Data mining is the same. <\/span>Raw data\u00a0needs\u00a0processing,\u00a0cleaning, and\u00a0polishing\u00a0\u2013 conforming\u00a0into\u00a0recognizable and inter-relatable\u00a0structures.\u00a0Only then can it packaged in a report and put on the \u201cmarket.\u201d\u00a0We know this process\u00a0as enrichment.<\/p>\nWhat is Enriched Data?<\/h2>\n
Many insurance companies have more than one source system.<\/span>\u00a0<\/span>Over the years, they<\/span>\u00a0acquired other companies that use a different system for policy, claims, or billing.<\/span>\u00a0<\/span>Many systems, that have run their company for a decade or longer, do not keep up with the insurance company\u2019s pace of change.<\/span>\u00a0<\/span><\/p>\nNew products, new processes, and poor performance under an\u00a0<\/span>ever-increasing<\/span>\u00a0load<\/span>\u00a0<\/span>force<\/span>d<\/span>\u00a0companies to\u00a0<\/span>migrate to new platforms.<\/span>\u00a0<\/span>Even within the same systems, errors\u00a0<\/span>are<\/span>\u00a0identified and fixed, but the existing dat<\/span>a<\/span>\u00a0still reflect<\/span>s<\/span>\u00a0some of these errors.<\/span>\u00a0<\/span>The data is not the same across all systems<\/span>,<\/span>\u00a0and within each system<\/span>,<\/span>\u00a0there remain<\/span>\u00a0data anomalies.<\/span>\u00a0<\/span><\/p>\nHow can you have a\u00a0comprehensive\u00a0report if the underlying data is in different\u00a0structures with disparate codes and questionable quality?\u00a0You must enrich the data.\u00a0<\/strong><\/p>\nThe Value of Enrichment<\/h2>\n
There are many ways to discuss value.<\/span>\u00a0<\/span>Let\u2019s talk about the cost of not\u00a0<\/span>correctl<\/span>y enriching the data.<\/span>\u00a0<\/span>An insurance company\u00a0<\/span>looked<\/span>\u00a0into agency performance.<\/span>\u00a0<\/span>They set up criteria to measure which agencies\u00a0<\/span>brought<\/span>\u00a0in the most profitable business and which agencies\u00a0<\/span>cost<\/span>\u00a0the company money.<\/span>\u00a0<\/span>They identified several agencies\u00a0<\/span>writing very few P&C policies<\/span>,<\/span>\u00a0and in general<\/span>,<\/span>\u00a0the loss ratios were very high.<\/span>\u00a0<\/span><\/p>\nThey decided to put these companies on a development path.<\/span>\u00a0<\/span>There were some rather harsh adjustments recommended.<\/span>\u00a0<\/span>Seems like a good use of data, right?<\/span>\u00a0<\/span>Not so fast.<\/span>\u00a0<\/span>Let\u2019s close this loop.<\/span>\u00a0<\/span>T<\/span>he company\u00a0<\/span>failed to<\/span>\u00a0realize\u00a0<\/span>their data warehouse did not include its L&A book of business.<\/span>\u00a0<\/span>Several of these\u00a0<\/span>\u201cpoorly\u201d performing agencies were not\u00a0<\/span>actively\u00a0<\/span>marketing P&C insurance.<\/span>\u00a0<\/span>They\u00a0<\/span>only sold<\/span>\u00a0P&C insurance to existing L&A clients as a service to keep the more lucrative L&A business.<\/span>\u00a0<\/span>One<\/span>\u00a0of these \u201cpoorly\u201d performing agencies was the largest, most profitable L&A agency the company had.<\/span>\u00a0<\/span><\/p>\nThe lack of proper enrichment led this company to conclusions that were incorrect and put a very valuable relationship at risk.\u00a0<\/strong><\/p>\nThe Cost of Enrichment<\/h2>\n
Enriched data is good.<\/span>\u00a0<\/span>Enriching data is hard.<\/span>\u00a0<\/span>In this blog series<\/span><\/a>,<\/span>\u00a0we have discussed some of the risks and challenges.<\/span>\u00a0<\/span>Data governance drives the enrichment process.<\/span>\u00a0<\/span>Data governance is <\/span>a\u00a0<\/span>load for the business side and<\/span> the technology side.<\/span>\u00a0<\/span>For every source system,\u00a0<\/span>every object,\u00a0<\/span>every attribute, and\u00a0<\/span>every measure there are\u00a0<\/span>multiple<\/span>\u00a0steps and\u00a0<\/span>many<\/span>\u00a0people working together to ensure accuracy and\u00a0<\/span>comprehension.<\/span>\u00a0<\/span>This process is not cheap.<\/span>\u00a0<\/span>Nor is it fast.<\/span>\u00a0<\/span><\/p>\nTo ensure proper review and thought, data cannot\u00a0become\u00a0enriched overnight.\u00a0<\/strong>Companies without a careful process might find their properly enriched data co-mingled with improperly enriched data that cause a complete loss of trust in the entire data set.<\/p>\nThe Future is Hard to Predict<\/h2>\n
From where will tomorrow’s challenges emerge?<\/span>\u00a0<\/span>Unfortunately, there is no crystal ball.<\/span>\u00a0<\/span>A down<\/span>turn in the economy may turn your primary three<\/span>–<\/span>year thrust from expanding into new ventures toward lowering costs\u00a0<\/span>or<\/span>\u00a0cash flow belt-tightening.<\/span>\u00a0<\/span>While both are great goals, only one\u00a0<\/span>is<\/span>\u00a0the\u00a0<\/span>primary driver<\/span>.<\/span>\u00a0<\/span><\/p>\nBecause the future is hard to predict, data governance –\u00a0that drives data enrichment and in turn\u00a0drives formal reporting – must remain flexible.\u00a0<\/strong>Also, the value propositions for finding new ways to glean information from raw data facts is not a straight road.\u00a0A<\/span>n<\/span>\u00a0insurance executive was talking the other day about how the idea that\u00a0<\/span>\u201c<\/span>credit scores could predict claim<\/span>\u00a0costs\u201d\u00a0<\/span>was not universally accepted when first introduce<\/span>d<\/span>.<\/span>\u00a0<\/span>Many insurance executives<\/span>,<\/span>\u00a0at th<\/span>at<\/span>\u00a0time<\/span>,<\/span>\u00a0did not understand how the credit score\u00a0<\/span>relates<\/span>\u00a0to if an incident would happen, or how severe it would be.<\/span>\u00a0<\/span><\/p>\nBut now we know that credit score is a very\u00a0<\/span>useful\u00a0<\/span>predictor<\/span>.<\/span>\u00a0<\/span>How many other ideas didn\u2019t pan out?<\/span>\u00a0<\/span>How many\u00a0<\/span>were other<\/span>