Why I Still Do Customer Support Myself

First thirty minutes of every workday: I open the support inbox. Not Slack, not email, not the dashboard. Support.

We have people who could do this. We have people who do do this — they handle the bulk of volume. But I'm still in there every day, picking up some tickets, reading the ones I don't pick up, and watching for patterns nobody else would see.

Here's why.

Support is the only honest feedback loop

Customer interviews are biased. The customers who agree to be interviewed are usually the ones who already like you. NPS surveys are biased. The people who fill them out are at the extremes. Sales calls are biased — those people haven't bought yet, so they're describing a hypothetical version of using your product.

Support is the one place where you hear from people who actually use the thing, when something just broke, in their own words. There is no filter. There is no "how would you feel if" framing. There is a person who tried to do a thing and the thing didn't work, and they're typing fast.

If I miss that signal, I'm running the company on second-hand information.

The patterns are obvious only to me

Our support team is great at resolving individual tickets. That's their job and they do it well. What they can't easily do — because they're heads-down in queues — is notice that three different customers this week described the same friction in three different ways.

I notice it because I'm not trying to close tickets. I'm scanning for patterns. Last quarter, that scan caught a feature gap that became a $40k/month expansion deal once we shipped it. The team would have eventually surfaced it — they would have done a sprint retrospective and tagged the tickets and run a report. By then we'd have lost the customer who wrote the first one.

Loom beats paragraphs

Tactical thing I do that nobody trained me on: any support response that would take more than three sentences to type, I record a Loom instead.

Time to write three paragraphs: maybe 10 minutes. Time to record a 90-second screen recording of me doing the thing: under 2 minutes. The customer gets a better answer and they feel like the founder personally cared. Response rate to the follow-up question "is that what you were looking for?" goes up roughly 4x.

I keep a folder of the most-watched Looms and turn them into help docs. Half our knowledge base started as a one-off Loom for a single customer.

The argument I keep losing

Every advisor I've talked to has told me to stop doing support. "It doesn't scale." "It's a $50/hour task." "You're the bottleneck."

They're right that it doesn't scale. They're wrong that it's a $50/hour task. The hour I spend in support produces better roadmap decisions than the hour I spend in a strategy meeting. It's the highest-leverage hour of my day. The fact that it looks like a $50/hour task on a job description doesn't change what it actually produces.

When I genuinely can't keep doing it — when the volume crosses a line I can't physically meet — I'll stop. We're nowhere near that line.

If you're a founder doing this

Three rules I've landed on after years of it:

• Pick a window and stick to it. Mine is 9-9:30. Outside that window I trust the team. Otherwise support eats the whole day.
• Never close a ticket without writing a one-line internal note about what it taught you. The pattern only shows up when you can search those notes.
• Reply as yourself, not as "the team." Customers behave differently when they know they're talking to the founder. They tell you more. That's the whole point.

Talk next week,
— George