The Analysis/Decision Chasm

The Analysis/Decision Chasm

Jason McDonald

Blogpost Image
home_blog_20150928_chasm_bridge_560x320.jpg

There is an overabundance of terms thrown out today that “everyone” “must” embrace when it comes to data. Here are a few of the most popular:

Big Data, Data Visualization, Data Science, Hadoop, Self-Service BI, BI Center of Competency, Data Mining, Predictive Analytics or is it Descriptive Analytics, or wait what about Prescriptive Analytics, etc…

As you can see, there is an ocean of keywords that every vendor, systems integrator and analyst has embedded in the metadata of their website to ensure they pop up in Google. However, there’s one big problem – every term is a technology-centric view of data, as opposed to a business-centric view of the decision-making process.

Unfortunately, this tech-centric view does not help end users who have daily decisions to make. For example, in any organization there are only a handful of employees who:

  • are trained in the science of data analysis;

  • have time during the course of their daily jumble of decisions to drill into data and look for unique patterns, trends, etc.; and

  • have a desire to use simple or complex visualization tools that take them out of the tools they use on a daily basis.

All the keywords, tools, practices, etc. are focused on the handful of users described above, not the majority of decision makers within an insurer. There is a chasm between the analysis and the decision. We need a bridge.

Given the majority of decision makers who lack the training, time and desire to comb through and analyze data, it’s no wonder that there is still only a smattering of insurers in the insurance market that can claim a great deal of success using any one of the aforementioned technology keywords.

To illustrate this analysis/decision chasm, I was in a conversation with a Commercial Lines VP for a tier one insurer and the comment he made that stuck with me was something along these lines:

“We have a group of analysts who take all my data, run their models on it once every six months to a year, and tell me how I should change, or where I did things well. We incorporate the changes and get a response six months to a year later on whether or not we screwed up again or if the changes were meaningful. The feedback loop is too long, it’s too solitary, and frankly, I need tools that help me and my underwriters make the right decisions today, when I’m actually making the decision.”

Where does this leave us? It leaves insurers with awesome, but often unused, visualization/analysis tools and large amounts of “big and small” data patiently waiting for discovery by a group of overburdened data scientists and/or analysts. In larger insurers, these same groups of data scientists remain sealed off behind closed doors, producing results well after the decision is made by the frontline business user. This calls for a different approach. As opposed to focusing on how big the data is or which technology to apply, we must focus on the user and the decision.

Take for example an underwriter by the name of Scott. Scott needs information from a plethora of sources in order to make a decision on a commercial property risk he’s reviewing. He needs crime, weather, flood, agency, past claim history data, etc. However, the information Scott needs isn’t what’s important. It’s the decision that Scott is making. Period. He needs to decide if the property is worth his company underwriting. This makes us look at the problem differently.

As opposed to the insurer worrying about or focusing on the data, the platform needed to support “underwriting,” and the visualization/exploration tools along with the training required to use the new set of horizontal tools and technology, what if the insurer could simply get an app that focuses on Scott’s question: Should I underwrite this commercial property risk?

This app should combine the needed data on a single screen and integrate with Scott’s day-to-day workflow within his policy administration system (i.e. PolicyCenter). The app should give Scott a severity guideline based on his or his company’s underwriting rules; the exploration capabilities to make sure the property isn’t in a flood or crime zone; and guide him when he uses his years of experience to make this complex decision. Sounds great, right?

I believe this is the right focus for vendors to help insurers. This is Guidewire’s focus. We are focusing on the role and the decisions within the role. We already have an application that does what Scott needs: Spotlight. Giving users a guided experience for the decisions they need to make, harnessing all the technology buzzwords that I alluded to before, and hosting the complexities of big data, visualization, etc., but assembling it in a way that makes it accessible, within Scott’s daily workflow and intuitive.

We will be building more apps like Spotlight over time to bridge the analysis/decision chasm. Each will focus on a problem. Each will focus on a role. Because at the end of day, it's not the data that matters, it’s the people and the decisions.

Tags