Q:Adrian, tell us a bit about yourself and Dynamo Analytics?

Adrian: I co-founded Dynamo Analytics after a 15 year career building and implementing financial and statistical models for global insurance companies. I’m a qualified actuary with diverse practical experience and have established and led technical teams in both the UK and South Africa. As a business, Dynamo Analytics employs actuaries, mathematicians and software designers. This mix of skills means that we have a unique ability to operate across the whole spectrum of model conceptualisation, design and operation, including validation. We help clients use and industrialise detailed analytics, processes and predictive modelling in a real world, business-as usual context. 

We have a presence in the Nordics through our office in Stavanger (Dynamo Analytics Nordic AS), in the United Kingdom through our office in London, and in South Africa through our office in Stellenbosch.

Q:The key topic of discussion these few years has been on how to harness the full potential of data. Can you share your thoughts on the subject?

Adriain:We have helped a number of organisations harness the power of their data by industrialising their financial and analytical modelling processes, and so change their process from slow, fragile and opaque to tightly integrated, fast, robust and collaborative. A couple of examples:

- Reporting Model: We used our software environment Psicle to build the actuarial and statistical reporting process for an insurance company. Psicle integrates tightly with their policy and claims administration systems, as well as their planning, assets and expenses data sources. This means that the moment these systems close, the data is transferred and the complex modelling begins. This modelling process takes the raw data from source systems and performs the actuarial and financial analysis in an automated fashion, producing the output within minutes rather than days or weeks. The internal team can then, by exception, examine and tweak the modelling results. Psicle then integrates tightly into the company’s financial and reporting systems, automatically delivering the results of the modelling back to the business. 

- Analytics Model: Psicle has been used as the analytics engine for a retail bank. An extremely large dataset of their client survey data, as well as other demographic data, was loaded into the Psicle environment to apply a variety of generalised linear models and other statistical tools to the data. This helped highlight the ‘surprising facts’ and hidden causal factors, and meant that scenarios and what-ifs were easy to run. Psicle’s tight integration with source systems, and robust and fast architecture, make it ideal for large scale, repeatable scenario generation and reporting. 

Q: From your perspective what are the biggest or the most common hinders or challenges organisations face when it comes to Big Data and Data Driven Innovation? And what would be the implications of not solving this challenges?

Adrian:We are entering the era of big data, sophisticated algorithms and rapidly emerging technology. We are collecting more data these days than we know what to do with, and are starting to understand how to build and use machine learning techniques within businesses. Financial and analytical modelling is moving to the heart of how businesses are being managed and assessed. The problem is, these models are not being built in platforms that are fit for purpose. The result is that boards, regulators and management are being asked to rely heavily on models that are slow, that are fragile, and that are opaque. 

One of the key reasons that businesses find themselves in this situation is that they have used legacy systems because that’s what was available at the start of the project, or tactical business applications (Excel and the like) because it was easy. However, these aren’t suitable for large scale business critical financial modelling. The key issue is that these business models are still essentially being run as proof of concept projects, not mature processes, and this needs to change to get executive buy-in and for the results to be reliable. 

Q:For those organisations that just started their journey in Data Driven Innovation, what would be your recommendation for speeding up the process of turning data into insight and action? 

Adrian: Organisations need to take a mature, project based view on understanding and working with their data. While tactical, on-the-fly analytical and modelling will always have a place for unusual problems, or for scoping a solution to an organisational question, we believe the days of these solutions being used for regular, robust, business-critical financial and analytical modelling are gone. 

We have developed a unique and powerful general financial and analytical modelling environment, which will be at the heart of how businesses will run their business-critical actuarial, statistical and financial models. Psicle, our model industrialisation platform has been developed over several years, with the aim to revolutionise the way businesses approach their data modelling. Psicle is taking large modelling processes that are slow, fragile and expensive and transforming them into fast, robust, auditable, transparent and reliable processes. We have been using Psicle to solve a wide variety of business challenges that would be difficult or logistically impossible in standard environments. The challenges have ranged from simply managing and manipulating data in a robust and secure manner, to applying leading edge statistical processes to business data to derive actionable insight. 

Q:How big part technology plays in this?

Adrian:We feel technology will play a huge part in data driven innovation. Stakeholders are increasingly expecting organisations to engage in more complex financial and statistical modelling of their business. But simultaneously they need to deliver these analyses faster and more robustly, with fewer errors and resources. We believe that the key to the future of business profitability and sustainability lies in the industrialisation and automation of financial and statistical processes. The only way to address these two competing demands is to use technology to do the heavy lifting. We have therefore invested heavily in building Psicle, our model industrialisation platform.  Q:Considering the amount of data, the technology available, the competence and the new/old processes in place, where do you see the biggest impact of Big Data and data-driven innovation? 

Adrian:  There has been an explosion of data. Every day we are collecting more data, from more sources, with greater variability and complexity. And this data is increasingly driving how businesses are being managed and assessed. For example, in insurance, big data is transforming how insurers price risks and manage customer relationships. More granular data is increasing the transparency of risk behaviour, enabling more accurate pricing and better risk management. Connected homes and the Internet of Things are providing intelligence previously unavailable. For example, through a smart home connection, insurers can encourage risk mitigation and anticipate claims earlier. The application of actuarial techniques to this expanded data set is enabling the derivation of deeper insights, helping highlight those ‘surprising facts’ and hidden causal factors, upon which businesses can make smarter decisions.

Q:Looking at the data-driven innovation from technology point of view, what can we expect in the next 12 months.

Adrian:We expect to see significant development over the next 12 months in the technology available and the depth of the analysis possible. However, the biggest area of improvement we see is the integration of analytics, that is, businesses taking the cutting-edge modelling that they are doing, and beginning to integrate these models into mature enterprise processes. Only through this integration can data and analytics become a core part of the business, and the models something that executives and stakeholders can rely on.