3 immutable qualities of successful data science

Data scientists in the marketing world have something in common with the nerd characters on police-procedural TV shows, according to speakers at a DigitasLBi-hosted event in Singapore last week.

Nearly every police-procedural show has at least one character stereotyped as the ‘nerd’, such as Spencer Reid on Criminal Minds, Abby on NCIS and Felicity on The Arrow. A common trope during scientific conversations in these shows is for a non-science lead character to interrupt the ‘nerd’ with "In English, please"—prompting the ‘nerd’ to simplify their spiel so it's more easily understood by the team.

Apparently, life for data scientists in APAC is much the same. The DigitasLBi event, titled "Learning to love data", included speakers Sunita Kaur, APAC managing director at Spotify; Ruben Lawrence, marketing director for Asia at NBCUniversal; Natalie Gee, creative director for NBCUniversal; Joel Fisher, APAC director of display advertising sales at TripAdvisor; Grace Tang, data scientist at Uber; and, from DigitasLBi, Chris Clarke (chief creative officer), Dan Hughes (head of data), and Annette Male (APAC CEO).

Like the nerds on TV, data scientists often take on the role of not only understanding data but also translating it for simplified and actionable consumption by their colleagues, clients, and partners. Conversations with the data scientists at the event revealed that succeeding in this niche field requires three immutable qualities.

1. Curiosity

Can it be taught? The verdict is still out on that one. A successful data scientist is one that can enter Batman's detective mode and conduct exhaustive data prep. Even the most advanced machine learning algorithm (as of writing this on 8th June 2017) cannot upset the curiosity of a human data scientist. The right combination of skills with curiosity is climacteric. The good news is, "people are getting more tolerant of creepiness", Kaur said.

2. Pragmatism

In a world where creative continues to be dominated by gut feel and intuition, the data scientist that points out the datasets that reveal the truth behind customer behavior continues to be vilified. It's like being the CFO of time and energy, justifying tasks and upsetting the norm. Data scientists that can prioritize the ROI of a project before themselves are in the best position to succeed in the long term. Every successful data scientist seeks to think about and map the possible as well as likely outcomes. They then map out the effort, scope, resources, and time required to achieve a set of outcomes. Considering the big picture opportunity cost is critical, even more, so projects that may not have a feel good factor but in the end result in greater value generation for the organization, as is often the case for projects that reduce customer attrition or churn rates.  "We need to be gaming the dopamine reward path through the devices in our pocket," said Clarke. "Disruption is not always going to be for the good of the end user, but for the good of the end shareholder."

3. Communication

Truly the most requisite skills essential for success in every profession, the data scientist of today must be able to speak the language of the layman. Breaking down explanations to their simple form and all. In the field, a data-scientist struggles to present a data set, project update, and even project case study that is true to the science but is also easy to understand and comprehend for the listeners from all academic backgrounds. By the core of their training, a data scientist is primed to rely on charts and graphs to deliver the point. When working for a brand marketer or agency, the visual aids are expected to become more localized for that industry. These will continue to be human tasks, said Tang. "AI is nowhere close to coming up with creative or media campaigns," she said. "It's very good at specific, quantitative problems, but terrible at qualitative tasks. Humans still needed to interpret AI. AI is best used in conjunction with human intelligence."

It's not an easy job but someone has to do it. As the world continues to produce more data every day than it has in the past century, we will need data scientists to help us make sense of it.