In many industries, points of access to goods and services are all rapidly becoming digital. Take the example of film distribution: first consumed only via cinemas, then television and video tapes, later DVDs and now through streaming. An exponential increase...
Today, success depends on the ability to have or interact with digital points of sales, indeed in some sectors it may be the critical element. Consumers expect to make purchases or have services online in a seamless and simple manner. This does not just benefit buyers, companies that focus on digital points of sales find that their ability to better understand their clients and to create value is vastly improved.
In a digital world, the one component that links both is data. Improving and maintaining the quality of the data and delivering data efficiently is therefore fundamental.
The investment fund industry is not immune to this shift and expectations of today’s fund buyers are high because of their online experiences in other areas. Fund players understand that the product creation and distribution ecosystems must change. Digital transformation, big data techniques, machine learning and artificial intelligence will enable them to vastly improve the value chains and to have a strategic and real-time overview of their businesses. For this to happen effectively data quality is of paramount importance.
But all too often, much time and resources are wasted, because ensuring data quality has been overlooked or put aside to the future in the rush to other priorities.
It should go without saying that if the data produced is unreliable then actors in the ecosystem will have difficulties implementing digital transformation initiatives on their own. But all too often, much time and resources are wasted, because ensuring data quality has been overlooked or put aside to the future in the rush to other priorities.
Ensuring data quality with advanced analytics
Investment fund data and data sets cover a wide range and are from a variety of actors. In the fund industry, data producers are also data consumers and data flows from one to another within a complex network. This network includes fund buyers, fund distributors, transfer agents, fund administrators, asset managers and many others.
As for the data itself, it may be structured or unstructured, quantitative or qualitative, human or machine generated and either infrequently or very frequently updated. It may also be needlessly duplicated and not standardized. Of course, any part of these records can be corrupt or contain errors or be incomplete.
This is when big data techniques can be applied. Here, by big data we do not mean dealing with extremely large amounts of unstructured data to discover insights and predict trends, but rather using analytics to ensure that the basic data underpinning the European fund and asset management industries is as accurate and robust as it can be. This is also when data lakes, which are for storing all types of data (structured, unstructured, etc.) and for supporting all types of users, can be used most effectively.
Mutualise for value creation
Checks are important in this complexity and we can look at a simple but revealing example from Fundsquare. A file that is updated daily may contain 300 data fields and 1,000 lines of products and therefore 300,000 data points. For each of these, ten controls must be performed, meaning 3 million data checks. Such a file may be from only one fund company and so with 100 such files in a workday, this means 300 million data checks. Clearly this beyond human capacity and big data analytics is ideal.
For many companies, verifying and guaranteeing data quality are decade-long programs that have long been imperfect to varying degrees.
The fund industry is seeing the sheer amount of data and information increase rapidly as well as the expansion of digital points of sales, with the attendant control and checks growing almost exponentially.
For many companies, verifying and guaranteeing data quality are decade-long programs that have long been imperfect to varying degrees. They are costly, labour-intensive and people are not as accurate as machines. Additionally, data quality is often seen as the preserve or responsibility of a particular team, usually the IT department.
Mutualised data quality processes in combination with a holistic information management approach is the way forward for fund companies who want to reduce internal costs and create real value.
Continually enhancing and enriching data
Data collection, data preparation and data cleansing can be automated by using big data methods and machine learning, in whole or in part. More importantly, data analysis and standardisation can be carried out in these early phases. It can also be enriched from other sources.
The focus is on data production and, therefore, before the data has even reached the data consumer, it has been prepared for optimal use for the user, who might be the final investor or fund producer or any other entity in the fund distribution chain.
Low quality data carries with it many risks.
Furthermore, this is an iterative process that runs in real time, with feedback from the data consumer to the data service provider, who can then enhance or update the information. This enables industry-wide innovation in products and services and provides strategic and real-time overviews of the relationships between actors in the chain.
Low quality data carries with it many risks. In addition to the obvious reputational risk and compliance risks, there is the very real possibility that decisions will be taken and acted upon based on unreliable data, thus impacting operations. Companies must master this risk but there is low value creation in doing so.
Holistic management will, ultimately, create value for better analytical approaches by players in the fund distribution chain and will support an effective digital transformation in the investment fund industry.