Implementing technology in the Investment Management industry
Further to big data and application programming interface (API), consultants and futurists are predicting transforming changes due to artificial intelligence (AI), blockchain and distributed ledger technologies (DLT), and robotic process automation (RPA) in the financial industry. Different position are taken by investment managers regarding these new technologies: the watchers, who try to stay informed of changes; the testers, who launch minimum-viable-products (MVP); and finally the doers, who incorporate technology within production. What are the obstacles that prevent a watcher from beginning to test, and a tester to switch to production?
Predicting the future
Futurists have regularly made predictions that have proved(1) to be wrong, regarding telephones, cars, planes, the atomic bomb, computers, space travel or the year 2000. The end of unskilled work was forecast a century ago after the introduction of machine tool automation. And driverless cars, not to mention flying cars, were already on the front pages of national newspaper in the 1930s.
Technology may come to life without it having an immediate clear application. The fax machine was invented in 1842, well before the telephone in 1876, and it took over one hundred years before it became commonplace. More recently, it took nine years to begin using lightly sticking glue for sticky notes ("Post-Its").
Further to big data and API, consultants and futurists are predicting transforming changes due to AI, Blockchain and DLT, and RPA in the financial industry.
Traditional change analysis is split into
internal impacts usually leading to work reorganisation, process transformation and cost savings;
external impacts tackling client demand and breadth of offering;
community impacts involving common infrastructure to improve global security and market efficiency.
Digitalisation bridges the gap between internal and external processes. It enables front-to-back processes with end clients directly triggering production, straight-through-processing (STP) in the financial industry.
At the same time, with the introduction of the Internet, radically new business models underpinned by network effects2 have appeared with: open data (OpenStreetMap) and open software (Linux, mysql) for which consulting firms can promote their expertise, free, usually with hidden revenues from advertising (Facebook) or data (Gmail), freemium where a small number of users who want an advanced service pay for most users who are using the generic service for free (online gaming), or collective (crowdfunding).
Clearly, the introduction of blockchain with bitcoin could open radically new organisations with the creation of trust without intermediaries between users.
Very few applications are already live in the financial industry. Much less disruptive, robotic process automation (RPA) is a technique, rather than a new technology, to lower the operational risk that arises from the local automation introduced by operational teams to improve their efficiency using a variety of tools, including Microsoft Excel and VBA. These automation scripts are very sensitive to any change in data model or user interface. RPA consists in choosing a single IT infrastructure to industrialise the development of these “macros” and to have them maintained by the IT team rather than the operational teams. Only the less versatile macros are eligible for RPA.
API and data are commonplace
APIs are not recognised as a new technology anymore. They offer the amazing possibility of extending the functionalities of an existing information system, provided the resilience of API provision can be managed.
Data, whether big or smart, has become a buzzword even though the value of information has been known for centuries. The difference lies in the current possibility of handling large amounts of data. Data previously deemed insignificant can become valuable.
Three positions towards new technology
Leaving aside consulting firms, freelancers and futurists, different positions are taken by investment managers regarding new technologies: the watchers, who try to stay informed of changes, the testers, who launch minimum-viable-products (MVP), and finally the doers, who incorporate technology within production. There are only 9% of doers, according to the SGSS’ “Taking the Long View” survey, while the watchers represent 43% and the testers 48%. So the question worth asking is how to move from watcher or tester to doer.
If new technologies have all the merits that the futurists are describing, what are the obstacles that prevent a watcher from beginning to test, and a tester to switch to production? Some consultants are as bold as to say that the investment management industry is lagging behind by five years in the implementation of new technologies compared to other industries.
There is no real financial obstacle to begin testing new technology. All these technologies are low barriers to entry: knowledge is widely available, hardware systems are not expensive and cloud technology is a way to pay per use avoiding investment, and many students are happy to put what they learn in ‘real’ life into practice. In addition, there is what consultants are calling ‘servicialisation’: missing components can be used as a service from third parties through APIs.
As it is becoming easier to test, it is also becoming faster. This makes it possible to shorten timeframes to select the tests to continue and those to discard. So there is no excuse for investment managers not to start testing new technologies.
Difficulties for Doers
Going from test to production is tricky, as it requires mastering security. However, the ratio between the testers3 (48%) and the doers (9%) is over five, according to the SGSS survey. This means that many projects do not find their business case, and there can be many different reasons for this. The obvious ones are linked to a lack of process mastering: availability of business experts and poorly-modelled processes, and to resistance to change.
The other reasons relate to excessive expectations regarding the technology: poor assessment of technology’s limitations. New technologies should be seen as additional tools in the toolbox rather than a Swiss penknife. New technologies should not be considered to be solutions in search of problems. Further to these business analysis problems, difficulties may arise from laws and regulations. In AI for instance, machine learning is a terrific analysis tool to support a human decision, but it has shortcomings when the logic of an automated decision must be retrieved and explained to a regulator.
Without going so far as to celebrate failure, taking a failure as an investment, learning from this failure and pivoting the business model, are strong factors of success, as long as there is a purpose.
The doers seem to have in common a precise knowledge of their business and of the technological kit that they are using. Technology is only a medium to achieve their purposes. Doers share a very practical approach to technology and they incrementally introduce innovations with a quick turnaround time. This does not mean that their business will not radically change as the futurists are predicting. It means that their approach is to take only one small step at a time so that they can take one giant leap. Go for it, test, accept errors and pivot, or stop early?
1. Pan Am took bookings for commercial flights to the moon from 1968 to 1971, with a first flight expected in 2000.2. Network effects are explained by Robert Metcalf: the value of a network is proportional to the square of its users, and David P. Reed: the utility of a large network can scale exponentially with the size of the network. 3. www.securities-services.societegenerale.com/uploads/tx_bisgnews/LONG_VIEW_MAGAZINE_FINAL_DEF.pdf