#28 No More Surveys, Please
A couple of weeks ago, I met up with the CEO of a local bank. We talked about the future of banking, the benefits and challenges of being a local entity, and, of course, the importance of customer experience.
At one point, I turned my screen to show how we typically measure business performance and customer experience with a journey-centric dashboard. He studied it for a moment, then frowned.
"Oh, we already do a customer satisfaction survey, and I really don’t want to send out more of those. It’s super irritating with everyone wanting to know what I thought about everything—from a package delivery to my last grocery store visit."
That’s when I realized—the example I was showing relied heavily on CSAT and other survey-based metrics.
And honestly? That’s pretty accurate for most businesses. Too much of what CX teams do revolves around survey-based research—which isn’t just inefficient but often counterproductive. Constant survey requests will likely lower a company’s standing with its customers rather than improving it.
Balancing Surveys with Automated Metrics
Surveys have long been the default method for gathering customer insights, but they often suffer from low response rates, biased answers, and survey fatigue. Many of the most commonly asked survey questions can be replaced with automated behavioral metrics that provide a more accurate and real-time view of customer satisfaction. Instead of asking customers to report how they feel, we can analyze what they actually do.
Just as an example, here are three common survey questions and how they can be replaced with automated metrics:
1. "How satisfied are you with our service?"
Rather than relying on self-reported satisfaction scores, we can use sentiment analysis and interaction analytics to determine how customers truly feel about our service. AI-powered tools can scan support tickets, chat logs, call transcripts, and social media mentions to detect positive, neutral, or negative sentiment. Moreover, user engagement data—such as frequency of logins, feature usage, and whether they complete key actions—can complement the picture.
If customers suddenly reduce their engagement or express frustration in chat logs, that’s an important signal—and an opportunity for further research and proactive outreach.
2. "How easy was it to … ?"
Rather than asking customers whether a process was easy, companies can detect real friction points by analyzing customer interaction data. If customers frequently abandon a form at a specific step, require multiple attempts or take a long time perform a common task, or contact support for help at a specific stage or place, it indicates issues that need immediate attention.
This data allows us to pinpoint exactly where customers struggle and thus where to quickly make improvements to remove friction.
3. "Would you recommend us to a friend?"
Instead of relying on self-reported intent through the commonly used and often criticised Net Promoter Score (NPS), we can measure actual customer advocacy through referral activity and engagement in loyalty programs. If a customer is truly satisfied, they are more likely to refer friends, leave positive reviews, and engage with branded content on social media.
These behavioral signals provide a more reliable measure of advocacy than simply asking customers if they “would” recommend a service.
What Survey-Based Questions Could You Replace?
By shifting from survey-based data to behavioral insights, we can get a continuous, real-time understanding of customer experiences without creating unnecessary friction. These automated metrics will never capture the full picture, but they serve as excellent early warning signals—helping us detect when customer satisfaction drops, engagement declines, or friction emerges.
Rather than relying on static, retrospective survey responses, we can use behavioral insights to trigger targeted qualitative research at the right time—asking “why?” when it really matters. This approach allows teams to investigate customer pain points before they escalate and respond faster, more effectively, and with real impact.
The question isn’t whether surveys should disappear completely—though it might feel tempting when we receive our third “Help us improve our services” request of the day. But it’s important to use them only when they’re the best or only viable option.
So, regardless of your industry, what survey-based questions could you replace by observing and interpreting customer behavior instead?