Anti-Bias Policies That Really Work in Customer Service

AI is transforming the world. And it starts with the right compute. ARM is the AI compute platform trusted by global leaders, proudly NASDAQ listed, built for the future. Visit arm.com discover. Before we start the show, I have an announcement. Some of you know that this podcast has been around since 2006.

Hello and welcome to the very first edition of the HBR IdeaCast, a soon to be bi weekly audio show from Harvard Business School Publishing. My name is Paul Michelman. I'm an executive editor here and I'll be your host.

Between that first episode and this one, number 826, there have been some changes to the podcast icon, to the way you listen to the show, probably to the hosts. The theme music has gone from that acoustic guitar to driving electric guitars.

Welcome to the HBR IdeaCast from Harvard Business Review. I'm Sarah Greene.

And on to the strings of this episode's musical theme. Despite all these changes over the years, some things have remained the same. IdeaCast has always offered insightful conversations with leading thinkers in business and management. And another constant has been someone named Adam Buchholz. You hear Adam's name in the credits. He's had his hand in this podcast for 13 years. He produced hundreds of episodes before stepping up to manage HBR's podcast portfolio. Now he's moving on to a new challenge in a new organization. Adam, you'll be remembered and missed. All the very best wishes to you in the future. And with that, onto our show. Welcome to the hbr ideacast from harvard business review. I'm kurt nickish. The list of companies paying the price for bias against customers is a long one. Like the $50 million settlement that JPMorgan Chase paid after its mortgage brokers charged black and Hispanic borrowers more than white ones. Or the lawsuit that Sephora settled with Asian customers whose accounts had been blocked, suspecting them of purchasing items in bulk to resell them. Or the reputation cost paid by Starbucks. When a manager called police on two black men who didn't order anything, the manager assumed they were lawyers. Turns out they were waiting for a friend. On this podcast, we've talked a fair amount about bias in hiring, in who gets promotions or who gets funding for their startup. But one of the biggest places bias is at play is in the everyday interactions between companies and their customers. And with greater awareness in recent years towards these risks and real costs, many companies are looking for concrete ways in which they can improve these customer interactions for the better. That's where today's guests come in. Alexandra Feldberg is an assistant professor at Harvard Business School. And Tammy Kim is an assistant professor at the University of Virginia Darden School of Business. They've been studying how to combat bias at the customer interface. And they're the authors of the HBR article, Fighting Bias on the Front Lines. Ali, thanks for coming on the show.

Thank you. It's great to be here.

And Tammy, thank you too.

Yeah, thanks for having us.

So I'd love to start just by finding out what drew each of you to study this to look deeper at companies and bias at this customer relationship level.

Part of what I love about the research that Tammy and I have been able to do together is that we bridge disciplines, I think, about people within the companies. Tammy, being a marketing professor, thinks about the consumer side of things. And so we found this research to be sort of a great synthesis of our interests.

I think also aside from our academic experiences, I think both Ali and I can quite confidently say, as avid shoppers, we've had many interactions where oftentimes you would just psychoanalyze a lot of interactions, like, why do we get a certain discount in this particular interaction? There are just countless number of interactions that we have as consumers for both very big and small purchases. And that's also one of the reasons why we thought it was so important to explore this domain, given the sizeful amount of time we spend with service providers.

Yeah, absolutely. And I think that the other thing that I think has been captivating for us has been just the really subtle ways that bias can manifest in these exchanges. So it's not just about whether an exchange occurs, but it's about, you know, how much does someone go above and beyond to be helpful to you? What's the manner in which they actually deliver the service? And so these elements of the exchange, I think, are often really subtle, but really important to our experiences as consumers.

What are some of the main ways that you've seen bias show up in customer employee interactions?

We key in on three main ways that we see bias in these interactions. So the first is just kind of the exchanges, the sort of provision of core products and services, you know, actually like going and being able to get that vegan poppy seed muffin that I got this morning. The second is in terms of how willing people are to go above and beyond. We call this sort of the extras that happen in the interaction. So, you know, as someone, you know, when you go to a hotel, are you going to be offered a room upgrade for free? Then the last is, is etiquette. So really, this is the manner in which service is Delivered. So, you know, like if someone's smiling at you, do you feel like, you know, the body language is inviting and warm?

In your research, you find that customers with more white sounding names, for instance, can be called by honorifics more often. Can you give some more examples of just things that you find really emblematic of bias?

We have a bunch of studies that really think about the information that service providers, people at the front lines are willing to give and sort of the volume of information. One study where we look at hotel workers who are the front lines, we basically reached out to these service providers with people with different names to signal things like race and gender. And we ask really simply for restaurant recommendations. And the thing that I always find so interesting about our findings here is that there are actually differences in the number of restaurants that people with different names get recommended. So it's fascinating to me again and again that there should be no difference in the number of restaurants that people get recommended. And we see that that is the case. So to me, that's the kind of thing that I find so interesting. So even if you're willing to help someone, or even if people are willing to provide the help, the information is quantifiably different.

In addition to the number of restaurants, we also saw that when emails came from white sounding names, hotel service providers were much more likely to go provide extra information. So it'd be things like, you know, check out these museums in our local, in our city, or like, you know, there's a library around the corner that's like lovely to check out. It'd be like information like that that in our email inquiries didn't ask for, but still kind of going back to what Ali was talking about, how the information was quantifiably very different. It wasn't just in response to the direct inquiry, which was about restaurant recommendations, but also so much more than that.

What do you think is happening in the minds of people who are doing that? What's the root thinking there?

From my perspective, I do not think that the differences that we see across these service providers is conscious. You know, people aren't intending to systematically give different information or resources to customers. I think it's our experiments aren't really picking up because they're kind of across different organizations. They're not picking up sort of one person's bias. It's sort of in aggregate. And so, you know, in any, for any one person, it's not, it's not like someone is necessarily even being unhelpful or in some ways it's Sort of like people are just willing to give more information or better services to some people than others. But I don't think it's intentional.

You know, oftentimes we think about discrimination as like somebody being hostile against someone. But I think in this case, you know, we're seeing that discrimination can also happen from people being nice, because here it's people are being nicer to certain groups of people.

Some people might say this is a human thing. It's hard to tackle. It's very nuanced. Why should companies buckle down and address this?

Simply put, companies are not taking advantage of the resources, the talent they have at their disposal and leveraging the people at the front lines to, you know, like, if they're not ensuring that they're sort of consistent, equitable service. So if you can act in a really favorable way toward one person, ideally you want, you know, that to be happening across every service interaction. You know, to find new customers, to bring people kind of into your business. And so I think that that's, to me, that's the business case for this.

So how do you start as an organization? How do you start recognizing and confronting this issue and. Yeah, and addressing it?

We, you know, emphasize the importance of actually first diagnosing the problem. Oftentimes we see that companies like to jump to solutions. They might create like a diversity task force or immediately order, you know, shut down stores and order all the employees to go through implicit bias training, like things like that. But without really understanding what's actually going on in your front lines, there's just no way that companies can effectively address that problem. So what we advise in our article is before you do anything, you first need to diagnose the problem. And you can do this in a few different ways. But I think in our minds, the most important step is to just go out and talk to your customers and make sure that you're talking to a diverse set of customers. So not just your prototypical group of customers, like the customers that tend to shop most often at your store, but, you know, go out and talk to people who don't necessarily aren't your prototypical customers to figure out what their experiences are like, if there are any specific experiences that or anything about your company that's turning them away.

I mean, that's a classic, you know, management advice, right? Identify the problem before you go out and try to solve it. Is it that people think they know what the problem is because they see the effects and just kind of jump to conclusions? Or why is it so hard to step back and assess what the problem really is.

There are probably a few different things that are going on. One of them is like sometimes a big scandal erupts like in the case of Starbucks, and they feel as if they need to do something to address that. So it's a very reactive process. But you know, just because they, I guess like, at least in the Starbucks case, like just because there were two black men that were arrested doesn't mean that's the only problem that's happening at the front, that might be happening at the front lines. What would probably be more effective is to for companies in those types of situations to take a step back and actually like fully assess the situation to come up with more comprehensive solutions. But there's also probably, you know, in light of many different social movements that have happened over the past few years, there's also an increasing amount of pressure from consumers.

Yeah. There was a survey in 2020 by LendingTree that found that quarter of those U.S. respondents they surveyed had stopped patronizing a company because it was accused of racism.

Right. So there's an increasing pressure from consumers as well because they're no longer caring about what kinds of products or services they're getting from companies, but they actually care about what kinds of values companies stand behind. You know, from that pressure. Maybe companies feel like they should act.

Yeah. How does a company know that it has a problem like this before it gets to this reactive stage?

In some ways there's a positive thing here which is that companies are better equipped than ever before because of the data that they collect to know, you know, how consistently and equitably they're delivering service. Given all of the data that many companies collect today, I think ways to identify a problem are really to make sure, like Tammy was saying, that there's a kind of an ongoing conversation with a diverse set of customers that is happening to really understand the customer experience and especially the service they're receiving. It's important also to figure out ways to monitor and really be vigilant about monitoring the consistency of service that customers with all sorts of different attributes receive. And so that I think is a really important kind of an ongoing step to take. We talk a little bit about this in the article, but I think this can't be a one off thing. I think it's an ongoing process that we hope companies engage in.

So what are some good ways forward?

There are kind of few different interventions that we suggest in the article. One of them is exposure. So just making sure that as a company you are exposing your service providers to as many diverse backgrounds as possible and that they are able to serve, like, as diverse set of customers as possible. So maybe this entails them revamping their training materials so they're reframing their prototypical customers to be heterogeneous rather than just one specific type of customer. Another one is standardization. If companies can make sure that there are specific ways that service providers should behave when interacting with customers, you know, especially as it comes to three E's that we specified, which is exchange etiquette and extras, then hopefully they can start to get at minimizing the gap in the types of service that are being provided to different types of customers.

Yeah, it's interesting to think about the scripts that servers at restaurants follow, and it can seem almost automatic, but it's fascinating how that standardization does create a more consistent and perhaps more equitable experience for people.

You know, there are some companies that. There's a national restaurant chain that specifies that instructs their all their frontline workers to greet their customers a certain way. In many of our conversations with different study participants, what we've seen is that oftentimes participants feel as if they're being discriminated against at, like, at storefronts or when they're boarding airplanes because they simply don't get greeted. So companies, for instance, could standardize the ways in which they greet customers or the ways in which they approach their customers such that there's as little variability as possible.

And I think that often is. You know, I think excellence is something that many or most companies strive for in the service that they deliver. But I think that equitability in service, I think to the extent that we can start to see that as either on par with or the same thing as excellence and sort of consistency, that I think can go a very long way. The last piece that we highlight in the article is the importance of making sure that people within an organization at the front lines are held accountable. So studies have shown that people are less likely to behave in biased ways when they think they're going to be held accountable for how fair their actions are. And so to the extent that you can instill a sense of responsibility and maybe it's a metric that allow you to bias and you can talk about both signals, that you're serious about eliminating discrimination and also can help people understand their own behaviors and sort of be accountable for them, I mean, you've been.

Studying this as the COVID 19 pandemic has been impacting customer interactions. I just wonder if online retail is better or worse. When it comes to bias and customer interactions.

Yes. So actually a lot of the work that we've done has been before the pandemic. But, you know, I think one thing that's important to point out is that, you know, pandemic or not, whether interactions are occurring online or offline, there are just many ways that biases can manifest. It just that perhaps the ways in which companies can capture the ways in which those biases may manifest might be different. So, for instance, you might think about. Let's suppose that because of the pandemic, a lot of these service interactions have more moved online. Then the ways in which companies can track these or where biases could manifest might be in the types of language that's used or the number of exclamation marks that are being used in like service exchanges, things like that, as opposed to the tone of voice being used or extent to which service provider smiles at you if it were in offline context.

Ali, you talked about data earlier. Is this something that big corporations have an easier time doing with their resources and their bent for standardization, or do startups and small businesses have a better chance at changing behavior because they're just that much closer to the front lines?

I think that. I'm not sure they're necessarily. It needs to be seen as a trade off between big or small. I think that sometimes there just might be. The data that's being collected might not be the right data to understand whether bias is even existing. And so, you know, I think we kind of think about it as sort of data collection. What data are you collecting? And then with the data that you have, how are you looking at it? And I think any company can be empowered to see, okay, what information can I get on an ongoing basis about the behaviors happening at the front lines? And, you know, what are creative ways that I can look at it, analyze it, kind of link it to customer attributes like Tammy was talking about earlier, that might be very revealing about things that we might not even be aware about.

Suppose that in your conversations with different customers, like in focus groups, you find out that many customers complain about call time, how long they have to wait when they have to chat with a service agent.

Yeah.

In which case, you know, that would direct you to your data to see if you're actually collecting information on that. If you have it. Great. But have you looked at the whether call time rates vary across different customer attributes? So oftentimes what we're seeing is that companies don't really dissect their data. They just kind of see it as aggregate, like, oh, what is our average call time? Wait this month. And then has that improved or has that gotten worse? Rather than dissecting it by looking at, looking at how that might vary across different customer groups. So that would be one example of ensuring that you're using your data effectively to make sure that you're incorporating equity into how you're looking at service performance.

A lot of this sounds like it has to come from the organization, it has to come from the top. It sounds a little institutional. If you're part of that chain that's bringing this action to the customer interface, what's your role?

I think it's a great point. There's so many sort of. We're taking kind of an institutional perspective in this because that is a really important level to be affecting change from for leaders and managers. I think it starts with values. So again, this gets back to this idea of emphasizing consistency and equitability in service. So really making those things kind of important, well known, clear across the organization as things that are really valued and how customers are treated. I think that that's kind of a first step. I think the second thing is getting back to this whole point about diagnosis. Really making sure that, you know, if you're in a position where you have a lot of data at your disposal, where you manage a bunch of folks who are interacting with customers, make sure that you have a thorough diagnosis of what's going on. And only then, once you understand the situation, would we kind of really recommend intervening. But I think those two things, I think the values and then following a process of diagnosis and then interventions would be the thing that I think could cut across any organization.

We started this episode with some examples in the introduction of companies that took big missteps and paid a price for it. Are there companies out there? Are there examples out there that you look at and you think these are great models for what companies can be doing to take steps in the right direction?

We've seen a lot of focus that from organizations on the issues of diversity and inclusion in terms of in hiring. But we actually aren't seeing that much effort and focus being given to issues of diversity, equity and inclusion at the front lines. What we're hoping with this article is that we can engage in more conversations with companies to make sure that we can encourage them to embrace change and to start thinking about these issues and start engaging in these dialogues going forward.

Allie?

Yeah, I think that, again, I think I'll maybe kind of take a positive perspective on this. There's so much data that companies have that they can take advantage of to kind of try to understand these interfaces. And there is a lot of attention hiring. I think we know a lot about hiring and where bias can happen and some interventions at points of hire that can be really effective. And I think we can take those learnings and apply them to people at the front lines. And so I think this is a place that I think both of us have been found interesting that there's not, there's not as much knowledge about it. It just hasn't been so much of a focus from our perspective, at least in kind of the academic literature. And so, yeah, I think, like Tammy said, we're excited for this to be a place that companies increasingly focus.

Ally and Tammy, thanks so much for coming on the show to talk about this.

Thank you for having us.

Thanks for having us.

That's Alexandra Feldberg of Harvard Business School and Tammy Kim of the University of Virginia Darden School of Together they wrote the HBR article Fighting Bias on the Front Lines. You can read it in the November December 2021 issue. And@hbr.org one company they mention in that article is Ritz Carlton, where any employee can spend up to $2,000 to improve a customer's experience. We did an episode on that titled Setting a high bar for for your customer service. That's episode 7 17. This episode was produced by Mary Dew. We get technical help from Rob Eckhart. Adam Buckholtz is our audio product manager. Thanks for listening to the HBR Ideacast. I'm Kurt Nickish.