By Nicole M. Hill
This work was a collaboration between Ling Hu, Dr. Kevin Nolan, and Dr. Nathan Carter. This piece is based on a presentation at the Strive 2019 Conference called, ‘Measuring trust when conversion isn’t enough’ by Ling Hu and Nicole Hill.
In Part 1 we discussed how Groupon was experiencing a tension between improving customer experience and conversion. The team believed that doing right by the customer was more important than short-term impacts on conversion, but we needed a way to measure the impact of following that direction. This led us to employ a psychometrics approach to measure consumer trust in Groupon’s ratings and review platform. In this article, I will make a case for why measuring trust is important and I will illustrate some different ways that our trust scale can be applied.
The case for measuring trust
When the project team started building our trust scale we weren’t thinking about trust broadly, we were just thinking about it in the context of our ratings and review platform. But we did have a sense that measuring trust was important when thinking about the quality of our relationships with customers, and we thought keeping trust intact was important when thinking about the lifetime value of a customer.
As it turns out, our intuitions were correct. Researchers who study trust at the macro level know that trust in institutions, such as the media and government, is declining and at an all-time low. We also have a trust gap within various segments of our population; with older people trusting more than younger people, not simply due to their age but their generational cohort, and educated people trusting more than the mass population.
Businesses are not immune to this phenomenon of declining trust. A study by Cognizant talks about how trust is breaking brands. It found that 57% of people will stop doing business with a company that has broken their trust and 40% plan to switch to a competitor due to trust.
Trustworthiness is not just a subjective opinion, it is a competitive advantage.
High-trust companies are outperforming in their sector with regard to stock price by an average of 5% and, in another analysis, the “Top 10” most trustworthy public companies have outperformed the S&P 500 by over 25% .
A recent study of over 7,000 companies by Accenture found that trust-related incidents are on the rise. 54% of the companies that they tracked experienced a drop in trust in the past two and a half years.
According to their analysis, when companies experience a drop in trust there is the potential for a drop in revenue as well. Their model conservatively estimates this drop to be $180B in lost revenue.
What’s more, trust incidents are public events that are more visible than ever before due to social media, the proliferation of blogs, and the 24–7 media cycle. Accenture claims that a trust incident is not a matter of if but when. To prepare, you must measure trust to know where you stand.
Clearly, trust is important. But why, as UX professionals, should we be concerned with trust?
First and foremost, as experts in user behavior, I argue that we are uniquely qualified to measure trust at our organization and we should be leading the way. When evaluating a new product or feature, we often think about success, what that looks like, and how to measure it. Many times, that boils down to conversion, other financial metrics like refund rate, or measuring page interactions. So it is up to us as user experience researchers to reframe our metrics by asking, “What does it mean to win with our customers?” From that perspective, winning is not just a matter of selling more short-term, but building deeper, long-lasting relationships that will build customer lifetime value. Trust is one thing to measure, undoubtedly there is more, and as researchers, it is our job to help define ‘winning.’
You had my curiosity but now you have my attention. (Django Unchained)
If you read part 1, you understand that building a psychometric (trust) scale is a time- and resource-intensive process that is best undertaken through collaboration with experts. Once you have created your scale, it can be applied in several ways, including driving qualitative research. Below are some ways that this scale can be applied.
Understand the nomological network
The nomological network is an additional step that can be used to assess the construct validity of our psychometric trust scale. It assesses how well the trust scale correlates with things it should theoretically be related to, such as different attitudes, intentions, and behaviors (Cronbach & Meehl, 1955).
We did this by deploying our scale along with additional questions that assess customers’ attitudes regarding Groupon’s ratings and review platform, Groupon and eCommerce in general, and various intended behaviors like purchasing or reviewing a deal.
Connect trust measures to business metrics
The power of the scale is that it allows us to explore the relationship between it and other business metrics. This can be done at any time, but an excellent time to do this is when you are assessing the nomological network. Any customer-level data you have, such as total spend, type of items purchased, engagement with certain online features, are fair game. I recommend socializing these types of scales throughout the organization in order to determine if there are other metrics that should be included for analysis.
Tracking trust over time and measuring the impact of change
At the end of the scale development process in part 1, our trust scale was composed of three subscales (i.e., trust in Groupon, trust in reviewers, and trust in content) each composed of four scale items. The most obvious use of this trust scale is to periodically measure shifts in trust over time such as every quarter. But you can also use the scale or a subscale to measure the impact of changes to your product. For example, if we added a feature to guide our customers in writing detailed reviews we could measure the impact of that change by using the Trust in Content subscale. We could administer the entire scale in this project or, for brevity’s sake, we could only administer the subscale.
Identify the largest gap
Another product of the scale development process is a set of mean baseline scores for each survey item. This allows you to pinpoint the strongest and weakest points in your experience. Just like how qualitative research informed the scale development process, now the reverse is true. You can conduct qualitative research to understand why trust is strong, weak, and everything in between for the specific items. From there, you can build a strategy to remediate the poor experience and test the success of this strategy by deploying the scale again.
Develop catered product strategy
This scale can also be used to look at differences in context and populations. For example, are there differences in trust based on the life cycle of the customer, new, old, and lapsed (i.e., those who have not shopped in some time)? In the case of Groupon are their differences between those who primarily shop certain types of deals (e.g., travel vs. restaurant)? If so, in the case of deal categories, we may want to consider changing the ratings and reviews experience to better support the nuances of shopping in those categories.
Benchmark against competitors
Finally, this scale is robust enough to allow you to benchmark yourself against competitors that have similar experiences, in our case e-commerce ratings and reviews platforms. We simply need to replace all references to Groupon with another company such as Amazon.
In addition to trust, what else should we measure? What other ideas do you have for applying these types of scales? Have you created a psychometric scale for your organization? If so, what kind and how are you using it?
Please leave your ideas and questions in the comments section.
Cronbach, L. J., & Meehl, P. E. (1955). Construct validity in psychological tests. Psychological Bulletin, 52(4), 281