Call Me Maybe

As we head into Hanukkah, Christmas, and the New Year (followed by what is expected to be a fairly stormy winter), flight disruptions will be inevitable. If you’re experiencing flight delays or cancellations, and need to reach your airline’s call center, you may find this more challenging than usual since some airlines are “shutting them down.” Well, they’re not really shutting them down, they’re merely removing the option to call:

“The airline got rid of its customer service phone line, Frontier said. Passengers will no longer be able to call about reservations, or more urgent issues such as delays, cancellations and lost bags. 

Instead, Frontier passengers needing assistance will have to use an online chatbot, a 24/7 live chat available on the airline's website, social media channels or WhatsApp, the airline said.”

The phone line was eliminated ahead of the busy Thanksgiving travel period, and you may suspect that Frontier decided to shut down their call center phone line due to idleness … but alas, I assume it’s actually the opposite. Frontier has more complaints than any other airline… by a wide margin:

I know the number of complaints is not the same as the call volume, since many of the complaints are not directed to the airlines themselves, but to the TSA, etc., but I assume that there is a high correlation between the two. 

In fact, Frontier is almost synonymous to bad customer service. The so-called ultra-low-cost carrier is known for its penchant for cost-cutting, but the decision to eliminate their phone-based customer service is unique and the first among all US-based airlines. Breeze Airways, another ULC carrier  founded by David Neeleman (who also founded JetBlue), doesn’t have a call center either: 

“At Breeze, our mission is to save you time and money. That’s why you won’t find a call center phone number on this site but you will find plenty of other ways to get answers and reach us quickly and efficiently.”

(Note that you can find comments online on how to contact them, but I’m not a customer and didn’t want to test it). 

One imagines that the main reason for eliminating a call center is to cut costs (and we’ll get to that part), but it might be slightly deeper. At least, Frontier claims so:

“‘We have found that most customers prefer communicating via digital channels,’ an airline spokesperson said in a statement provided to TPG, noting that customers could still use those options to ‘get the information they need as expeditiously and efficiently as possible.’”

Of course, the issue is broader and doesn’t pertain only to airlines. Call centers have been a mainstay of customer service for the last 60 years. But are we observing the beginning of the end? 

The trend seems clear as many organizations are switching the term from call center to contact center (which may or may not provide a phone number). 

The question: Are there other options, and why are they not good enough...yet?

In other words … Why are reports of “the death of the call center” being greatly exaggerated?

The Numbers

As of 2022, the size of the call center market in the US was $23.9 billion, which amounts to 29.5% of the global call center market as of 2020. 

Alternatively, the contact center market, which includes all email, chat, text message, and in-app support, is much bigger. According to some estimates, it is worth $339.4 billion, and is expected to reach $496 billion by 2027.

Finding the exact number of customer service employees is not a trivial task to achieve, but according to some estimates, there are 3.4 million contact center workers in the US alone, compared to India and the Philippines (1.1 million and 1.15 million, respectively).

Alternatively, it is estimated that globally, there are currently 1.5 million call center employees —a number that is expected to rise 15% by 2022, reaching 1.8 million.

But as mentioned, call centers are becoming less about calls and more about “contact.” According to industry reports, 29% of all businesses have installed a live chat software, and 42% of businesses have incorporated a help desk system to improve customer experience. 

What’s the Tradeoff?

The tradeoffs are clear when comparing the four types of systems: traditional phone call vs. the “contact” options of live person chat, chatbot, and email.

The asynchronous methods (i.e., email or even live person chat) allow firms to reduce cost and pool resources over time (since issues can be addressed when there’s downtime, at least in the case of email). Additionally, everything is recorded (to the benefits of both sides), and there are no misunderstandings due to different accents, use of terminology, or general language barriers. 

More specifically, firms are able to use machine learning to automate a big part (including the option of automating everything through a chatbot), and according to industry reports, the global call center AI market, which was valued at $960 million in 2020, is anticipated to reach $10 billion by 2030. 

The downside of these asynchronous methods is that they increase the total resolution time significantly, and actually decrease the likelihood of a resolution altogether. They may be faster (usually there is little hold-time for a live person chat or a chatbot), but speed is not what consumers really care about. 

According to a 2017 survey, 27% of customers said their number one frustration with poor customer service is the lack of effectiveness, and 90% of customers worldwide consider the resolution of their problem as their top concern. Additionally, 90% of customers also believe that it’s very important to receive an immediate response.

There isn’t adequate research regarding what people like or dislike about these chats, but we can borrow from how people feel about automated menus, and extrapolate from there. In particular, 88% of customers prefer voice calls with a live agent instead of navigating an automated phone menu. The things that people find most annoying about automated phone menus are i) listening to irrelevant options (69%), ii) the inability to fully describe their concern (67%) and iii) the complete lack of human interaction (43%).

I haven’t researched this topic yet, but I have studied customer behavior in call centers extensively in the past, and more specifically, the notion of customer retrials (i.e., the reasons why customers call again even though their call was answered). In our study, we identified that a significant portion of customers call back, and ⅔ of those do so because the initial response was not satisfactory (the rest are due to high congestion). This isn’t to say that that is the case everywhere, but it’s a data point on the fact that people who call again are looking for a resolution. In other words: while customers who contact a call center want a fast response, they prefer that their issue be resolved in that single phone call (even if it takes a little longer) rather than have to call again. 

At this point: resolution time is still faster through a call-based call center. 

Why? Because many of the responses provided by live person chats and email are scripted, which means they essentially copy paste from a script, and that’s what makes them so economically efficient. And it’s not that call centers don’t follow their own scripted responses, but once they realize their script doesn’t offer the necessary solution, call-center representatives will try to find other ways to resolve the issue. As of 2018, 78% of customers associate better service with one that doesn’t sound like it was read from a script —a nine-point jump since 2014.

And indeed, 44% of US consumers still prefer phone or voice as their primary customer service channel —21% higher than any other channel available (with live chat coming in second at 23%).

Furthermore, firms that have tried chatbots, realize that they are a little premature: 

“Because of its limited customer service functionality, many companies are slow to adopt the technology. For instance, fashion retailer Everlane ditched the Facebook Messenger chatbot after it recorded high failure rates in 2017. Along those same lines, in 2018, Accenture reported that 53% of organizations ‘have no plans’ to invest in chatbots.”

But the trend is clear: call center volumes have dropped by 17% since 2015. 

The Future

Tech firms are trying to solve the fact that while very economically efficient, chatbots and live chats are not yet good enough. For example, MindSay, which was recently acquired by Laiye, is a conversational customer service chatbot, powered by AI. According to a testimonial of a customer relations executive at French Bee, Mindsay helped the airline reduce the number of incoming chats by 60%, and the number of projected agents by 50%. 

Another option is to add a human in the loop, just like Tymely, which recently completed its funding round: 

“Tymely, believes that human supervision in its processes provides a solution that enables human-level personalizations… The company, which today announced it raised $7 million to ‘make AI converse better,’ claims it uses AI-human hybrid tech to enable brands to provide email and chat support services in a more human, empathetic and precise way.”

Finally, another developing opinion is community-based support, such as the one used by Apple. In my paper with Neha Sharma (on the job market this year) and Achal Bassamboo, we study how to structure communities that help firms resolve support questions. The benefits are clear: cost reduction and the development of a community that supports itself while utilizing network effects. The issues: these groups are not known for their timeliness, and require building a thriving community, balancing different levels of expertise, and maintaining the willingness to ask and answer questions. 

And if you’re wondering about me … I’m generally impatient, and not much of a people’s person, so whenever possible, I try to use the live chat. By now, I’ve tried it with almost every firm I’ve interacted with, and I almost always end up on the phone. The systems are not ready, or maybe I’m too much of a complex customer. My mother always told me I was special, so maybe that’s it. Or maybe I know where to look for information (Google), and when I do need support, it’s usually complicated, which means I need to call. But as I’ve been experimenting with ChatGPT over the past few weeks, I’m as optimistic as ever on the future of chatbots for 99% of problems —but still not 100%. 

Coming back to airlines: this is one situation where timeliness and understanding nuances is crucial. Sure, we all try to avoid calling them, but if we call, we want to have a quick resolution. Otherwise, we may miss our next flight, or not make it to class on time…