Uncovering Hidden Biases Through Data Helps Symphony Orchestras and Financial Institutions Perform Better

May 29, 2017

If you stepped into a concert hall to hear a symphony orchestra in 1969, then entered the same hall in 2017, you’d notice some obvious differences—and not just in the musicians’ hairstyles. Nowadays you’d see an orchestra that’s quite diverse. Up until 1969, the composition of top American orchestras such as the New York Philharmonic was typically 100 percent white men. An experienced listener in 1969 and in the present would likely detect another difference too: The orchestra sounds better now. Steps have been taken to ensure that musicians who are hired these days are chosen purely for their skill, which wasn’t always the case.

What does any of this have to do with financial services? A great deal, surprisingly. A bank that overcomes entrenched bias in selecting its clients performs better. It earns higher revenues, and its agents make more robust commissions when they target clients based on data instead of stereotypes.

The bigger question is: How did orchestras root out prejudice from the audition process, and how can banks similarly identify and avoid biases to ensure a more diverse clientele, and better performance?

The first step is acknowledging you have a problem


nyphil 300x132 1
The New York Philharmonic Orchestra today

In the case of the orchestras, it wasn’t until a non-white musician in 1969 sued the New York Philharmonic for discrimination that the entire auditioning and hiring system changed. Although that musician lost the lawsuit—which charged that he was at least as skilled as his competitors but was passed over because he didn’t fit the Philharmonic’s stereotype of a “leading musician”—his legal action led to major changes in the way auditions were held across the U.S.’s “Big 5” orchestras.

In the wake of the lawsuit, hiring committees were established, which shifted the hiring decision from the sole responsibility of one person, the Musical Director, to a group of people. Musicians were asked to audition behind a screen. Over time, the hiring committees found ways to eliminate any cues to candidates’ identity during their audition, down to asking candidates to remove shoes before walking across the stage, to prevent the sound of women’s heels from signaling gender.

When the hiring directors didn’t know who the hopefuls were or what they looked like, they tended to recruit a far more diverse pool of musicians. Orchestras became more ethnically and racially diverse, and had a surprise side effect. Over time, the proportion of women grew to 44%. While the blind auditions were established primarily to counter racism, bias against women was also addressed. The auditions revealed that women were, in fact just as capable and sometimes better than men at playing classical music.

Bias can blind you

BETA Agents Male and Female
BETA Friends

The orchestra example holds important lessons for financial institutions: sales people—working under commission—will naturally seek out more profitable clients. However, they operate on the basis of assumptions about potential customers’ financial habits that are based more on intuition and perception than data. This instinct can cause both the salesperson and the institution to leave cash on the table by missing out on potentially profitable segments.

Women’s World Banking recently had the opportunity to examine the unconscious biases that can impact even those financial institutions that make a concerted effort to target women. For example in Nigeria, our partner Diamond Bank designed its BETA Savings accounts specifically to appeal to women, and feedback among women clients was very positive when the accounts launched in 2014. Two years after launch, however, only 37% of accounts were owned by women. Women’s World Banking took a close look at how the BETA Friends, or sales agents, recruited clients. We found that the agents’ underlying assumptions were leading them to overlook potentially active customers, namely women, who could increase the BETA portfolio’s performance and earn higher commissions for agents.

“But she isn’t a good customer!”

Women’s World Banking’s collaboration with Diamond Bank to identify the reason for the drop in the percentage of women revealed that the agents weren’t aware of how their unconscious biases were negatively impacting BETA’s goals. As far as the BETA Friends were concerned, they needed to make a living, so they sought out clients who seemed to fit the optimal customer profile to boost their commissions. Sounds reasonable enough, right? But the agents’ rationale was based on the misconception that men make more profitable clients. When we interviewed some of the agents about this assumption, they’d offer an anecdote about one or two of their “best customers:” male clients who had made particularly large deposits. In effect, the agents were defining who a good client is or looks like based on the activities of a few customers and pursuing people that fit that mental model.

Diamond Bank BETA Savings Client DataHowever, the actual data on average versus median deposits didn’t substantiate these assumptions. Women, it turns out, maintain higher median balances than men by a sizable 20%; as for men, the higher average deposits of a few clients skew the data. When we factored out the outliers, we found that active clients, both women and men, contribute the same value to the BETA portfolio. The agents who didn’t actively seek out women were missing out on a large segment of potential customers who could bump up their overall commissions. Just as the Big 5 orchestra directors’ assumptions had kept out candidates who didn’t fit the “leading musicians” image, bank agents’ underlying biases were keeping women out.

As musical directors did before that influential lawsuit, people tend to anchor their expectations based on anecdotal evidence and familiarity. Since white, male musical directors knew more white, male musicians, and observed them doing well at their jobs, then they routinely hired others who fit the same profile. Similarly, when bank agents accustomed to working with male customers noticed that some of them maintained extra-high deposits, then incorrectly generalized about their overall performance, they tended to prioritize men as clients even if they didn’t realize it.

BETA male female clients combined
This is what a good client looks like

Fight bias with data

Uncovering entrenched bias is just the first step. Maximizing performance means actively attempting to counter those biases. But there’s no magic to it: Often all it takes is a simple data analysis to reveal unconscious bias.

To reverse the gender trend in the BETA portfolio, agents have the opportunity to change their approach. As they walk through the markets to attend to customers and attract new business, their recruiting pool of attractive potential clients is suddenly much larger, now that they are aware that active women customers are just as valuable to them as active men – this has strong potential to be a financial win for both the bank and the agents.

Ensuring that financial institutions deliver better results for their clients, and a healthier bottom line for everyone? That’s music to our ears.