On April 21st, 2021, IIA held its third completely virtual Analytics Symposium, which was a celebration of diversity of opinion and perspective in the analytics industry.
The day began and ended on opposite ends of the spectrum, with a panel of C-Suite analytics leaders kicking off the event with a bird’s eye view of their organizations’ efforts to get better at analytics, while psychologist/science journalist David McRaney closed the day by offering his thoughts on the relationship between psychological theories of confirmation bias/faulty heuristics and data-driven decision making. We also learned from VC Investor Margaret Wu of Georgian about how a VC firm focused on AI measures a company’s AI capabilities -- a perspective that can help our clients improve their internal AI performance, as well as select better AI vendors. On developing analytic talent -- a perennial pain point in the analytics industry -- Fireside Analytics CEO Shingai Manjengwa spoke about working on a tiered approach to analytics talent, instead of expecting to hire a do-it-all data scientist.
Here are some of the key takeaways from each of our Symposium presentations:
Opening Remarks - Alistair Croll
Author, Lean Analytics & Chair, FWD50, Startupfest, Scaletech
What is your job, really? Many data and analytics professionals get caught up thinking that their job is to produce insights, which may be true, but Alistair notes that the challenge really lies in whether or not anyone will care. Getting buy-in on analytics is often the hardest part of the job, and Alistair argues that hard problems are hard because normal solutions don’t work -- you need to find abnormal approaches. Alistair provides tidbits of advice from his new book, Just Evil Enough: How subversive thinking can deliver the results you need. Mixing behavioural psychology, game theory, and lessons from the world’s most iconic brands, Alistair discusses how to create the right attention to help your projects succeed.
Founder and Chief Executive Officer of Fireside Analytics
Data science is a team sport – industry must rethink the Chuck Norris data scientist role and empower broader teams to solve problems and create value with data. Shingai’s talk goes through how to unlock value in non-technical business functions and move from dashboards and reports to cost reduction and revenue-generating opportunities. Her theory includes 5 key factors:
- A data driven culture starts with leadership. If you don’t work to actively cultivate and form a culture, one will form by itself -- and it’s unlikely to be data driven.
- Secure that the team knows the objective functions. Ensure that everyone involved in a project actually understands and keeps in mind the end goal (objective function), which is typically to maximize company profit. Do your team members understand how their work feeds into that end goal?
- Focus on solving problems and creating value with data.
- Highlight and reward problem solvers. And on the flip side, it’s important to not punish people when they fail, but recognize that mistakes are learning experiences that make for smarter employees and better models.
- Invest in targeting professional development and technical education, from literacy to advanced analytics. In order to effectively work cross-functionally, there needs to be a level of common understanding/language surrounding data.
Lead Investor at Georgian
The hype around AI and ML have corporations rushing to launch AI initiatives despite inadequate infrastructure, resource limitations and little connection to business value. Margaret discusses AI and ML from a VC perspective including insights on critical leadership skills, talent evaluation, identifying market opportunities, building competitive advantage and the continuing evolution of the cloud ecosystem.
Two key factors are data richness and AI capabilities. Some of the factors that Georgian considers when examining a company’s data richness include:
- Deep and broad data set
- Data that cannot be easily acquired
- No concentration of data sources owned by a third party, who could retract access at any time
- Exhaust data, which is a data set extended with new labels that further context
Being able to use customer data in an aggregated, anonymous, form results in a greater ability to develop unique and valuable product offerings.
When it comes to AI ability, Georgian looks at (and if you’re selecting an AI vendor, you can look for these things as well)
- Are there dedicated data scientists?
- Are they being led by a product manager with a dedicated AI roadmap?
- Is there a working R&D process so that new models can be smoothly transitioned into production?
- Is data science working in a silo to the rest of the tech team?
Science Journalist, You Are Not So Smart
Enabling data-driven decision making and building a data-driven culture are critical for analytics success, yet these remain elusive and difficult-to-realize objectives. You Are Not So Smart author and blogger David McRaney explains why overcoming the pitfalls of cognitive biases, logical fallacies, and faulty heuristics are critically important for analytics leaders. One simple but powerful takeaway from David’s is that we are smarter as groups. Psychological studies have shown that when given several options, you tend to pick the one that is easiest to justify, not the one that is "best." This can have negative consequences when it’s done in isolation, but working in groups can minimize that internal feedback loop that blocks innovation.
C-Suite Data and Analytics Panel
The complexities and challenges in developing and deploying advanced analytics are increasing on a daily basis while the pressure to deliver business value has never been greater. This panel explored perspectives from C-suite data and analytics executives as they lead their organizations to deliver on the promise of advanced analytics and AI.
Jeff McMillan, Chief Data & Analytics Officer, Morgan Stanley
Gina Papush, CDAO, Evernorth
Cameron Davies, Chief Data Officer, Yum Brands
Albert Marinez, Chief Analytics Officer, Intermountain Healthcare
Bill Franks, IIA CAO (panel moderator)
Don’t be afraid to hire out of your industry
Several analytics leaders confirmed what we know to be true at IIA -- facilitating cross-industry transfer of knowledge can deliver immense value in terms of innovation and creative problem solving. Cameron Davies, Chief Data Officer for Yum! Brands, and Gina Papush, Global Chief Data and Analytics Officer at EverNorth, both have had careers that spanned multiple industries. Gina was in finance and insurance before switching to healthcare, and Cameron Davies worked in entertainment before the restaurant industry, and both agreed that the switch to a new industry was energizing, and their cross-industry experience helped them ask new questions and see new solutions to old problems. Davies noted that he often looks for cross-industry experience in the people he hires, saying that “if you never hire anybody from outside your industry, you're never going to be smarter than anybody else in your industry.”
COVID is accelerating demand for analytics, maybe more than we realize
Albert Marinez, Chief Analytics Officer at Intermountain Healthcare, spoke about the value that their existing data and analytics structures had when COVID hit. Their models and analytical infrastructure allowed them to rapidly look at metrics like operational capacity, predictions for COVID infections, staffing needs, etc. Marinez noted that while their response to COVID was good, it also highlighted the need to continue to modernize old, legacy systems. Cameron Davies, Chief Data Officer for Yum! Brands, also discussed how COVID pushed them to make big investments in their digital platforms, which has now become a major part of their business ecosystem.
Use empathy to help get buy-in on analytics
Oftentimes, data and analytics professionals run into issues not because their numbers don’t add up, but because they are met with resistance from stakeholders or executives. To combat this, Jeffrey McMillan, Chief Data and Analytics Officer at Morgan Stanley, spoke about the need for those in data and analytics to embrace empathy. Is the analytic output you produce really connected to the problem that a person is having? Does your work shine an uncomfortable light on a business unit or individual? McMillan advises that remembering to constantly make an effort to think empathetically about your work can help drive analytics further in your organization.
Our Spring 2021 Virtual Symposium was a day of diverse perspectives and inspiring approaches, focused on growing analytics capability and analytics value. For more information about upcoming IIA Symposiums, go to iiasymposium.com.