Customer Discrimination in the Workplace: Evidence from Online Sales

Many workers are evaluated on their ability to engage with customers. We measure the impact of gender-based customer discrimination on the productivity of online sales agents working across Sub-Saharan Africa. Using a novel framework that randomly varies the gender of names presented to customers without changing worker behavior, we ﬁnd the assignment of a female-sounding name leads to 50 percent fewer purchases. Customers also lag in responding, are less expressive, and avoid discussing purchases. We show similar results for customers from different parts of the world and across workers. Discrimination directly affecting worker productivity has important labor market implications.


Introduction
Women work less outside the home, earn less, and run less profitable businesses than men on average, especially in low-income countries. Recent literature has focused on understanding why this might be the case (Blau and Khan, 2017). One potential contributor that has received relatively less attention is customer-side discrimination, a phenomenon whereby customers prefer to work or interact with certain types of workers. If customers are selectively biased against women by purchasing products less often or bargaining more forcefully with them, then female wages, promotions, and hiring prospects could be affected.

Methodology and Context
We investigate the role that consumer-side perceptions of gender play in determining the productivity of workers through an experiment at an online sales company in Sub-Saharan Africa. We study workers, specifically sales agents, who chat with customers online in order to answer questions and increase sales. Two aspects of the context provide a novel framework for estimating the causal effect of customer discrimination.
First, the names of workers, and hence their inferred gender, were randomised on a daily basis, providing plausible variation in customer beliefs about worker gender. Second, while agents were aware that their names were changed, they were unaware of the particular name assignment and hence their behaviour was not directly related to the name variation. Therefore, any change in consumer behaviour towards sales agents can only be a product of customers rather than the agents themselves.
The context of this study is important for three reasons. First, barriers to female labour force participation are higher in low-income countries (Jayachandran, 2015), and reducing gender-based pay differentials are central policy goals for governments and international in-situations alike (Bank, 2011;O'Donnell et al., 2020). Second, in Sub-Saharan Africa the high prevalence of conservative social norms that favour men over women as economic agents and business owners may contribute to heightened gender discrimination by customers. For example, data from The World Bank Development Indicators (WDI) show that households in Sub-Saharan Africa are more likely to agree that men make better business leaders than women and that women have no say in decisions on large household purchases (Jayachandran, 2015). Such norms may contribute to higher rates of customer discrimination. Finally, the service sector is growing across the continent, which makes issues related to gender-based differences in customer interactions increasingly relevant.

Findings and Policy Impact
We find that when sales agents are assigned a female sounding name, the likelihood of customers making any purchase falls markedly. Customers also purchase fewer total products, and the total value of their purchases fall. An exploration of mechanisms suggests these results are most consistent with customer disinterest with working with female agents, rather than differential bargaining or openly negative behaviour.
The results have a number of implications. First, these results suggest that female sales agents in our context may be more productive than their male counterparts if we hold fixed customer behaviour. This speaks to the "twice as hard" phenomenon whereby members of a discriminated group need to perform better than their counterparts in order to maintain their position in the workplace (Sofoluke and Sofoluke, 2021).
Second, workplace policies that give equal pay for equal work may not fully correct for the discrimination women face in light of customer discrimination. This is especially relevant in the service industry where employee pay is often tied to output (the number of sales, for example) through piece-rate wages. These results also imply that employers/institutions may have to take additional measures so that women don't feel compelled to create an online presence that obscures their identity.
Finally, these results speak to barriers that women may face in the labour market in Sub-Saharan Africa. Like many regions around the world, significant gender disparities exist in the formal sector, where the share of women who work full-time for an employer is below 15 per cent (Klugman and Twigg, 2016). At the same time, there exists a variety of models and empirical studies suggesting that improvements in gender parity can result in significant economic growth (World Economic, 2017). Governments across the continent are recognising this and have spearheaded a number of initiatives to address these issues, including powerful provisions supporting gender equality. Our paper provides additional evidence on an understudied barrier that institutions should address when designing policy responses.

Moving Forward
Moving forward the team is interested in investigating the extent to which customer side discrimination is exacerbated by the industry where workers operate, and by customers characteristics (including gender).