« Back to The Remarkable Credit Union blog

Technology & Anxiety: When Automation Isn’t a Good Thing

This piece, written by our CEO Cameron Madill, was originally published in the Credit Union Times.

When was the last time you felt anxious? You know the feeling: stomach churning, unable to focus, and pursued by a nebulous but frightening vision of the future.

Okay, now that you’ve latched onto that feeling, let’s ask a second question: would you ever want to make your members feel that way when they interact with you?

I assume the answer is “no,” but you might be inadvertently causing your members anxiety, even when you are trying to make their lives easier. Here’s how:

Two Fascinating Experiments

Does technology truly improve member services? This was a key question posed by Michelle Shell, a doctoral student at Harvard Business School, at Filene’s recent research event Inside Out / Outside In. The question was the title of her compelling presentation, but an equally relevant title could have been “Buyer Beware!”

Experiment #1: Text Updates for a Loan Application

Here’s the first experiment: let’s take a standard loan process, and let’s see what happens when we add personalized, automated text updates to the process. Presumably more information is a good thing, right? Don’t we all want more information? Well, it turns out, like most things in life, the answer is, “it depends.” Michelle and her team tracked three different scenarios, comparing them against each other on the basis of how many of the approved loans were actually funded (in other words, the member accepted the loan and received the money). Here are the scenarios:

  1. Standard loan process, no text updates
  2. Text updates at each step of the loan process that included the name of the loan officer
  3. Text updates at each step of the loan process that included the name of the loan officer and their phone number, telling the member to reach out if they had any questions
Graph of probability of loan acceptance
Shell, Michelle A., and Ryan W. Buell. “Mitigating the negative effects of customer anxiety through access to human contact.” Harvard Business School Technology & Operations Mgt. Unit Working Paper 19-089 (2019).

Isn’t that fascinating! The automated text updates — even including the personalization touch of including the loan officer’s name — slightly decreased the success rate in getting members to accept their funded loans, whereas the texts that included contact information and the opportunity to make human contact dramatically increased the success rate. A 12% increase may not seem like much, but it could potentially transform a credit union’s financial results for the year.

Shell hypothesized that the reason for the drop is that a loan application is an anxiety-inducing situation. Automated reminders that you are being evaluated heighten than anxiety — ultimately damaging trust and hurting your business results. Yet in Scenario #3, the option of human contact had the opposite effect — it built trust and improved results. Also noteworthy was the fact that only a small percentage of the loan applicants in Scenario #3 actually reached out to the loan officer. Apparently just the reassurance of knowing that you have the option makes all the difference.

Experiment #2: Self-Service

Self-service has long held the promise of letting consumers do whatever they want, whenever they want, wherever they want. But it is not quite so simple. In the second experiment, Shell dug into the concept of anxiety. She created an experiment where people could use an interface to invest a hypothetical $100,000 in a mixture of stocks, bonds, and cash, and then returned updated results over a series of rounds by updating the portfolio based on the performance of the market (more on this below). To make the experiment more real, participants had the possibility of winning real money if their portfolio outperformed other participants.

Portfolio screenshot
Shell, Michelle A., and Ryan W. Buell. “Mitigating the negative effects of customer anxiety through access to human contact.” Harvard Business School Technology & Operations Mgt. Unit Working Paper 19-089 (2019).

There were two groups here:

  1. For one set of participants, the market performance was based on randomly picking years from the last 76 years, resulting in overall positive portfolio performance (call this Team Contentment)
  2. For the other set of participants, the performance was based on the worst years over the last 76 years, resulting in generally very negative performances (call this Team Anxiety)

Then each group of users in each scenario were shown three self-service options:

  1. The normal interface
  2. The option to chat with an expert
  3. The option to chat with a peer

The results were fascinating. For Team Anxiety (those getting very poor stock market returns), having the ability to chat with either a peer or an expert greatly improved their satisfaction. But for Team Contentment (those getting good stock market returns), the option to chat with a peer greatly reduced their satisfaction! They appreciated the opportunity to chat with an expert, but somehow, knowing that peers were out there decreased their satisfaction. Once again, very few users actually used the feature to reach out. Shell hypothesized that the negative impact was due to a contented user suddenly being made aware that there were other investors out there, many of whose portfolios were probably performing better.

The moral of the story? Offering the option of a human connection to your credit union is compelling, especially for an anxious consumer, but done in the wrong way it can create anxiety where none existed before.

The Risks of Automation

Shell also noted that the desire for automation has not always fulfilled its promise: “In financial services, self-service technology use has been associated with simultaneously higher use of full service channels; and in healthcare, web portal use stimulates higher call volumes for complex, ambiguous tasks.”

Shell’s presentation confirmed my long-held belief that while it is painful to not offer a service or channel that members are asking for, it is far better not to show up at all in a channel than to show up poorly.


Here are my three core takeaways for credit union marketers or anyone involved in implementing member-facing technology:

  1. At high-anxiety digital touch points (this includes pretty much any loan product), you must provide a pathway to human contact. It could be chat, a phone number, or even a link to a branch listing with hours and addresses. With third-party online application tools, this not always easy, but don’t be afraid to put some pressure on your vendor. Adding an alert bar at the top or bottom with your phone number is not a complicated programming request and something they should at the very least be able to implement..
  2. Self-service technologies like chat, online banking, websites, and online applications, can stimulate more contact depending on the situation and implementation. For example, consider the scenario of being emailed your full medical test results from blood work… now you have even more questions for your doctor than before! Make sure you are equipped to respond to additional questions, and whenever you can, present information in clear, jargon-free, member-facing language.
  3. The presence of an option to connect with a human can significantly improve the business results you are getting and increase consumer trust… even when no one is using it!

As you explore and expand your suite of self-service solutions, remember that your goal, first and foremost, is to make your members’ lives easier. You may not get everything perfect right away (in fact, you probably won’t), but incremental improvements to your digital tools based on these principles — including member surveys, focus groups, and secret shoppers — just might get you a substantive increase in funded loan applications, along with some very happy members.