What’s the best way to give feedback that prompts action?

Interventions
Survey feedback
Experiment Type
Field Experiment
Goals
Increase engagement
Outcomes
Reduce expenses
Focus Areas
Product
Behavioral Concepts
Salience Framing Scarcity Information avoidance
Partner
WSECU
Partner Type
Bank/Credit Union

What Happened

It partially worked. While the forms of feedback (i.e. numbers, faces, or terms) did not affect a respondent's likelihood to take action, showing sub-scores slightly increased such likelihood.

Lessons Learned

Subscores may provide respondents more concrete ways to imagine taking action compared to just than the overall score, which feels more abstract. Additionally, financially vulnerable individuals may be overwhelmed by a poor situation and may tend to avoid confronting the problem.

Background

Many financial institutions and service providers have drawn on the increasing availability of financial health metrics such as the Financial Health Network’s (FHN) Financial Health Assessment to benchmark their members and clients.

While these metrics can be useful as a way of summarizing past behavior and comparing someone’s circumstances to others, they often lack the opportunity to take action based on that feedback. Even worse, sharing these results in the wrong way could have negative consequences: sharing negative feedback at the wrong time potentially could demotivate individuals from engaging in behaviors that would improve their situation.

We partnered with Washington State Employees Credit Union (WSECU) to better understand how we might display the results of a financial health survey in a way that is meaningful and provide targeted recommendations for actions that members can immediately take based on their responses.

Key Insights

In early 2020, WSECU offered members the chance to take the Financial Health Network’s Financial Health Assessment. The assessment worked as follows: after completing a survey, members were given an overall score based on their responses. In addition to the overall score, members are also given four sub-scores based on specific questions related to spending, savings, borrowing, and planning. A member’s circumstances across these categories are further categorized as “vulnerable,” “coping,” and “healthy.”

In addition to allowing a credit union to have a better understanding of their members and their financial needs, the individualized survey assessment results had the potential to increase a member’s motivation to change their behavior. However, there are several reasons that the survey’s current design might actually discourage members from seeking strategies for improving their financial health:

  • The survey currently offers little to no context for the overall financial health score. Without contextual cues as to how to interpret their responses, members may struggle to understand what the score means for them.

  • Relatedly, the results of the survey are presented in categories that may be unfamiliar to the member. This could further exacerbate the tendency to shut down and avoid undesirable information.

  • The survey may also provide members with too much information. Past research has found that “information overload” can increase the likelihood someone shuts down as well.

Experiment

We hypothesized that communicating the results of the survey in ways that showed less information and presented the results in a more familiar form would increase the likelihood that the survey increases motivation.

In order to learn more about how to design the feedback for WSECU’s financial health survey, we designed a survey-based experiment where respondents received feedback about their overall scores or both overall scores and subscores using numbers, faces or words.

Results

We found that presenting the score using social cues of acceptability like faces or terms was not any more effective than the numeric score (p=0.27 & p=0.36).

The analysis found that including sub-scores slightly increased the likelihood to take action (p=0.12). We believe that the sub-scores might provide respondents with more concrete ways to imagine taking action compared to just the overall score, which feels more abstract. More generally, the results also show an interesting insight into who is more likely to be responsive to the results of a financial health assessment.

Respondents that were classified as “healthy” and respondents classified as “vulnerable” were significantly less likely than people with “coping” profiles to seek additional information (p=0.07 & p<0.001). These results reinforce the idea that financially vulnerable individuals may be overwhelmed and may tend to avoid confronting the problem.

These findings will inform an intervention we are currently designing with WSECU focused on the financial health score that will launch with their members in 2020.