New insights from MIT Sloan Management Review explore a gender gap in experiencing toxic culture, how AI can improve performance measurement, and when it makes sense to consider subtraction as a lever for growth. And an examination of Northwestern Mutual’s AI initiatives shows how the financial services firm is staying fresh in the face of competition from fintechs.
Toxic culture impacts women more than men — no matter where they work
Women are 41% more likely than men to experience toxic workplace culture, as reported in Glassdoor reviews, according to a five-year analysis conducted by MIT Sloan senior lecturer and his CultureX co-founder, Charles Sull. This gap only worsens with seniority: Women in C-level roles were 53% more likely to mention toxicity.
In the Glassdoor reviews, women were 2.8 times more likely than men to speak negatively about sexual bias, discrimination, or gender-related exclusion in the workplace. But they were also more likely to mention many other elements of toxicity, from bad leadership — whether disrespectful, abusive, manipulative, or unethical — to a culture that thrives on fear and favoritism.
And the occupations with the largest toxic-culture gaps tend to have a higher percentage of male employees, such as electrical engineering or the military; those with more female employees, namely child care, social work, and mental health, have the smallest gap. The gap also tends to be higher in industries with a high percentage of hourly workers, such as retail, transportation, and food service.
Sull describes toxic culture as a “cancer.” One key step to removing it is finding the microcultures where toxicity is often concentrated, whether it is a specific business unit, regional office, or team, and acting “quickly and effectively.” Companies that succeed in fixing toxic culture identify these problems by collecting data frequently, mining responses for hidden insights, and aggregating feedback tied to an individual manager or business unit. And they communicate changes that have been made after assessing the findings.
Consider subtraction as a catalyst for growth
Business growth often defaults to the addition of new products or more price promotions. In industries such as retail, though, these decisions can have negative impacts downstream. Productivity slows as employees rearrange store shelves and field questions from confused customers. Lower productivity leads to lower wages, which increases turnover, contributes to burnout and toxicity, adds to staffing challenges, and worsens the customer experience.
MIT Sloan professor of the practice advocates a growth strategy of subtractive change. This involves removing distractions or other tedious and unproductive tasks that hinder customer service. Fewer sales promotions leads to less time spent updating prices and to more predictable demand; fewer products means shelves and stockrooms are easier to fill, which cuts supply chain costs. And subtracting hours of operation can have a positive effect. When stores open later and close sooner, employees can work more stable schedules and devote more off-hours time to training, which reduces employee turnover and boosts morale.
At large companies, these decisions are made upstream, in a corporate office many miles from the retail stores. Leaders may avoid subtraction as a strategy because it could conflict with short-term sales goals or the incentives in their compensation packages. Silos between business units may cause leaders to abandon changes that seem small but could have a larger impact across the organization. The only way for leaders to know whether a subtractive change will work, Ton writes, is to go to front-line employees and ask them.
Use AI to fine-tune performance measurement
Just as the purpose of performance measurement has evolved beyond increasing worker productivity, the data driving performance measurement has also evolved over time. In fact, writes Michael Schrage, a research fellow at the MIT Initiative on the Digital Economy, leading organizations increasingly believe that key performance indicators aren’t just static measurements of performance.
By using artificial intelligence to refine existing KPIs and generate new ones, leaders can continually adapt and optimize KPIs to improve performance, profitability, and growth. Schrage and his co-authors outline three implications of applying AI to KPIs:
- Shifting from tracking progress to driving action. Companies traditionally track churn as a lagging indicator, but it’s possible to predict churn and explore options for preventing it. Analyzing customer data lets firms identify at-risk customers and determine how much effort it will take to retain them — or let them go if it’s not worth the effort.
- Measuring how KPIs perform. AI makes it possible for companies to assess KPIs on a regular basis, ensuring that they provide leadership with the right metrics to drive decision-making. AI algorithms can also look at relationships among multiple KPIs — no small feat when disparate business units generate frequently siloed and sometimes competing KPIs.
- Aligning KPIs. When KPIs conflict, disagreements can arise — for example, if finance wants to cut costs, but marketing believes that increased spending will attract customers. AI can see where KPIs overlap, making it possible to resolve underlying conflicts and generate shared KPIs that better align with the organization’s overall goals.
Explore Northwestern Mutual’s ongoing experimentation with AI
Stability is important for financial services firms. Amid economic uncertainty and increased competition from digital-first fintech firms, Northwestern Mutual has spent the past five years focusing on how AI and data science can improve the customer experience. Momentum picked up considerably when the firm hired chief data officer Don Vu in 2020.
Thomas H. Davenport, a fellow with the MIT Initiative on the Digital Economy, and Randy Bean, an innovation fellow at data consultancy Wavestone, describe three initiatives that illustrate how Northwestern Mutual is aligning its AI work with its business goals:
- Underwriting. By analyzing data in existing digital medical records, the company can eliminate the need to conduct home visits and collect blood and urine samples from life insurance applicants. In addition to helping Northwestern Mutual keep underwriting operations moving during the pandemic, it also cut the time it takes to issue new policies from four weeks to three days.
- Advising. Individuals who hold life insurance policies are customers for decades — longer than the tenure for most financial services advisers. Northwestern Mutual is developing predictive models to help identify clients who are likely to want new or upgraded products, which should help less-experienced advisers serve longtime customers.
- Training. Northwestern Mutual’s foray into generative AI has focused on making its vast internal training documents easier to search. Information is currently spread among several intranet portals, and better synthesizing these resources will benefit its advisers.