Many companies falsely assume that ramping up analytics is a matter of hiring enough data scientists. Maybe that’s why job postings for this role have increased by over 250 percent since 2013.
McKinsey cites the example of one large financial services firm, where the CEO was an enthusiastic supporter of advanced analytics. He was especially proud that his firm had hired 1,000 data scientists, each at an average loaded cost of $250,000 a year. Later, after it became apparent that the new hires were not delivering as expected, it was discovered that they were not, by strict definition, data scientists at all.
[ How are you screening for communication skills? Read also: The most important skill to look for in data scientists. ]
Data science roles that you can’t skip
A common problem in the industry is that there is no clear understanding of the data scientist role, nor of the other critical roles on a data science team. This is a key reason that many analytics projects fail.
Roles such as data engineer, machine learning engineer, and visualization designer are becoming more prevalent. But unless executives prioritize hiring for three other important skills, they can expect lost data science opportunities, failed initiatives, or projects that don’t get adopted at scale. Let’s explore those specific skills and roles:
1. Data science translator
You must start data science initiatives by picking the right business problems. Translators help you identify the most impactful projects. They sculpt the business challenges into a shape that can be solved by data. Translators possess a strong understanding of the domain, have data fluency, and are effective communicators.
Translators act as a bridge between business users, data engineers, data scientists, and visual experts. They infuse the soul into the solution and act as a glue that binds all the roles together in a team. Their role continues well after project delivery to help users adopt the solution and scale it organization-wide.
“Hire as many data scientists as you can find — you’ll still be lost without translators to connect analytics with real business value.” - McKinsey
Signs you are missing this role in your organization:
If your analytics projects aren’t delivering value, you may be picking the wrong initiatives. If your business and data teams work in silos, information may be lost in translation. If your projects die a slow death, you might be falling short on internal advocacy. All of these are signs that you’re missing these internal champions.
Fix this problem by hiring or training data science translators. People with a business analysis background fit this role best. McKinsey projects the demand for this role will reach two to four million by 2026 in the U.S.
Given the skill gaps in the industry, you will likely have better success training internal candidates with the right business acumen.
[ Read also: 3 reasons data hoarding may not pay off. ]
2. Behavioral psychologist
Machine learning models are good at identifying patterns from data. But you still need humans to interpret the many patterns that big data footprints often lead to, and to pick those few hidden gems that deliver business value. Today, most data science applications aim to make sense of human behavior. You need experts who understand why people behave the way they do.
[ Need to chat with non-technical colleagues on this topic? Read also: How to explain machine learning in plain English. ]
Behavioral psychologists can help understand purchase decisions or why customers churn. They can validate actions to promote user engagement or influence lifestyle changes. Experts in humanities and anthropology have a rising role in the data science team as AI enters more areas of our life.
Signs you are missing this role in your organization:
Do your predictions and insights on human behavior miss the mark? Are your teams struggling to interpret the human aspect of their decisions? These are signs that your products aren’t market-ready. If your teams aren’t having these conversations, that’s a bigger red flag.
Carve out roles for people with psychology and social science backgrounds in your data science teams. Have them participate in discussions to define your AI solutions and product features. Let them validate insights about human actions and determine what the decisions mean for the users.
3. Data storyteller
Some leaders naively assume that putting analysts and engineers into a data visualization bootcamp transforms them into storytellers. But data storytellers are not the same as visualization specialists. Data stories are not only visual; they also provide context to the user on what has happened so far. They then add a narrative to summarize the insight and drive business action. Gartner says that data stories need to have all three elements: data visualization, narratives, and context.
As psychologist Daniel Kahneman famously said, “No one ever made a decision because of a number. They need a story.”
You need data storytellers on your team to infuse life into your data insights.
Signs you are missing this role in your organization:
Gartner reports that 50 percent of data science projects fail due to bad storytelling. If users struggle to understand the insights generated by data scientists, you are facing a data consumption problem. If decisions are disconnected from the dashboards, storytelling might be the missing link.
Help your data visualization experts understand the business better. Upskill your data translators to communicate information more effectively. Train both on the techniques of storytelling with data. Validate the actionability of your data science initiatives and measure the return on investment.
[ Read our related article: The most important skill to look for in data scientists ]
The complete data science team
Data science is a team sport. The team needs to be proficient in technical, business, visual, and soft skills. The three missing roles covered here call for expertise in overlapping areas and are often missed.
The role of a translator is critical to help define the business problem and shape an approach to solve it. The role of behavioral psychologists is crucial in interpreting human patterns uncovered by machine-learning algorithms. Finally, data storytellers weave these insights into an interesting narrative that can propagate and promote action across the organization. Without any one of them, your data science efforts could fail.
[ How can automation free up more staff time for innovation? Get the free e-book: Managing IT with Automation. ]
