Coming Soon to Analytics Teams: Analytics Translators
An overlooked skill is needed now more than ever to remove roadblocks to understanding analytics results.
- By Wendy D. Lynch, Ph.D.
- September 29, 2022
As anyone in data science knows, the past two decades have propelled us into a new era. Big data, and its ever-evolving suite of new, open source methods, have produced breakthrough insights in almost every industry. From the way Amazon recommends your next book or how Visa detects fraud to start-ups facilitating early detection of cancer or manufacturers reducing waste, companies are leveraging big data analytics in hopes of reaching new customers and improving profitability.
However, despite some highly visible success stories, there are troubling signs that data science in business is not living up to its promise. Although there are limited studies of the exact rate of failure, we can find some data:
-- Gartner projected that only 20 percent of big data analytics projects will deliver value to business through 2022.
-- Business journals highlight the level of disappointment with analytics, including MIT Sloan Management Review, Harvard Business Review, and Forbes.
-- In a 2021 survey of corporate executives, over 90 percent responded that the biggest barriers to successful implementation of big data were people, communication, and culture, not technology.
The underlying message: failure is common. Whether we believe the failure rate is 80 percent, 50 percent, or even 30 percent, it has huge implications.
On top of that, data analysts will leave their jobs every 20 months, on average, because their advanced skills aren’t being used or they don’t see a promising future. So at a time when data science should be revolutionizing businesses, projects are floundering and employees are looking elsewhere.
The Big Problem We Don’t Talk About
Reviews of what goes wrong emphasize that the problem is not technological -- workers have the tools, equipment, and resources to do analytics work. Nor is it a lack of analytics talent. Instead, the primary culprits are organizational and cultural, a failure at the interface between business and analytics teams.
A communication gap between these two groups should not surprise us. The two have different styles, terminologies, training, and biases, and they usually operate in separate divisions. Their objectives are often misaligned, if not in direct conflict. Communication styles are different -- marketing teams favor simple, definitive statements but data scientists emphasize nuance and limitations, qualifying their observations -- yet neither business nor analyst teams place a high priority on communication training. Marketing teams favor simple, definitive statements. Data scientists emphasize nuance and limitations, qualifying their observations.
Combining these factors, misunderstandings and dissatisfaction seem almost inevitable. According to an online poll, only a third of either team feel positively about their interactions with the other. Analysts describe a relationship where requests are one way, with little context or invitation to provide input. Business professionals describe frustration getting the results they need in a format they understand.
Advances in Analytics Will Only Make It Worse
Although advances in big data elevate analytics, they also compound the level of difficulty to explain results to those with no statistical training. A person who is only vaguely familiar with the basics of Pearson correlation coefficients and linear regression will not follow how importance plots in GBM reflect not only direct associations but also second- and third-order interactions, squared and cubed terms.
Essentially, the better we get at applying machine learning and AI, the wider the understanding gap between analytics teams and business teams becomes. Data scientists -- whose main interest is developing increasingly elegant and accurate models -- may feel irritatingly handcuffed by requests to simplify their results for leaders or clients.
A Dedicated Translator Can Help
Now more than ever before, there is a need for individuals who focus intentionally on bridging this communication gap: analytics translators. This role, first introduced by McKinsey in 2018 (also referred to as analytic translators, or data translators), dedicates effort to converting business terms into analytics language, and analytics results into business outcomes.
Trained in both analytics and communication, analytics translators facilitate better interactions in both directions. They interpret analytics findings in ways that business professionals can apply effectively. Just as important, they know how to ask questions that reveal the underlying business needs, increasing the likelihood that analysts answer the right questions the first time.
- Essential communication skills for analytics translators include:
- Recognizing where interactions go off track
- Listening for signals that there is more to the story
- Asking effective questions to elicit different levels of information
- Selecting and presenting relevant information each team needs
A New Option for Analytics Careers; An Essential Member of Analytics Teams
Previously, data science careers provided two main advancement options: become an expert in increasingly complex methods or become a manager of other analysts.
Analytics translation provides a third option. Analytics translators reduce wasted time and effort spent on rework while making work more meaningful for analysts. Their goal is not only to convert ideas into understandable language for each team, but also to build alliances and mutual appreciation. In this way analytics translators become essential team members in ways that encourage greater collaboration while delivering more relevant and effective analytics results.