With contributors from: Carl Lambert, Vice President, National P&C Business Intelligence at The Co-operators and Hashmat Rohian, AVP, R&D, Innovation at Aviva

“In the 90s, actuaries wanted pre-processed data but now they’re saying that they want raw data to work it themselves. That’s a big difference,” said Lambert.
Not only is this change impacting data governance, but it’s also changing the complete architecture of analytics platforms. There is now a need for an entirely separate protocol from dashboards or reporting because actuaries and data scientists need more control.
“It needs to be faster and give the user much more control so they can use their own tools, run their own analysis and integrate any data they want,” said Lambert. “Speed is where you gain insight.”
And in the Canadian insurance market, this insight is a key differentiator. If a company is taking a quarter or a year to process data and learn from it, then they’re behind the market. Speed is of the essence, and that will likely mean improvements in infrastructure or a private-cloud architecture to keep up. 
Not having a system designed to meet the needs of data scientists is among the quickest ways to fall behind.
Actuaries precede data scientists by many years, and this legacy may have caused insurance companies to fall behind in new areas of analytics and their application.
The most-cited sector by our experts and the industry at large is in the area of social media. Social media has proven to be a successful arena for analyt¬ics to determine general company perceptions, happiness of customers, customer service needs, risk and fraud. 
Social data mining presents the insurance industry with the opportunity to better align its pricing structure with demands and customer sentiment. Marketing materials also have the ability to provide a more immediate impact. The evolution of Facebook and Twitter advertising units now allows marketers to track users across multiple devices and websites, allowing ROI measurements to become more certain and specific.
In all of these client-facing services and channels, data analytics will serve as the backbone for identifying and validating trends. Analytics teams have the opportunity to apply their expertise to new areas and increase the company’s bottom line, making them more valuable as a distinct business unit.
A concern in implementing public-facing business units and applications is the patchwork structure that governs insurance. 
The regulatory landscape varies greatly across insurance segments as well as local jurisdictions, often inhibiting innovation. With pricing and policy information often required to be public information, companies may find it hard to innovate in the areas most visible to consumers.
“Regulators have also proven to be less-than-accepting of innovation at times, so you get push back there too,” said Lambert. “It’s also tough to get a full 360-degree view of the client because you can’t align all available customer information in many cases because it’s not allowed in instances such as auto insurance.”
Analytics also shines a spotlight on some regulator-approved metrics that aren’t as useful as new data options. Credit score, for instance, is seen by legislation as a certified measure of risk, when it is essentially a measure of behavior that’s just proxy for the risk.
Some analytical innovation is stifled by this regulation and a legal process that’s slow to adapt itself to the nature of the Internet, cloud computing and other Big Data advances. However, these technologies also have the potential to help insurers to better respond to changes as they finally arrive.
The government is still responding to the advent of Big Data and it’s likely that new legislation around data access and consumer privacy will continue.
This shouldn’t be viewed as the government, or industry, playing catchup, but instead as part of a continual flow of information and regulatory changes. In 2011, IBM stated that 90% of the world’s data was created between 2009 and 2010, so it likely that insurers will remain in a state of flux.
However, the same analytics programs governed by current regulation may also prove to be the best way to adhere to regulation. Analytics platforms are now robust enough that they can provide insight around both data trends as well as individual units. 
Insurers can take advantage of granular processing capabilities by apply¬ing it to data governance around information types and characteristics. Currently, information on home and commercial lines can typically be shared when an insurer owns both policies, but there are limitations to what data can cross over for auto policies.
As legislation continues to change – such as the Insurance Companies Act (S.C. 1991, c. 47) being updated this June – rules can be applied within a system to categorize data and apply restrictions or controls to limit or allow sharing. This immediate change could be used to ensure that all actuaries and data scientists are only accessing the allowed information, which is especially helpful as analytics programs expand.
An added benefit to the data governance expansion and classification would also be a reduction in system-wide replacements or updates that removed software or connections. The shift to a permission-based model also takes the burden off of IT by making changes an adjustment of access rules as opposed to creating limitations connection-by-connection.
For companies and business unit heads looking to start their own analytics program or expand a current offering, it’s best to start by helping others. Don’t focus on imposing an analytics mindset, but instead position analyt¬ics as a support structure for other operations.
“Be humble and try to see the different departments of an organization and ask them what their challenges are. Walking in with your own idea about how to fix things may not be well received,” said Lambert. 
Big Data analytics is popular among actuaries because it’s the latest name for a well-known practice. However, in business units that aren’t familiar with the term, it could be threatening. Analytics has the potential to change practices and automate some work, so it can be seen as a competitor to the work of business units that analysts are trying to help.
“When you talk with a provider or other partner, they’re essentially telling you: ‘help me buy, don’t just try to sell something to me.’ Take this same approach to analytics. If other company units have more success because you work with them behind the scenes, analytics will be seen more favor¬ably by your company at large,” said Lambert.
A major part of that structure is hiring the right team. Each data unit must have a mix of scientists and those who understand the business and can speak its language. The blend is important because the funding conversa¬tion always comes back to the return on investment.
“It’s a safe assumption that customers are always looking for better value propositions. They don’t stay the same from day to day as they navigate through their lives and careers. They expect continuous improvement on the proposition and organizations have realized that they need to deliver incremental value through disruptive innovation and predictive analytics to avoid being disrupted themselves,” said Rohian. He suggests treating the C-suite like that customer and proving direct value and return on invest¬ment. “Once you have sold them on the ‘why’ it is good for them, then it is just a matter of going in and executing in an agile manner.”