Earlier writing. I wrote this during my time in financial services, and it’s left here as it was published. More earlier writing
Three Zeros and a Love Letter
Last week I closed out day one of the NEFMA Fall Conference, and the moment that landed hardest was a survey response I’ve been carrying around for years.
An online banking survey came back with three zeros. Likelihood to recommend, zero. Ease of use, zero. Overall satisfaction, zero. On paper, that’s a customer with one foot out the door, and if you only looked at the scores, you’d flag her as at risk and maybe send her a win-back offer.
Then you read the comment.
“I love your branch staff, especially Linda. She’s been my banker for 20 years and I would never bank anywhere else. Put the bill pay button back on the home screen.”
That’s a love letter, with one complaint about a button.
This happened all the time when I ran survey programs. People would give us a score that told one story and a comment that told a completely different one. And look at what she did. This was an online banking survey, and she told us about Linda anyway. People tell you whatever’s on their mind, no matter which question you asked. The score answered the question we asked, and the comment told us what she actually wanted us to know.
The feeling and the driver
When I talk about sentiment analysis, this is what I mean. You read what people wrote and figure out two things: how they feel, and why. The feeling is what they tell you. Frustrated, grateful, anxious, confused. The driver is the reason behind it. The bill pay button. The banker who knows your name. A loan process that drags on for weeks.
Drivers come in two kinds, and it helps to know which is which. Some only get mentioned when they break. A loan that closes on time never gets a thank-you note, because it’s expected, but when it goes badly people tell you about it, and then they tell their neighbor. Those are the ones to fix, and they don’t need to be fancy. They just need to be smooth.
Others get praised without anyone asking. Linda is one of those. That’s what you feature, because your happiest people are already telling you what makes you different, in their own words, which is usually better copy than anything we’d write ourselves.
You already have the data
For a long time, reading all of this meant a printout, a highlighter, and a few weeks to find one theme. By the time you had an answer, the frustrated customer was already gone. That’s changed. The AI tools a lot of organizations already pay for can read every comment, tag the tone and the topic, and help you find what’s driving it, in an afternoon. You don’t need a data science team to get the juice anymore.
And the best place to start is probably the one you think about least: your call center. People tend to call when something’s wrong, but on the phone they’re kinder, and they say more than they’ll ever type into a survey. That’s exactly why phone surveys cost so much. Your agents are already on the phone, every single day. Have them ask one or two questions before the caller hangs up. How do you feel about us? What would you change? Write it down. You just built the most valuable survey you have, for the cost of a script change.
It works at the door, too
I’ve spent the last few weeks in a very different world. At The Outreach Team, the conversations happen at real doors, with organizers who hear what people care about all day long. It’s the same lesson. The people closest to the ground already know what their communities need. The job is making sure what they hear actually reaches the places where it can change something.
So here’s my ask. Pull a handful of comments this week, from a survey, a call log, wherever you have them, and read them for the feeling and the driver. I’d bet you find a Linda in there.
And if you want a head start, the prompts I use are free on my site: Prompts that listen.
