10 Traits of Top Business Analysts
When you recruit business analysts, what traits should you look for?
- By Ted Cuzzillo
- June 29, 2011
We used to laugh back in the '90s at job postings seeking Web producers with "10 years of experience." Today we're confused again, scurrying to find business analysts to interpret our business data -- a challenging selection when skill with analysis tools is the bare minimum.
What would I do if I were recruiting? Disregarding skill level and tools, how do we recognize a good business analyst -- or someone who could grow into the role -- when we meet one? I asked several experts.
They're "five-legged sheep," says Frank Buytendijk, trying to be helpful with the Dutch expression for a very rare animal. Buytendijk is the Netherlands-based author of several thoughtful books, most recently Dealing with Dilemmas (2010, Wiley).
They're "hunter gatherers" who "stalk the wild data resources of the business, seeking out new and unusual facts, and building from them deep insights into the meaning of business life and events," wrote business intelligence consultant Barry Devlin in a 2009 white paper for Lyzasoft, Collaborative Analytics: Sharing and Harvesting Analytic Insights across the Business.
Analysts come from all disciplines. I know a former accountant, a social psychologist, and a former help desk worker who all thrive now as analysts.
A few traits experts offered came as no surprise. Good analysts:
Have a scientific sensibility. Two attributes are key to a scientific attitude, according to physicist Alan Sokal: a willingness to accept what you find and a willingness to discover that you're wrong. For the discussion this trait requires, see Jonathan G. Koomey's book, Turning Numbers into Knowledge (2008, Analytics Press). It should be on the reading list of every new analyst.
Continually challenge their own mental ruts. Everyone is prone to lazy thinking. Our everyday interpretations of data tend to fall into patterns that hide critical ambiguities and shades of meaning. Good thinking, and valuable analysis, tend to be messier than that. That tendency is a major concern among intelligence analysts, according to Psychology of Intelligence Analysis by Richards J. Heuer, Jr. (1999, Center for the Study of Intelligence, CIA).
A related problem is thinking in simplistic stories, as if to tie a neat bow long before data's been sufficiently analyzed. Bad analysts think in stories too soon, and good analysts often resist stories to the end. "This is what makes truly amazing data scientists so rare," writes Venkatesh Rao, a blogger at ribbonfarm.com and an analyst himself. Data's fine shades of meaning often fit badly within a cute, punchy PowerPoint show.
Are skilled at abduction. The ability or willingness "make a leap of faith to explain something," says Buytendijk, is important. Abduction is otherwise known as "inference to the best explanation," according to the Stanford Encyclopedia of Philosophy. Others call it guessing. Whatever it is, it sounds logical to me that a good analyst should be willing to make a good guess at the cause of what the data says.
Have a deep knowledge of the business ecosystem and the world around it. A good analyst arrives knowing the world and business trends and acquires profound knowledge of the processes of the business.
Lack of knowledge shows up in little things. One 22-year-old analyst in India was excited to tell his boss, a banker in New Jersey, about the correlation he found: one firm's revenue correlated strongly with the number of employees named Bob on the payroll.
Are passionate about data. They love to wallow in it, get it all over themselves, see what they can do with it, and see it change state and morph and eventually crystallize into facets of meaning.
Are practical and resourceful. Barry Devlin wrote to me in an e-mail message that good analysts use "whatever tool or data can do the job."
Their resourcefulness may surprise you. One occasional analyst I know used an ironing board. Terry Baird analyzes data now and then as a member of a thriving cooperative bakery, one of a group of cooperatives near San Francisco that includes the Cheese Board Collective in Berkeley, Calif's "Gourmet Ghetto." About 10 years ago, he helped evaluate a new store location. The new customer base couldn't overlap the Cheese Board's. However, he had no data.
Over three busy mornings, he set up his ironing board by the front door with a map of the area. Each customer got two stick-on dots, one for home and one for work. By the end of the week, the dots marked market boundaries. He saw that several of the possible locations would work.
Are tenacious. Learning a new tool is tough. "You can't become an expert overnight," writes Andy Cotgreave, senior data analyst at Oxford University. "Without complete knowledge [of a tool], common problems will give you serious headaches, and you might create a massively inefficient workaround to solve them."
Other traits are less obvious. Analysts:
Can be troublemakers. Smart people with spirited intellects and strong egos often have disagreements. Analysts stalking multiple versions of the truth can be rebels or disrupters of the status quo. I keep thinking of Apple's old TV ad that began, "Here's to the crazy ones...". Colleagues who insist on a "single version of the truth" or strict data governance are especially prone targets.
Have impact. Good analysts know how to exert influence, says Ken Rudin, vice president of analytics and platform technologies at Zynga. "Analytics is not about insights," he says, "Analytics is about impact. If no one changes behavior, there's no impact."
Some of the impact comes in storytelling to connect with audiences and decision makers. Andy Cotgreave cites the "amazing bubbles" Hans Rosling used in his famous visualization. Amazing as they were, it was Rosling himself who made it so compelling. "Even the best visualizations," he says, "almost always need commentary."
Are impartial and free of "hidden agendas." I admit, this is a reach, but it's worth observing. Eleven distinct archetypes may ebb and flow in an analyst's mind, writes Venkatesh Rao. Only three of the archetypes seem up to the job, the most attractive being the "data philosopher." The other archetypes help to distort data out of personal bias, distract the mind with trivia, manipulate data just to prove others wrong, or simply to lie, among other things.
Next step: figuring out how to detect these qualities in interviews -- and to refine them as the craft matures. To start with, though, it's easy enough to spot a five-legged sheep.
||Ted Cuzzillo is a journalist and industry analyst focused on analysts' tools and needs as well as the environments in which they work. You can contact him directly at
firstname.lastname@example.org. If you analyze data, he'd appreciate your participation in his survey; you'll receive a free preview of his report when it's complete. |