAs, like the ever-ubiquitous spreadsheets, that provide institutional-grade functions, but lack the appropriate levels of institutional controls, governance, maintenance, or documentation. To be clear, these applications may be a blind spot for many managers, yet represent significant fiduciary, operational, key-man, regulatory, and cyber risks.
The outputs from these assessments help to define a strategy and actionable plan, inclusive of objectives, timelines, staffing, and budgets.
At BNY Mellon, one such output of this assessment led to the creation and expansion of Asset Management Operations, a centrally provided platform that delivered a shared suite of technology solutions and operations services to the boutiques. While not all migrations were easy – due to a combination of divergent agendas, integration hurdles, and differences in methodologies – the platform ultimately delivered considerable savings to the overall investment management franchise.
DEVELOPING A PURPOSE-BUILT ARCHITECTURE FOR MULTI-MANAGERS
Finally, it’s critical to develop a data, integration, and workflow strategy that effectively bridges centrally provided services and edge services. Common application and data licensing terms can also be pursued. There will be certain functions that may lend themselves quite well to a shared model, whereas others are better maintained within the underlying managers.
Technology infrastructure is an obvious choice. Security master and other reference data services, for example, can also offer improved capabilities at lower cost than a federated alternative…but only if they adequately support the broad set of products, strategies, and data attributes required by the underlying managers. We also increasingly see the application of centralized, or even outsourced trading desks where execution is treated as a utility.
At BNY Mellon, we developed a multi-product/multi-tenant security master that provided such broad support. The platform was not only adopted by the investment affiliates but is now a core component in BNY Mellon’s middle-office outsourcing suite offered by the investment servicing side of the house to third-party asset management clients.
We also developed an innovative platform, Tapestry, that enabled the first of its kind, semantic integration. Here, integration pathways were derived at runtime based on the functional relationships between data providers and data consumers. For example, we know that an accounting system consumes trade data from a trading system. Based on meta-data describing this relationship, Tapestry was built to inject transport-level details (MQ, file-based, REST, etc.), along with data enrichment and transformation at runtime, ensuring the timely delivery of data in the consumer’s lingua franca.
DATA ARCHITECTURE IS KEY
An enterprise data architecture is especially important. This helps drive improved platform oversight, risk management, operations, analytics, and client/user experience. It’s also the most important prerequisite to the adoption of data-driven investment processes and the application of AI that leverages your own data sets. Ultimately, and especially as users become more technically competent, we want data to be delivered as a utility that provisioned users can access as needs dictate, much like a common electrical socket provides consistent current to any number of applications.
There are many design paradigms to consider, but I tend to prefer a data mesh for multi-managers with suitable technical competencies. This is a decentralized model that is largely self-service and allows the functional domain experts to deliver their respective data services as products, including associated APIs, meta-data, maintenance, and documentation. Standards and governance are still centrally defined, ensuring a consistent level of discovery, interoperability, and access.
For other organizations, a data fabric may be the preferred path. The fabric is centrally designed, integrated, and deployed, creating a cohesive view of disperse data assets across the full organization. This is not to say that larger, more complex organizations can’t also benefit from this paradigm, but such organizations are more likely to already have a federated operating model that befits a mesh.
While both approaches may utilize data warehouses or data lakes, efforts to consolidate data into a single warehouse typically fall flat. Invariably, they become so homogenous and difficult to change that few constituents are ultimately well-served. Similarly, lacking any governance, data lakes can quickly devolve into data swamps where data structures and attributes are known only to those that first deposited them in the lake.
The same is true when trying to force very different investment teams onto the same application. For example, an OMS may be very product specific: private market assets will have very different requirements than for public market assets.
The key is to embrace an architecture and patterns that facilitate a hybrid model of centralization and specialization. The capabilities that warrant centralization will emerge from the prior assessment and will offer a win-win for the franchise and underlying managers, alike. Even the capabilities that belong in the respective investment teams will adopt standards that allow reuse across the enterprise.
Here, I believe the carrot works better than the stick. Rather than trying to force-feed shared services down each investment team’s throat, it’s better to create a collective sense of ownership and proper incentives to drive more strategic, economical, and operationally prudent outcomes. Not doing so pastes a target on the backs of those trying to create such a platform, exemplified by little interest in manager adoption, blown budgets, and a flurry of complaints when things don’t work out as planned.
Further information about data architecture for asset managers is provided in this prior article.
CONCLUSION
Multi-manager investment firms find themselves at a critical juncture, balancing a delicate dance between diversification and economies of scale. The challenges and opportunities lie in the ability to deliver a heterogeneous suite of strategies and customized investor solutions at scale, further exemplified by a consistent client experience and fewer exceptions.
The pursuit of a shared-services operating model can deliver considerable benefits for the franchise, but must also deliver best-in-class capabilities that are well-tuned to the needs of the diverse set of underlying managers. The model should also recognize its boundaries and not try to be all things to all people, as this risks dilution of the exact diversification that platform managers (and, most importantly, their clients) covet. Investment teams should ultimately be incentivized, rather than forced to embrace the platform.
For those who navigate this space strategically, a promising combination of scale, efficiency, diversification, transparency, and superior client service awaits.
About Author
![Gary Maier](http://www.fintovapartners.com/wp-content/uploads/2024/07/Gary2048x1947.jpg)
Gary Maier is Managing Partner and Chief Executive Officer of Fintova Partners, a consultancy specializing in digital transformation and business-technology strategy, architecture, and delivery within financial services. Gary has served as Head of Asset Management Technology at UBS; as Chief Information Officer of Investment Management at BNY Mellon; and as Head of Global Application Engineering at Blackrock. At Blackrock, Gary was instrumental in the original concept, architecture, and development of Aladdin, an industry-leading portfolio management platform. He has additionally served as CTO at several prominent hedge funds and as an advisor to fintech companies.