Читать книгу Analytics for Insurance - Tony Boobier - Страница 20

CHAPTER 1
INTRODUCTION – THE NEW ‘REAL BUSINESS’
1.3 How Do Analytics Actually Work?

Оглавление

It is in the nature of insurance people to want to know how things are done. They want to understand how business intelligence happens from a technological point of view, how predictive and prescriptive intelligence works and what really sits beneath the covers in cognitive analytics. That is not to say that they want to be able to do it themselves, but rather in understanding the basic mechanics they are able to recognize the key issues and also the limitations of the technology. It will also help them in terms of the implementation conversation.

Let us start by saying that this is not a simple matter nor was the concept of Big Data and Analytics invented overnight. Rather that the insurance industry finds itself in today's analytical environment as a result of evolution, sometimes also an element of the step change and also from time to time, as a result of different thinking. In insurance which already has a legacy of analytical thinking as a necessary result of actuarial processes, new ideas increasingly find their way into the industry from other sectors such as retail or telecom. The use of ‘Smart Meters’ and predictive maintenance of machinery is already present – but how can this thinking be adapted and extended to the insurance sector? What comprises innovation for the insurance sector may be relatively ‘old hat’ for other industries. For insurance practitioners, it is critical that they maintain a 360-degree view of all that is happening in the wider world of analytics to be able to take full advantage of the opportunities before them, and then to be able to take that thinking and apply it to their own industry.

1.3.1 Business Intelligence

The starting point of any discussion regarding business intelligence arguably goes back to the concept of measurement and control. Without measurement there can be no control, and without control there can be no improvement. Such straightforward thinking found its way into the challenges of industrial productivity of the automotive and other manufacturing lines of the 1920s and later, and was subject to continuous refinement both in process and methodology. As organizations drove for increased profitability, the management of activity and its translation into activity-based costing (which identifies activities within a process and assigns cost to each activity) started to dominate, and up to the present day still heavily influence our thinking. Cynics reasonably argue that cost rather than value is being measured, and that the measurement process drives a quantitative rather than qualitative agenda.

But regardless, the essence is that measurement of operational activity in the form of performance metrics became prevalent and to a great degree remains so. What organizations have come to realize is that the metrics which drive performance improvement also drive changes in behavior, and that these changes are not always beneficial. Individuals measured against performance metrics often seek ways of manipulating data to show themselves in the most positive possible light, for example, in the case of sales progression. The psychological linkage between performance management and individual behavior cannot and should not be underestimated. To counter this, some organizations are also building behavioral traits into the assessment process, although like ‘soft benefits’ these behavioral traits may struggle to avoid a degree of subjectivity.

The topic of ‘conduct risk,’ that is, how we manage the performance and behavior of our sales people for example, becomes increasingly important especially in the shadow of Dodd-Frank and other consumer-orientated legislation. Analytical capability can sit behind the sales process not only in terms of sales performance management but also in the way that sales are conducted. If performance metrics drive behaviors (as might remuneration packages of sales staff) then analytics can be used as part of the solution in ensuring sensible behavior.

In essence, information collected can be assembled and structured to create management information. Historically this has been through tabulation but increasingly has been managed through spreadsheets. Information from outside the immediate organization, for example from the supply chain, can be obtained by ‘enforcing’ the supplier to provide information in a prescribed format as part of the supplier contract, so that information from many suppliers can be merged and consolidated to give a view of the broader environment. In other words, the procurement process can form one of the tools whereby suppliers provide information in a consistent way allowing the insurer to gain greater insight into multiple factors, such as cost, value and customer satisfaction drivers.

By collating this information from many suppliers as part of an RFI (‘Request for Information’) or ongoing process, insurers and others may find themselves with a clearer view of particular parts of the industry than some of the so-called expert vendors themselves. The challenge for such insurers is to recognize that knowledge in such circumstances is power, especially in a vendor negotiation process. Procurement professionals have recognized this for some time and use it ruthlessly in the negotiation process.

For many insurers, a spreadsheet approach to management information remains a critical capability but even spreadsheets have evolved. Those who were once experts by being adept at creating a pivot table now find themselves needing to be conversant with the advanced capabilities of spreadsheets with better visualizations and analytical capability. Spreadsheets, like the whole topic of analytics, continue to evolve and allow the user to have greater insight and improved visualization. The challenge for users of spreadsheets is most probably not only that of data capacity but also the increasing complexity of the business operation and its interdependencies. If it is argued that the insurance business is too complex to be managed either by intuition or experience (or both), then we are increasingly reaching the tipping point (if we have not reached it already) that it is also too complex to be managed by spreadsheet especially in the larger organizations.

Business intelligence is more than a form of enhanced management information. Rather it is a fundamental tool which allows insurers to understand if they are on track to meet their strategic objectives and where appropriate provide early warning signals that corrective action needs to be taken. If the sole purpose of executives is to ensure that the strategy of the organization is achieved, then it is critical that they have access to information which tells them, in indisputable ways, that the organizational ambition is on track or whether corrective action is needed. It follows that the metrics of business intelligence should align themselves to the strategic objectives of the insurance organization. There is no point in measuring things which are irrelevant to the strategy of the insurer.

The mood is increasingly for relevant information to be placed in the hands of the decision-maker, from a common source, so that, regardless of personal interpretation, there can be no doubt as to the source of the information and its veracity. This is often described as a ‘single version of the truth.’ To do this requires an information infrastructure which places data at the heart of the organization, centrally stored and accessible but with access appropriate to need and clearances. Many if not all business intelligence systems provide such capability which has become in effect a hygiene factor in creating a BI or ‘Business Intelligence’ solution.

The concept of a ‘single version of the truth’ lends itself to a data warehouse where information is held for the common benefit and is accessible accordingly. Such warehouses are usually in the form of an OLAP ‘cube’ of inter-related or ‘relational’ data and in most cases form the foundation of a business intelligence strategy for the insurance organization. (‘OLAP’ is an acronym for ‘Online Analytical Processing.’) Think of this as data held in a form of ‘Rubik's Cube’ rather than held in a flat form, and allowing the user (with appropriate access) to cut and slice the data according to their needs. In some cases, the data may be so large as to require the need for an intermediate data warehouse, or ‘datamart,’ so that the most important data allows more rapid interrogation.

Such an approach also demands a change in the function of the IT department who are also transforming from being the gatekeepers of the organization to the facilitators of information. They have a role in terms of the integration of systems and capabilities but as the demand for information increases it is critical that the IT department are seen as the ‘key enablers’ rather than any form of bottleneck. The days of waiting for reports from the IT department are rapidly disappearing. The most insightful organizations are transforming the function of the IT department into a role which fully supports the business – perhaps something they themselves have always aspired to be. The change is as much around culture and capability as it is around attitude.

Such an approach requires the IT department to be much more aligned with the business issues of the organization as with the technological requirements. It requires technology experts to sit with line of business decision-makers to understand their issues and to consider how technology can help an insurer reduce cost, improve profitability and reduce risk. It is a conversation about business issues converged with technology rather than forcing new technologies into business scenarios. Technologists increasingly need to understand that in the future (perhaps current) world of insurance, business and technology need to work absolutely hand in hand.


Конец ознакомительного фрагмента. Купить книгу
Analytics for Insurance

Подняться наверх