top of page
Writer's pictureThe Tech Platform

The Recession’s Impact on Analytics and Data Science

There has been a huge demand for data scientists in the past decade. Is that about to change?

The outbreak of the COVID-19 pandemic is having a dramatic negative impact on economies in the U.S. and worldwide, and unemployment rates are soaring. Given the economic disruptions, it seems likely that many countries in the global economy will experience a recession.

Organizations are beginning to grapple with how the economic slowdown will influence investments they are making across the board. One question we wonder about is whether the demand for analytics and data science resources will remain heavy or slow down. You don’t have to look far to find evidence that the focus in this area has been strong: A 2017 report by IBM, for instance, predicted that the number of analytics and data science positions in the U.S. alone would increase by 364,000, to 2,729,000 by 2020. In 2019, LinkedIn ranked “data scientist” the No. 1 most promising job in the U.S. based on job openings, salary, and career advancement opportunities and reported a 56% rise in job openings for data scientists over the previous year. The exponential growth of data — and industry’s desire to use that data for better business outcomes — has been widely cited as a reason for the increasing demand for analytical talent.

Will the current recession slow the growth in demand for analytics and data science? Will changes in organizational goals and focus make job losses in these fields likely? Any lessening of general demand would be good news for aggressive AI adopters and AI-focused vendors, many of which are already hiring laid-off data scientists and engineers. But for the average company, lower demand for data scientists will be a signal that less data science is going on within their organizations, meaning there will be a continued reliance on intuition and other less-powerful guides to decision-making and action.

Predictions: Influences on Analytics Investments in the Next Year


To understand what managers are thinking regarding where to go with analytics and data science in the coming year, we reached out to a number of company leaders in one-on-one conversations and reviewed aggregate demand on job boards. Based on what we learned, we think that as organizations scramble to imagine a new COVID-19 economic reality, four factors will determine their decisions on continued investment in analytics and data science.

Proven return on investment. ROI is one of the first metrics used when companies turn to cost cutting in a recession. This will be true for analyzing investments in their analytics and data science groups. Individuals, groups, or projects that don’t show a clear return on investment will likely be placed on the list of potential cuts for cost savings. This could result in a decrease in the growth of investment or an actual reduction in employment.

ROI is a tough standard for data science, in part because many algorithms never get deployed into production applications. By some estimates, 85% of big-data projects fail. Achieving success in analytics and data science and properly documenting the ROI can be challenging: A report from McKinsey states, “While investments in analytics are booming, many companies aren’t seeing the ROI they expected. They struggle to move from employing analytics in a few successful use cases to scaling it across the enterprise, embedding it in organizational culture and everyday decision-making.” Groups that have been doing mostly descriptive and predictive analytics will be at risk if they have a small percentage of deployments or have concentrated only on building models.

On the other hand, data groups that have proved their value could thrive. John Morris, managing director of operations decision science at Delta Air Lines, told us that “executive leadership will most certainly look to the analytics/data science group for guidance during a recession because analytics has a proven track record for adding data-driven value.”

Pre-COVID-19 leadership support. Existing C-level support for a data-driven culture will be another important factor determining a company’s level of continued investment during a recession. If C-level support has created a data-driven culture, then the use of analytics is likely to be pervasive and a part of the core strategy of the company.

However, a 2019 Deloitte survey of executives of large companies found that 63% do not believe that their companies are data-driven. Similarly, only 31% of the C-level executives of large U.S. companies surveyed by New Vantage Partners classified their companies as data-driven. The lack of strong analytical leadership and a culture that supports it presents a significant risk factor.

According to Charles Thomas, who has led data and analytics groups at USAA, Wells Fargo, and now General Motors, strong leadership of data and analytics groups is critical to their fate in economic downturns. “CEOs often form centralized analytics groups by merging smaller units because they believe more value can be generated if brought together,” he told us. These groups, even if they receive some senior-level support, still have to fight to highlight their incremental benefit to the business in order to gain resources and demonstrate their validity. The best of the leaders of these analytics groups, he noted, have honed their ability to focus on business outcomes: “[They] got to where they are because of their ability to communicate their value (on top of technical prowess) to the overall organization.” An analytics leader who is doing her job properly, he said, “has forged inroads with business partners.”

Analytics maturity. How far along is a company in realizing value from its data? Many so-called maturity models measure analytics development by an organization’s progression from using descriptive statistics to predictive data to more complex prescriptive analytics. Under such models, descriptive and predictive analytics are most at risk because they do not directly impact decision-making unless they are bundled with or embedded in either a rule-based system, a machine learning-based scoring model, or an optimization — that is, unless they are deployed.

In a recession, when there will be an increased emphasis on cost cutting and efficiency, there should be an increased demand for prescriptive analytics. Optimization will be applied to everything from production to logistics to human resource management, and analytically mature organizations should see an increase in demand for data science services. The reality is that analytics groups in analytically mature organizations that have succeeded in creating production deployments for their algorithms are safer in a recession.

The organizational structure of the analytics group. How analytics are delivered and deployed may influence their perceived value to an organization. A centralized structure, most common in larger organizations, consists of an internal consulting group or a center of excellence. An embedded structure allows individual functions to bring on their own analytical support. A hybrid structure has some members in a central group and others embedded in the business units.

A fully embedded group might be more susceptible to cutbacks if there are across-the-board cuts over all business units. Centralized structures could also be quite vulnerable, particularly in less analytically mature organizations and especially if the group has not proved its ROI. In an analytically mature organization, however, the opposite could be true: As GM’s Thomas observed, mature and centralized groups often have stronger leaders who are skilled at delivering and communicating value.

A centralized group with strong leadership, proven ROI, and C-level support could, in fact, experience an increase in demand. Jack Levis, who headed an analytics group at UPS for many years and over several recessions, pointed out that mature and well-connected centralized groups might be asked to employ analytics for targeted cost reductions. “During the [2008] recession, everyone wanted to find cost-cutting measures,” he told us. “My team was buried with work. We were fortunate, as we already built many tools for network modeling and didn’t need to start from scratch. We were running alternatives day and night.”

The Most Likely Determinant of Demand

After discussing this topic with several senior leaders, we concluded that demonstrated ROI, or a lack thereof, is likely to be the biggest determining factor in whether organizations will strengthen or contract their data science and analytics efforts. For those data teams that have demonstrated strong, positive ROI, the demand for analytics might increase in a recession. Based on data from Burning Glass Technologies’ Labor Insight tool on the three-week average for the period that ended April 18, 2020, growth in new U.S. job postings has slowed, with rates of decline varying across industries such as finance and insurance, non-store retail, passenger airlines, and air freight.

However, although new job postings in data science and analytics have declined overall, they currently appear to be declining at a slower rate than that of most other occupations. And within the finance and insurance industry, new job postings in the analytics and data science space have actually increased.

We are on the leading edge of the decline and must be careful in interpreting this movement. We feel strongly that data science capabilities are critical to competitive success, so we hope that businesses don’t let those skills die out or let the best data scientists go to work for a few IT companies. Analytics and data science professionals have experienced extreme demand for their services over the past decade. Whether they continue to be sought by organizations will depend to a substantial degree on the factors we have discussed here.


Source: paper.li

0 comments

Comments


bottom of page