Note: To download this article in the original PDF research article format it was published as, please Click Here
The Niche Corner
This is potentially the most exciting area a Fintech can start in. Alternatively it is also the one that is most likely to lead nowhere. Companies in the niche corner are discovering things which 2-5 years ago wouldn’t have even existed, yet now the market is not only there but growing. A very prominent example of this are companies dealing in new and alternative data sources and analysis. One such company is Neudata, a provider of new and alternative datasets for hedge funds, discretionary and quant asset managers. Traditionally data used by these clients consisted of financial and market data provided by a few key players in the market data industry (Bloomberg, Reuters et al.) however with quant funds tending away from the traditional “quant” business of derivatives pricing and valuations that dominated a decade ago and into the new world of investment insight through data science companies like Neudata have now found valuable business opportunities in the sourcing and provision of that data. What Neudata additionally do which adds to the data value and impact, is rather than being just providers of data they also employ financial and data science experts which are able to determine any alpha potential (and even size of the alpha potential) of datasets, allowing the funds to potentially save valuable time in sourcing alpha generating data and implementing strategies around it.
What impact does this have on traditional, large asset managers? Well, as is being observed in the current low returns environment that traditional managers are struggling with, the one area in which dominates is being able to get and act on unique actionable data insights. What if you could predict the footfall of customers in a major FTSE100 retailer, or predict the current oil surplus floating across all the world’s oceans, or even track the spending shift of a countries customers by analysing spending on a granular transaction-level dataset? Well these are exactly the types of datasets that are available, or are able to be derived these days. By utilising the vast array of satellites circling the globe, the “exhaust data” from 3rd party apps that millions of people use on a daily basis or the growing stream of data coming from integrated sensors or the Internet of Things (IoT) Fintech startups are harnessing all of this data and selling it on as a clean and structured dataset which makes it simpler to analyse. Some startups are offering a data science layer on top of this data, allowing for fully actionable analysis (e.g. analysing the shadows of oil tankers in the ocean to calculate the mass of oil or natural gas they’re carrying, leading to a surprisingly accurate forecast of oil/gas surplus). With this informational advantage, funds can forecast important factors ahead of any official releases the traditional players rely on (e.g. being able to forecast the earnings of a corporate before their quarterly earnings report) which naturally puts them in a more actionable position. Neudata say that as far as the consumers of their service, none of them are traditional asset managers with the primary consumers being hedge funds, of which, they say, the US funds are the ones most advanced, being 5 years ahead of everyone else in this field. The reasons for this are varying, from restrictions in strategies imposed by clients on traditional players, to the technological incompetence of large asset management IT departments to deal with dynamic upstarts. Neudata say that most of the funds they sell to are able to turn a contract for the data they’re interested in around within 24 hours and have a fully functional strategy and technological framework in place for consuming that data within a few days. This is an orthogonal image to the timeframes of the traditional asset managers which take 6-12 months to just settle on the data contracts themselves and then provision a project on a timeframe of several years to be able to consume and take action on the data.
Whilst Neudata themselves have no plans on acting on the data they provide themselves, it’s easy to see how new funds could spring up with successful strategies based on this new data science way of investing, and when investors see that this environment of uncertainty and low returns is here to stay (as many believe it is) I can imagine even most traditional pension funds might migrate into new strategies.
The Middle-Man
Another area in which a lot of Fintech’s are currently appearing in is as an intermediary layer between B2B and B2C businesses. This is actually less of a disruptive stance on traditional asset management and more of a complimentary service to traditional asset managers. Startups such as Closir and Obsidian Solutions provide modern tools and an easier framework for services which traditional asset managers used to do manually, or with outdated and overcomplicated spreadsheets. Closir, for example, allows companies in emerging markets to connect with investors which they normally might not have had a chance to connect with, and gives investors tools to search out companies they might be interested in, in economies which traditionally did not offer much in the way of access. Obsidian Solutions offers something which a lot the intermediary startups are leveraging through their ability to create and deploy modern and innovative technological solutions from the ground up, which is a customer relationship management (CRM) system for sales and fund managers to build more concrete relationships with customers. These are technologies which in other sectors have been leaps ahead of asset management, but the traditional financial institutions are only just now starting to integrate the use of these platforms. Another example of intermediary Fintech solutions I have come across is the need for traditional institutions to communicate with each other but having out of date technologies and systems in place. There’s typically a very lethargic drive for the update of these technologies, even though it is well needed, and I don’t see this changing very fast, especially with most of the large institutions cutting back in vital areas such as IT and technology due to the margin and profit squeeze being felt across the industry. This however opens up the opportunity for startups to provide a funnel that’s able to ingest the data outputted by these legacy systems and integrate/transform this data into other systems to provide a greater overall functionality and flexibility to the client and consumer.
The New Kid on the Block
These Fintechs have one of the hardest and most ambitious goals, but also one that is most likely to disrupt the traditional asset management and wealth management industry. Their aim is to take the business from the traditional asset managers by offering better, more modern and timelier solutions for clients. Examples in this field include Nutmeg and MoneyFarm both of whom are a new breed of “robo-investment managers”. They are essentially a passive, or semi-passive fund manager for retail investors that allocates assets in a systematic manner based on a number of input factors chosen by the customer. The novel thing about these are that, as technology startups they have the huge advantage of being able to build on a modern framework of new technologies. This enables them to work to the today’s expectations from clients. Having a sleek and easy to use interface is a huge advantage that a lot of the traditional players are underestimating. As is having the ability to view real time analytics for a customer’s funds. Think about it this way; what other area of modern life would you be prepared to wait 1 month or longer to find out the status on something, only to then be able to see its value 3 weeks in the past? Modern apps and startups pride themselves on speed and real-time integration. Facebook pushes instant notifications to your phone, location apps are able to notify you in real time if you’re near something interesting based on your preferences and just this month Google was able to broadcast the winner of the 2016 presidential election on its election platform integrated into the google search page within just 30 seconds of the winning votes having been counted! With speed like this, waiting 30 days+ for a valuation of something as important as your assets seems madness.
Unfortunately the state of change within traditional asset managers on this issue seems to be almost stagnant (due in many parts to politics, lack of funds, unwillingness to undertake the risk of new technology and a general lack of capacity to implement technological change) which is where these new players to the game could slowly chip away at the business of traditional asset managers. And as they chip away, more and more startups will form and also start chipping away, all in unison at the large players, until suddenly they will have chipped away enough to collapse the iceberg very suddenly. The current state of these startups is such that their impact on industry is minimal. But it might not be for long.
The Dark Horse of AI
One final area, which I didn’t include in my original three areas of current Fintech startups, is the area of Artificial Intelligence. I didn’t include it for a good reason, as I believe from all that I have seen in the current state of startups 99% of those who label themselves as using “Artificial Intelligence” are not actually using AI, or are using it in such a limited capacity that it doesn’t deserve to bear the buzzword label. In my experience, there are very few key players in the world to date who are using AI in a full capacity, enough to warrant labelling themselves as an AI company. Those include manufacturers of robotics and self-driving cars, Google, Facebook and a select few others. The only area within asset management I have come across AI in a full capacity are top quant hedge funds, those which use AI alongside the alternative data mentioned in the Niche Corner section of this article through having the capacity to employ the best team of PhDs and AI researchers as well as give them use of the powerful hardware that’s needed to train highly complex AI networks. Those funds have fully realised the power of the knowledge they hold and hence keep their cards very close to their chest. Outside of that capacity however I’ve seen startups try to label themselves as AI based in capacities ranging from more “intelligent” customer relations to better “insights” into data. And whilst I’m not saying that these startups don’t provide intelligent and actionable value, especially when compared to the competition of traditional asset managers they are up against or selling to. A lot of the back-end functionality they claim as AI is usually part or fully manual, or at best utilises some traditional algorithms which could best be described as dumb machine learning (not an oxymoron).
However, I’ve mentioned this section in this article for good reason. I believe, alongside many of the top AI researchers and leaders in the AI field, that we are standing at the bottom of a rapidly growing exponential curve of development in AI. There are many papers and articles to be read which explain the revolution we are about to face from the utilisation of AI technology. I won’t divulge into that too much here, but from what we’ve seen we have gone from being able to classify handwriting to beyond human level in the late 1990’s, to being able to identify and label pictures to a beyond human level in 2015. Then from 2015-2016 Google (who arguably have the biggest and most complex AI research and capacity in the world) created AI [AlphaGo] which managed to achieve a feat long thought impossible by many which was to win games of the board game Go against the world grandmaster in a show of AI that acts and behaves far more human in the way of thinking than any narrow AI system (through the revolutionary use of reinforcement learning neural networks). Since then google has also announced it is now able to translate some aspects of some languages at a level beyond human translation. This has been achieved within only the last two years. So with this exponential rate of growth, I predict within the next 5-7 years the capacity of being able to apply real, broad and intelligent AI networks will be far easier than it is today for new startups to achieve. Those AI networks will also be far more intelligent than we can observe today and hence the impact they will have on every industry, especially asset management, could be significant if Fintech’s utilise them in the right way.
I will end on this closing thought, which has occasionally been highlighted and then quickly dismissed by a few people over the last couple of years: If Google, with their world leading capacity in artificial intelligence and coupled with the biggest and vast collection of structured alternative data they have access to, decided to focus that capacity away from advertising and user behaviour predictors and refocused it on asset management, they could have a very real chance to become one of the biggest players, if not the biggest player, in asset management in an inconceivably short space of time.