When you work with ICA, data reveals its secrets
An interview with Stephane Rio, CEO and founder of ICA, the Big Data analytics and visualisation fintech from Societe Generale’s Global Markets Incubator
ICA is a fintech firm which helps banks to achieve the highest returns, for themselves and their customers, on all the numerical data they capture. “ICA developed its data warehouse and analysis technology specifically for financial-services data,” says ICA CEO and founder Stephane Rio. “Because of this, clients can analyse large numerical and financial datasets with total flexibility in real time. This allows them to extract all the value locked up in that data. It helps, for example, to improve risk management, regulatory compliance or execution in the markets.”
What market problem does ICA address?
Storage solutions are not designed for real time analytical purposes. The data storage and processing technologies (typically Hadoop technology stack) which many banks use, were not built for financial and numerical data. Most (typically in memory) also weren’t designed for the massive data volumes we see today. As a result, they can be slow to retrieve data, may need expensive hardware to hit their required performance and may not be able to provide access to all the data a project needs within the required time frame. Given the huge data volumes generated by banks today, this heavy approach adds ill-afforded time and prohibitive cost.
What does ICA do differently?
In summary, ICA empowers the user to have an efficient access to 100% of the data in order to leverage their embedded values.
To achieve that, primarily, the data is stored differently. ICA uses a massive parallel-processing database that is columnar and leverage array types. The data is organised using a data model specifically designed for financial-services data and has been built to enable powerful analytics. To give a simple analogy, using traditional databases to enable financial-services analytics is like trying to do a calculation with Microsoft Word. You don’t do that. You use Excel.
Another way in which ICA has an edge over the traditional data solutions used by financial institutions, is in the easy access to the full amount of data in contrast to traditional solutions, which can usually only deal with part of a dataset at any time due to the technical and hardware costs limitations. When a company works with ICA, any non-IT user is empowered through an intuitive front end and a low code API. We expose only the business data model (the one the user understands), not the physical data model (the one that is hard to understand if you are not intimately involved on the IT side) and the underlying complex queries made to the database. It allows the user to easily access all the data they require and perform powerful analysis.
Finally, this helps to concentrate on the financial industry matters, without engaging in technical discussions with our clients, mainly focusing on use cases such as market risk (Including new complex regulation like FRTB, counterparty credit risk (including risk weighted asset calculation), liquidity management, wholesale credit, pre-trade execution optimisation, retail marketing, Solvency 2 for insurances, etc.
How do traditional data technologies limit banks and how does ICA overcome this?
Banks typically have made compromises in the way they can use their data. Sometimes the technologies they use require to pre-load large volumes of data into memory, sometimes they require preemptively to prepare the data. This forces banks to make assumptions on what data users might need and how they will want to use it.
Moreover, both methods come with serious drawbacks. Pre-loading data into memory involves buying large quantities of memory, which is very expensive. Trying to predict what data users will need is an imperfect science and often leads to data being unavailable when users need it, because it has not been prepared, pre-fetched, pre-aggregated or indexed.
What are the limitations of these traditional data handling approaches?
As mentioned, using in-memory solutions is very expensive: the cost of hardware of our solution is typically 20 to 50 times lower. For example, we worked with Societe Generale on one single use case to save annually several millions of euros of hardware.
Apart from the expense, the problem with these methods is that the data available is always a subset. Typically, banks will load in memory a limited history (a few days) and when using prepared data, they will expect to cover about 80% of queries users will perform. In a highly competitive arena such as capital markets, that simply isn’t enough. If you’re the bank that’s able to access 100% of its data in real-time and analyse it anyway, you hold a competitive advantage.
What value will the ICA approach bring to financial institutions and their end customers?
Our dedicated focus to the financial industry allows us to deliver very concrete use cases in a wide range of domains which helps them to improve and automate their processes, ultimately resulting in cost reductions and improved service to their customers.
Does this have any positive environmental implications for clients?
Compared to other database and data-analytics solutions, ICA is far less power hungry and requires far fewer servers per data volume to run. That means a smaller environmental footprint, which helps ICA play its part toward meeting the banks’ ambitious environmental targets, like those set by Societe Generale.
What is ICA’s relationship with Societe Generale and how has that relationship helped ICA?
At the end of 2018, and following a rigorous selection process, we were proud to be one of only six fintechs chosen by Societe Generale to participate in its first-round fintech accelerator programme, the Global Markets Incubator.
It was a wonderful experience for us. We needed a live example to showcase to potential customers what we could achieve. Working with Societe Generale for six months, we delivered our product to end users (traders and risk managers) in the bank, fulfilling real needs and helping them to extract greater value from their data. This proved that our model was useful. We then went successfully through an in-depth technical due diligence and a benchmarking exercise with the IT department and it convinced Societe Generale to invest in ICA to build a long term relationship.
What is your vision for the future and how does this align with Societe Generale’s ambitions?
We take our social responsibility and sustainability contribution very seriously. ICA’s model is not just about making financial services more environmentally sustainable, important though that is. It’s also about achieving other types of sustainability. We want to enable organisations to work more intelligently with today’s data volumes, for instance, allow the industry to become more resilient, manage better their liquidity and capital.
This makes it easier to extend loans and risk-aware products, which in the current context is required more than ever. Banks and other players will also be better able to withstand future shocks to the global financial system. That is a kind of sustainability worth working towards. It also aligns well with Societe Generale’s beliefs in bringing maximum efficiency for customers and commitment to supporting clients and partners in achieving sustainable goals.