Alternative Data

Superior Insights at Scale

Inventing New Financial Concepts

There are quantities in finance which financial professionals have long desired but it has been hard to achieve. For example, many asset managers would like to know how much their assets are exposed to certain events or individuals. By knowing what they're exposed to they can position better and increase their profits significantly. Quantifying concepts like exposure is hard though. This is because, to put a value on this concept they need to know how each of their assets are related to the source (or sources) of the risk and by how much. To make things more complicated, some of these relations are anything but explicit. Consider the exposure of a company to a new political wave. While the company's target customer segments may directly be related to this new wave, there are other, more subtle relations between the company and the new wave which could strengthen or weaken the exposure.

For instance, the donations of this company and each of its leaders (if any) to the political movement (or its opponents) is an insider indicator showing us how sensitive the company is to that political movement. Some company leaders may be in good relations with the leaders of the new movement which could change the picture. Alternatively, they could seem unrelated but invest in common causes. Also, the foreign policy the new wave is advocating could either disrupt or strengthen the company's supply chain. We could go on as such waves can be very complicated in nature.
Putting all this together and coming up with a quantity manually is hard and extremely time consuming. Maintaining consistency to produce it for other events is even harder. And this is not just about one metric, there are other metrics which have been on everyone's mind in finance but hard to produce. A few examples are quantities like the Synergy and Merger Likelihood between two assets or Environmental, Social and Political Impact of assets.

Re-defining Old Concepts

In addition to new concepts, there are also concepts everyone have been using so far which have significant room for improvement. Take risk (beta) for an example. This is an age-old, price-based metric and quantifies the investment risk associated with an asset by comparing its returns to that of a risk-neutral benchmark index.
Knowing the relations that an asset has helps us to calculate the risk much better. This is because while assets have intrinsic risks they're also affected by the risk of entities/assets related to them. One other aspect of risk which can be improved using inter-asset relations is it's interpretability. While market professionals believe that "every" information about an asset is priced in, it is not clear how much each information has contributed to the price. In other words, price is an irreversible aggregation of the impact of each piece of information revealed about the asset. At the end of the day, it is just a number and one cannot decompose it to its main factors. On the other hand a risk metric calculated through reltionship-driven analytics is fully intepretable. It is clear which entity changed the risk by how much.

The same applies to correlation too. As it stands today, the correlation between two assets are high if their price charts are moving well with one another. However, when a rational person is asked to measure the correlation between two assets, the reasoning behind their answer will be quite different than just comparing their prices. They will start a reasoning process and that reasoning will involve a comparison on products, customer segments, dependencies, leaders of each company and much more. In other words, the person will look into the world that these assets live in, their dependencies (relations) and exposure. Then they will try to estimate a measure of similarity between them. Correlation data generated through analyzing inter-asset relationships is strikingly similar to this reasoning process.


The Solution

Our solution to this problem is both fundamental and ground-breaking. Our knowledge graph stores all the relations between each asset and the rest of the world while our AI algorithm extracts its meaning and quantifies all these metrics in a few seconds through automated reasoning. We don't throw historical data away either. They too are stored in our knowledge graph and used in the process.
The process itself is very similar to the reasoning process we all are familiar with. Only that, it involves much more information about the assets than our brains can handle. The assets in the knowledge graph are all interconnected. This means that the reasoning process invloves all entities of different natures (financial, political, social, ...) at once and creates metrics which are relatable and remain comparatively valid when judged by human intelligence.

Assets 360°

Sometimes nothing compares to having a visual of the world you're operating in. In finance, this translates to seeing an asset and how it is connected to the rest of the world. Maybe looking at these relations gives you an idea or perhaps it could serve as a validation of a hypothesis you've been sporting with for some time. Whether its clarity or validation, visualization of inter-asset relations brings a lot of benefits to your analysis. Acknowledging this, we have created a responsive, interactive visualization of our knowledge graph which allows the user to navigate through relations. Data and charts relevant to each entity is embedded within each node which helps the user to perform click-through analysis with a lot of ease. We call this visualization Assets 360°. It is also assisted by our AI technology as for assets under our coverage, the user can see various metrics generated for it. It is the best of both worlds.

Whether it is our metrics or the Assets 360° which interests you contact us to arrange for a demo.

Generic placeholder image

Founder & CEO

March 1, 2021