Data Science is disrupting business—as surely as electricity,telephony, connectivity and mobility disrupted it. That's astatement insurers find in most every trade publication they read.It's often much closer to home: many have to look no farther thantheir own organization to see the role data science plays in growthand profitability. Marketing and underwriting, for instance,habitually analyze data to target advertising, price policies,assess risks and more.

It's not universally adopted though. Many claims organizationsin particular have yet to take full advantage of the value of allthe data they routinely generate and store each and every day.Depending on the scope and reach of your data network, that caninclude:

• Customer data.
• Performance data.
• Historical data.
• Competitive data.
• Industry data.
• Partner data.

Too often, this data is left untouched: the rich knowledge itcontains never harvested and analyzed.  There are a numberof reasons for this. There's the investment involved, first inequipment (or Cloud services) and the operational costs to manageit. The second, and more significant, investment is in people: thedata scientists and senior business analysts that uncover andextract value from data. There's often a cultural resistance—doingthe status quo seems like the most attractive alternative todisruptive change. And that resistance might reach all the way upto fundamental corporate philosophy—which limits executive supportfor the effort. The tendency to place external, urgent goals overlong-term business goals can halt data analysis initiatives intheir tracks. Fighting fires isn't what data science does; datascience prevents them. Data informs business. Data sciencetransforms it.

Still, no matter the reason, many if not most claims executivesare at the starting gate when it comes to getting the most out oftheir data. But now is definitely the time to get data sciencegoing if you're interested in realizing some of the potentialbenefits, including:

• Improving the accuracy of initial appraisals.
• Reducing time and money wasted in the claims processing.
• Establishing industry-wide benchmarks for assessing individualperformance.
• Creating predictive and prescriptive frameworks to improvedecision making.
• Opening the door to improved customer satisfaction andincreased customer loyalty.
• Enabling "small data" [1] (how to humanize data and respond tothe individual). 

Find Your Place on the Data Science InnovationCurve

Today, insurers that are using data science to drive specificprocesses are realizing excellent results, but for most, it hasn'tbegun to deliver its full value across the enterprise. CCC findsthat as you look at the value of data [2] and combine the value ofbig data with value of small data you can disrupt your products,your processes and even your business model.  Pinpointwhere you are on the Data Science Innovation Curve below; the moreyou can leverage deep historical and predictive data and apply itin an individual way in real time with the customer, the moredisruptive you can be.

Keys to Claiming Your Data

To be sure, change is required in order to embrace and benefitfrom data science. This change doesn't need to be overly costly orcause you to entirely reinvent your business, but it will need toinclude three key components to be successful.

Strategic Alignment

To get the full value of data science, there has to be astrategic alignment across the business. Field offices, regions,corporate offices, and partners have to be on the same page. That'seasier said than done. As you implement new processes driven bydata science, your people will have to learn to do things in newways and to operate under new standards. It's not compliance you'relooking for though. You want more; you want commitment. You want anengaged team asking the creative questions that improve processesand accelerate growth. 

While the debate about artificial intelligence is ongoing at thehighest levels of philosophy and science, we find that in mostcases it is the combination of the human and the machine thatprevails, not one alone.  Still when scale and speed areparamount, a fully automated solution could be in the future.

Dedicated Resources

You have to dedicate resources and money to this—data science isnot a second-level priority, using shared and borrowed resources.You need to dedicate an infrastructure on premise or in the Cloud.Data Scientist is not a part-time job: it's a strategic role inyour organization. Data Scientists understand your business, yourworkflows and your goals, and then uncover ways to improve themthrough data analysis, actionable insights and prescriptiverecommendations. The demand today for Data Scientists isskyrocketing. Leading researchers, including McKinsey GlobalInstitute, predict that by 2018 there will be more than twice asmany openings for data scientists than available talent.[3]

Executive Support

Executive support is a must-have. The trick here is not to tryand do too much at once—to stay focused and above all patient.Think of which specific processes in your business will show thequickest and most impressive results.  A quick win forexample could be a situation where you already have a tool toselect one of many alternative processes.  Making thattool "more intelligent" allows data science to shine without thecost of developing systems to deliver the model to the user, sinceyou would merely be swapping out the decision engine.

Moving First or Moving Last Decides Winners

If investment in a dedicated team is too much too soon, or yourdata isn't broad or deep enough or its buried deep in legacysystems, then working with an established provider canwork.  But getting started now is critical as the historyof business is filled with leaders who didn't recognize and takeadvantage of disruptive technology.  Forexample: 

• The President of Western Union viewed the telephone as a toyand failed to invest in it.
• Newspapers around the country were shuttered because they weretoo slow to adopt the web.
• In 2000, Blockbuster rejected an offer to buy Netflix for $50million (Netflix revenue in 2013 was nearly $4.5B).
• Taxi companies were blindsided by the impact of companies likeUber.

Hindsight is always 20/20, but you get the point that resting onyour laurels can quickly turn on you. 

The best way to avoid being a late mover on the data front is toact now. Carve out a small team that is protected from day-to-dayoperations, such as from urgent fires that distract from achievinglong term strategies. Turn that team's focus to finding predictivemodels with optimal use cases; important, practical projects thathave a high probability of success, and that align with a largercompany strategy. But keep the team on its toes, as your analystsuncover value in your data, they'll likely find new directions totake. You've got to be nimble so you can continue to move quickly.Speed doesn't mean haste, don't be impatient, stay focused on theuse-case processes where you'll generate the best results.

We've seen this measured approach enable some early data wins,while setting the foundation for future, broad applications of datascience. For example, one claim executive used our predictive MOIproduct, which recommends routing totaled vehicles for salvage andrepairable vehicles to the appropriate appraisal source.Because CCCbuilt a model and developed a use case that deployed the science inreal time in use with the carrier's live interaction with theconsumer, the carrier was able to realize more than 3 percent ofLAE savings.  The culture and executive support thatenabled this success did not happen overnight.  Yetresults grew exponentially. Success builds on success; adoptionbarriers get smaller; and support is reinforced. 

If you want to dig deeper into your own data, or to learn moreabout CCC's solutions, please contact me at [email protected].

[1] http://smalldatagroup.com/

[2]http://www.forbes.com/sites/gilpress/2013/04/23/the-big-data-landscape-revisited/

[3] McKinsey Global Institute. "Big Data: the Next Frontier forInnovation, Competition and Productivity." 2011