The Everyday IDO has been saved
The Everyday IDO
A front-line view of insight-driven organizations at work
Insight-driven organizations (IDOs) are able to put data-driven insights to work everywhere in their businesses, everyday. Sometimes this takes the shape of a big initiative. But the real mark of an IDO is its ability to put insights to work in small ways, hundreds or thousands of times a day. In our Everyday IDO series, we have gathered small but potent examples of leaders who are doing just that.
- The language of analytics
- Positive simplification
- Institutionalizing storytelling
- Hold the details
- A unified approach to maximizing data value
The language of analytics
Kimberly Holmes, XL Catlin
Jim Wilson, a lead data engineer at the insurance firm, XL Catlin, was chatting with his boss, Kimberly Holmes, about the people issues the company’s “Strategic Analytics” group faced every day. This is how Holmes described the role: “The business people know what data they need and can define requirements, but typically don’t have the skill set to design a data architecture that gives them the data they need. Technology people typically don’t understand the business requirements, but they can design the data architectures. Jim responded, “It’s like the people in IT speak blue, the people in business speak red, but we need people who speak purple to create an appropriate solution.”1
The name stuck at XL, so Holmes seeks “purple people” to translate the needs of the business for data and analytical systems into the high-level designs for those systems. Other people may actually develop the models and write the code, but the systems couldn’t exist without those who speak purple.
Holmes is passionate about the need for people who can bridge business and analytics. Because people with all the needed skills are like “unicorns,” she has wrestled with the question of which skill to insist on first. “Do you hire a good businessperson and teach them technical skills, or hire the highly quantitative person and teach them how to work with the business? I have settled on the former. You can teach analytics, but you can’t teach likability.”
Lessons from the “Purple People” story:
- Make sure that you have people who can speak in both business and technical/analytical terms.
- “Light quants” can be a valuable addition to analytical teams if they can work closely with the business and manage change.
- It’s useful to have a shorthand description of the complex talent challenges in analytical work.
1The term “purple people” has been attributed to Brian Stucky, who described “purple people” in a chapter he wrote for Business Rule Revolution: Running business the Right Way, published by Happy About, 2006.
Kevin Kelley, Great American Insurance
Kevin Kelley is a Divisional Vice President in the Specialty Property & Casualty (P&C) business unit of Great American Insurance. He’s the leader of an analytics group there that focuses on models for P&C underwriting—whether to accept an insurance risk, and how to price it. Kelley says that his group “uses stories to help analytics make more sense.”
The stories told by Kelley and his group focus on two objectives: making the results of analytics simpler and more positive for those affected by them. P&C underwriting models, for example, might involve as many as ten different risk characteristics for a prospect, which can make the result difficult to describe to nontechnical audiences affected by the decision.
Kelley’s group began to use “reason codes” for the major drivers of the technical price. These captured the most important reasons for the decision, although not the only ones. Reason codes could be generated automatically, which reduced the storytelling burden on the human analysts.
But Great American’s agents are independent, and can take their business to a variety of companies. Underwriters directly communicate with them about the results of an underwriting decision and will try to put a more positive spin on it. If they have to turn down a policy application, they can say, “You’ll be most successful if you bring us this kind of business.”
What Kelley’s group is doing is to try to identify “success patterns” for particular agents, which can lead to a positive and understandable story to discuss with them. While Great American’s models are increasingly automated, no current automation approach can identify all the positive aspects and success patterns in an agent relationship. It’s only the human underwriters that can play that role. Kelley notes, “It’s not just the best math in the model, but how do you get the information out there to those who are affected by it.”
Lessons from the “Positive simplification” story:
- Analytical models can be complex, but explaining them typically requires communication of the most important factor in the model.
- In order to improve and maintain customer and partner relationships, it’s important to look for “success patterns” in the relationship and tell a positive story about them.
- At the moment, only humans can find and tell such positive stories.
Pamela Peele, UPMC Insurance Services
Perhaps the most advanced approach to storytelling is to institutionalize it in the process of creating and disseminating analytics. Sometimes we forget that there are people who make it their profession to tell stories–journalists.
But we know of only one analytics leader, Dr. Pamela B. Peele, who has hired a journalist to leverage the storytelling efforts of quantitative analysts.
Peele noticed several years ago that her analytics group was making great progress at creating new knowledge from data and analytics, but wasn’t having the impact on organizational initiatives and policies that she had hoped.
So a few years ago Peele hired someone with a journalism background, primarily to write clearly and in plain language about what her group’s analysts discovered.
Peele felt that telling good stories was a way to help executives consume that truth in an understandable form. “We have a finite amount of attention in a world of massively increasing information. The challenge for big data is to not only produce new information but to weave information together to make it attention-worthy.”
The analysts do the design and analysis, but it’s the storyteller’s job to elicit the main points and how they will impact the organization. “The storyteller’s job, Peele says, “is to start at the end and only provide technical detail when absolutely necessary.
Lessons from the “Institutionalizing storytelling” story:
- Analytics executives need to focus just as much on the successful consumption of analytics as the generation of them—and stories are key to accomplishing that objective.
- Professional storytellers—also known as journalists—can be helpful additions to an analytics group and can free up quantitative analysts from having to tell their own stories.
- Analytical stories should begin with the implications of the analysis, and only provide technical detail when absolutely necessary.
Hold the details
Analytics Leader, Major US Bank
This experienced analytics leader at a major US bank knows when to involve his internal customers in the details, and when they’re better off being a bit insulated from those details.
The bank is going through a revolution in the technologies it uses to store and process data for analytics. It has committed to a long-term plan for using open source software and commodity storage solutions to help lower costs and achieve higher performance levels. It will also make it easier to comply with new regulatory structures that apply to all US banks.
However, many of the bank’s quantitative and business analysts are highly attached to existing proprietary software tools and methods. The goal of the analytics leader’s current work is to insulate the bank’s analysts from this change in technology. He and his colleagues are creating a layer of abstraction for analysts so that they won’t have to change how they access and analyze big data. If users are comfortable with a certain statistical package or database, they can continue to use the same language they are used to. What’s important to the analytics leader and his colleagues is maintaining a set of architectural principles that underlie the new architecture.
The user doesn’t need to know what is happening behind the scenes. The details of how those processes are executed are important to the bank’s technology organization, but not to the average analyst.
Lessons from the “Hold the details” story:
- Architectural transformations that are important to the success of an organization may not be perceived as important to the individual analyst.
- It may be useful to impose a layer of abstraction between what the user sees in an analytical process, and what happens behind the scenes.
- New processes for analytics can be useful both for offense—marketing and customer relationships—and defense, or the actions needed to comply with regulations.
A unified approach to maximizing data value
Peter Husar, TD Bank Group
Many Analytics professionals would agree that the degree to which we understand our customers facilitates our ability to connect with them and meet their current and future needs. Peter Husar, Vice President of Enterprise Customer Data & Analytics, and his team at TD Bank Group, are building upon that outlook to bring about enterprise-wide change within the bank.
Like similar organizations, TD is engaged in a once-in-a-generation change to transform analytics tools, processes and capabilities for teams and functions across the organization. These enterprise-wide initiatives with far reaching impacts can only achieve success through obtaining senior business leader sponsorship, and that’s exactly what Husar and his colleagues have done. Senior sponsorship enables Husar and team to consult with each business segment and corporate function across the organization to understand their unique priorities, relay any challenges existing in the current state, and obtain buy-in for work required to achieve the enterprise future state vision.
The team illustrates how leveraging data strategically will help every part of the bank: from building specialized analytics capabilities to sharing commonly used data. Doing so will build more robust customer profiles, offer deeper insights, enable cross-business referrals, and reduce inconsistencies. "The cases for each business segment and corporate function across the bank are informed by a current state assessment of reporting and analytics capabilities," says Husar. "We have established an Analytics Advisory Council, comprised of leaders of various reporting and analytics groups across the organization to provide input on those capabilities that will resonate within each unit."
Keeping all stakeholders aligned to a common vision is essential, and Husar and team require that senior leaders commit to a call to action, commitment of people and resources, and overall support on behalf of their business segment or corporate function. Taking a unified approach to leverage data strategically across the organization will strengthen TD's capabilities to address customers' needs, exceed expectations, and deliver legendary experiences today and beyond.