In an increasingly data-driven investment landscape, there is no doubt that integrated data, technology, analytics and visualisation are critical to deliver consistent, superior investment performance.
Cortex Investech is AMP Capital’s proprietary investech team, strategy, and platform responsible for supporting and enhancing investment decision making across the Public Markets business.
Dr Alistair Rew, founder of Cortex Investech and Global Head of Investech, Public Markets at AMP Capital talks to Fund Business about the solution, its origins and his plans for the future.
What is the Cortex Investech strategy and how does it improve investment decision-making?
Cortex Investech represents the integration of global investment experience, with the power of a technology enabled analytics and visualisation strategy. By bringing together ‘the human’ and ‘the technological’, we have created a solution that provides high frequency, high resolution, on-demand integrated analytics (portfolio, risk, return, exposures, performance and custom) that are tailored precisely to each investment team’s unique investment strategy, style and investment framework. This provides the foundations for robust, fact-based conversations that ultimately drive superior investment decision-making, from security selection, to position sizing, portfolio and risk construction, allocation, and integrated investment management.
As any firm about to embark on a data transformation project would know, reorganising data across multiple sources and streamlining it into an integrated analytics framework is easier said than done. How did AMP Capital approach this challenge?
It’s impossible to overemphasise just how important it is to ‘get the data right’, when embarking on any analytical-centric venture. Considering our ‘Insight Stack’ as a pyramid, ‘data’ and ‘technology’ sit side-by-side as the foundational stones, providing support to the entire structure (see image).
The next layer up is ‘analytics’, on top of that sits ‘visualisations’, and finally ‘insights’ represent the capstone of the pyramid.
Fail to pay attention to the foundational data layers and the entire structure and insight potential is undermined.
What this means in practice is that we need to fix data ‘at source’, which often means far upstream from where we sit. This is a radically different approach to that of the traditional, siloed culture of ‘local’ tactical fixes. It can be challenging but the fruits of fixing data at source, one-time, successfully, are beyond measure, as it enhances the quality and integrity of the entire business’ data layer.
As a solution, Cortex Investech is very much the product of a nimble and agile approach, rather than the result of a pre-defined business plan. Can you share some thoughts on what the architectural journey and stages of development looked like in practice?
When we started what we now call Cortex Investech in December 2016, it was just me and two founding members, who were working informally on it as a side project. We knew that the traditional approach of formal business plans and strategy would bog us down, delaying important progress, as well as anchoring us to something rigid, at exactly the time when we needed to be flexible. Instead, we adopted a start-up mentality.
In the early days, the technology build was completely driven by us, with modules built to provide us the insights we needed to deliver real value to our investment teams. Because our core principle was (and still is) to leverage data and technology to exact maximum value from the systems, analytics and calculators that we use, our first two hires were naturally experts in this space.
As we drove greater quantities of actionable insights from these systems, and demand on our services grew, we needed a business-orientated team member to drive our planning and prioritisation cycles, communicate with the business, and help with the strategy.
Just over 12 months into the journey, the basic tech build was complete and investment teams began requesting customised modules and analytics. This increased demand highlighted bottlenecks in two key areas: data optimisation and front-end development. The next two hires were experts in these fields; the first in database optimisation and administration, with strong relationships to the IT group, the second in visualisation and user experience design.
Today, our mandate covers equities, multi-asset and fixed income, requiring additional resourcing across data and analytics skillsets. Our new data expert will help build out and structure the complexities of our data stack, required to facilitate insights in multi-asset and fixed income, while the analytical expert will focus on getting the most out of our analytical systems, and developing new, meaningful analytics for these teams.
Today, Cortex Investech employs a team of ten, with expertise ranging from investment management, portfolio and risk management, analytics, data architecture, visualisation, automation, robotics, to software engineering and stakeholder management.
You clearly value flexibility and the ability to respond quickly to investment teams’ requirements. In that context, what is your view on insourcing versus outsourcing?
The build vs buy question is a timeless one, and they both have a role to play. We’ll generally focus our insourcing efforts on areas that require specialist knowledge of our investment teams and our broader business, our data and technology stacks, and where the types of analytics and tools we use are not easily accessible or provided through an outsource solution.
The other area where it makes sense for us to insource is in the customisation of analytics and visuals for the teams. This is something which cannot be easily purchased, or pre-canned from a vendor, for the simple reason that vendors typically provide mainly commoditised platforms. Our proximity to our investment teams and the close relationships we share with them, enable us to quickly and efficiently tailor many components which just wouldn’t be feasible with an off-the-shelf alternative.
On the flip side, many of the standard inputs to our integrated and customised offerings are derived from outsourced solutions. For us, this typically means traditional volatility-based risk calculators and performance attribution engines; there are many strong options in the market, which are very well supported. We are generally agnostic to which calculators we use, so long as they are mathematically sound and enable us to leverage their maximum potential in terms of the breadth, depth and frequency of the analytics they produce. We work these calculators in unconventional ways to generate hundreds of millions of data points each day, which we then integrate with custom analytics and visualises to create additional data sets, forming the basis of our Insight Stack.
User friendly visualisation is key to the appeal of the technology, with portfolio managers able to see near real time positions and risk exposures at the click of a button. Can you share some thoughts on how the user interface is constructed from a technology and design perspective?
For the most part, human beings are visual creatures, biologically hardwired to interpret visual information very quickly and efficiently. We are also tactile creatures, with touch-response mechanisms embedded in our brains. It’s from this biological starting point that we frame our design process, to make the transmission of the billions of data points we have at hand as seamless and as intuitive as possible.
It starts with placing yourself in the shoes of a portfolio manager, or analyst, and thinking about how they view their portfolio. This exercise informs the structure and placement of various analytics on each page.
Because investment management is multi-dimensional, we present portfolios along similar lines. It’s here that we introduce the notion of drill down navigation, which allows users to evaluate analytics at the portfolio and click through, or ‘drill down’ into, the portfolio through intermediate levels such as custom groups, regions, countries and sectors, before finally reaching the base level of securities, investment and financial data, custom scores, and factors.
A related navigation technique is the horizontal ‘drill across’, where users can flick between different analytics for any given level of the portfolios, including custom weights, risk contributions, expected returns, realised returns, etcetera.
We’re agnostic to the technology we use for the interface, so long as it’s fit for purpose. It’s typically written in Java script, and we leverage third party charting packages which can be further customised.
Your investment teams have clearly embraced the new technology, with a high rate of adoption among your teams. How do you cultivate that sense of ownership and technological adoption among investment teams?
The basic starting point for successful adoption is to recognise that each team begins the process with different attitudes, approaches, and needs. This calls for a Partnership Model, where we stand as equal partners, with decision enhancement at the core of the partnership.
In instances where initial openness may be low, we initially work with teams to create efficiencies across their core investment framework processes by using technology to develop reports and charts that may typically take an analyst days, often days per week, to produce. We find that most teams are open to this ‘efficiency approach’, as it frees up resource for them to focus on what they are really good at – analysis and portfolio management. Soon these teams are driving all new build requests, and that is when the adoption and partnership accelerates.
Looking ahead, what does the future of Cortex Investech look like?
We will be leveraging and integrating more and more advanced technological solutions such as voice, natural language processing, deep analytics, machine learning and advanced technology. The potential is mind blowing and we may even end up with virtual reality tools to improve investment decision-making. What we know for sure is that Cortex Investech will be at the cutting edge.