I asked 10 customers to describe their support experience from their perspective. Here's what I learned.

About three months ago I did something I hadn't done before.

I reached out to 10 customers who had contacted support in the previous 30 days — a mix of customers whose tickets had been marked resolved quickly and some that had taken longer.

I didn't ask them to fill in a survey.

I asked them to spend 15 minutes on a call walking me through what the experience was like — from the moment they realised they had a problem to the moment they felt it was resolved.

Here's what I heard that I didn't expect:

1. Finding how to contact us was harder than I thought

Three of the ten mentioned that finding where to submit a support request wasn't obvious. I had assumed it was.

We had a "Contact" link in the footer. Apparently that's not where people look when they have a problem. They look in the product itself, near where the problem is happening.

We added a contextual help button inside the app within a week of these calls.

2. The first auto-reply created anxiety, not reassurance

Our auto-reply said: "Thanks for reaching out! We'll get back to you within one business day."

Two customers mentioned this made them more anxious, not less — because "one business day" felt long when they were stuck on something they needed for work.

We changed the auto-reply to acknowledge the specific nature of the request and give a more honest estimate.

3. Customers didn't know if their ticket was still being worked on

Four customers mentioned a period of uncertainty where they didn't know if their issue was still being looked at or had been forgotten.

Our tickets were being actively worked — they just hadn't received an update.

We added a simple rule: if a ticket is open for more than 4 hours without a resolution, send a brief update.

4. "Resolved" didn't always mean resolved to them

Two customers whose tickets were marked resolved in our system still had the underlying issue.

They hadn't replied because they didn't want to be a bother — not because it was fixed.

We now send a follow-up 48 hours after marking a ticket resolved asking explicitly: "Is everything working as expected now?"

None of these were large changes. All of them came from 15-minute conversations that no survey would have surfaced.

If you haven't done a customer walkthrough of your support experience recently — 10 conversations will tell you more than months of CSAT data.

What method do you use to evaluate your support experience from the customer's perspective?