Why Anxious Customers
Prefer Human Customer
by Michelle A. Shell and Ryan W. Buell
APRIL 15, 2019
When people feel anxious, it’s human nature to turn to others for help. For example, an apprehensive
taxpayer might want guidance from an IRS agent in interpreting a new tax law. A distressed patient
might want to speak with a nurse when making sense of his blood test results.
This type of behavior is quite common. Yet many companies in high-anxiety settings – like fnancial
services and healthcare – are funneling nervous customers to self-service technologies (“SSTs”) –
kiosks, websites, and smartphone apps – isolating them at the precise moment when they’re most
keen for connection. It is clear that these technologies are less expensive to ofer than human
support. But what’s less clear is the toll these self-service interactions may take on customers.
Is it an efective way of helping customers deal with their concerns? Or is it exacerbating customer
anxiety and doing long-term damage to service relationships?
In our research, we set out to understand these questions. Through two lab experiments and one
feld experiment conducted in fnancial services, we found that anxious customers interacting
through self-service technology feel dissatisfed with their decisions even when those decisions
appear aligned with their goals. Their dissatisfaction reduced their trust in the service provider. But
our results also reveal how a simple, and surprisingly low-cost, change – ofering access to a readilyavailable human – can reverse the negative efects of customer anxiety.
The domino effect of customer anxiety
We set our research in the fnancial service industry because it is riddled with uncertainty and
complex decision-making known to provoke anxiety and distress for its customers. In our frst
experiment, we simply wanted to understand how situational anxieties – those outside of the
company’s control – can crossover to afect the way customers feel about their service providers.
We developed an online investing platform to simulate the retirement planning experience. Over 150
adult participants from across the U.S. were told to allocate a hypothetical portfolio of $100,000
across stocks, bonds and cash over a series of multiple rounds with the objective of growing the
portfolio. As incentive, we paid them cash bonuses based on their performance in the simulation.
Some participants were randomly assigned to experience normal market conditions, which we
defned as having an equal chance at a year’s stocks, bonds, and cash returns drawn from actual US
history during each round of the simulation. Others experienced a greater probability of drawing
from the worst stock market years in U.S. history. As a part of the investing platform, we gave
participants access to historical performance information for each of the asset classes and the ability
to track their own portfolio growth to help inform their decisions.
Every few rounds, we asked participants to rate how satisfed they were with a decision they had just
made, as well as how anxious they felt at that time. Unsurprisingly, those experiencing more
downturns reported feeling twice as much anxiety as those facing normal market conditions. They
were also less satisfed with their choices, even though their portfolios actually outperformed the
stock market they faced on average. (Interestingly, those facing normal market conditions, who felt
less anxious, reported higher satisfaction with their choices but, on average, their portfolios
underperformed the stock market.
That dissatisfaction people felt with their own choices carried over to infuence how they felt about
the company that provided the investment platform. When they were fnished investing, participants
who had been the most anxious reported trusting the company less, despite the fact that it had no
control over the market they faced, or the investment decisions they made.
How does the opportunity to connect with a person help?
Since customer anxiety in online settings undermines customer satisfaction and trust, we wondered
if ofering customers the opportunity to connect with a real person might help. We repeated the
experiment from before, but this time, we randomly split over 200 participants along another
dimension: some were given the option to “chat with an expert”, others to “chat with another
investor” and others still had no option to chat.
We found that when people had the ability to connect with another person – either an expert or a
peer – the deleterious efects of anxiety were ofset. Although people facing rocky market conditions
were still anxious and again produced portfolio gains on average, those who were given the option to
chat felt the same level of choice satisfaction and trust in the frm as people facing a normal market.
What really surprised us though was that very few participants took advantage of the opportunity to
chat with someone. Although those who felt most anxious before the experiment were the most
likely to use the chat feature, merely having the option to access a person seemed to be all most
people needed to feel supported. This implies that companies deploying self-service technologies for
anxiety-provoking tasks might be able to put their customers at ease, and enhance their trust in the
frm, with a relatively low-cost change in design. Just knowing that we can chat with another person
– even if we don’t choose to do so – seems to make a big diference.
The value of being able to connect with a person
Taking these insights to the feld, we partnered with a U.S. credit union that had recently launched an
online loan application process. They wanted to increase the percentage of approved applicants who
completed the closing process and accessed the funds for their loans.
Over 200 applicants were randomly assigned to one of three groups: (1) those who received no
contact from the credit union until their loan decision had been made, (2) those who received text
messages with updates about the status of their approval process (e.g., “we have begun our review”,
“we are pulling your credit report”, etc.), and (3) those who received the same messages throughout,
with each message including the name and phone number of their loan ofcer, and an invitation to
reach out with questions.
Sending applicants text messages with a play-by-play of their approval process performed worse on
average than sending no messages at all, in that people who received the text messages and were
subsequently approved were less likely to move forward with their loan. In a post-hoc study to
understand why, we found that reminding people they are being evaluated – even though the
evaluation is an expected part of the service – increases their anxiety.
In contrast, we found that the probability of approved loan applicants moving forward with their
loans jumped from 64% to 80% when customers receiving those same play-by-play messages were
also invited to connect with a loan ofcer.
As automated service processes are being deployed to engage customers, it has never been more
important to understand how to balance touch and technology to deliver efcient and satisfying
experiences – ones that lead to trusting, long-term relationships. Our fndings suggest that using selfservice technologies in high-anxiety settings can be costly. Anxious customers left to fend for
themselves are less satisfed with their choices, and less trusting of the company with which they are
interacting. Merely ofering access to talk to a person can be enough to restore customer confdence,
improve trust in the frm, and strengthen long-term relationships.
shaifali-chaurasiaHey there, the activities suggested by my team are really what i have been looking for. It would be great if you can provide some suggestion,your notes should be least 5 pointers ,you should mention your leaning and also should write about how we can implement all these activities. It would be of a lot of help to me.
Thanking you in anticipation