Over the course of the last 8+ years, the team at Outseta has responded to 21,623 support tickets—not that anyone is counting! And of those, I’ve personally replied to 10,000+.

As with my post on Product Management Lessons From The Trenches earlier this year, I’m not a Customer Service or Customer Success professional by training any more so than I am a Product Manager. As a Co-founder on a small bootstrapped team, I’ve absorbed these responsibilities as part of the job.

At Outseta, everybody on our team contributes to support without exception—so our entire team is on the frontlines responding to customer service inquiries daily. This post shares what I’ve learned about support over the last 8+ years of helping our customers.

Take a stance in terms of how you’ll approach support

I’ll start with this—I think it’s important that companies take some sort of stance when it comes to their approach to support. “We’re not going to offer any customer service” is every bit as valid as “We’re going to offer the best customer service in our industry.” I think the important thing is that you choose an approach that fits your business and stick with it.

The vast majority of companies seem hellbent on wallowing somewhere in the middle. They offer support, but it’s not a top priority and it ends up leading to frustrating experiences for customers more often than not. This is how you build a brand that people hate.

Don’t wallow.

Customer service directly generates revenue

Too many companies look at customer service teams as a cost center—and they just can’t rationalize investing in customer service when there’s product and marketing and sales functions competing for budget. This is how you end up wallowing in the mediocre support wasteland.

Customer service directly generates revenue—I know this, because I’ve seen it prove itself out over-and-over again. Customers have told me that they would have churned had I not helped them through onboarding. Customers have signed up and told me it’s because they were referred by someone else who had a great experience with our team.

This point should not be controversial.

The more important question to ask is “Will investments in sales or marketing generate more revenue than investments in customer service.”

And that may very well be true! Investments in sales and marketing may be much more scaleable, while customer service often feels like a slog where you’re only able to help one customer at a time.

This is often very true but I think what’s needed most is a mindset shift. If you’ve decided to invest in customer service, you need to constantly be looking for ways to help customers self-serve whenever possible. You need to spend your time figuring out how you can scale your support processes just as you spend time trying to figure out how to scale your marketing channels.

From helping customers succeed with your product to creating positive brand experiences that drive word-of-mouth, great customer service is a massively under-appreciated growth lever.

Don’t troubleshoot technical issues live

If you have any sort of technical product, customers will inevitably ask you to jump on calls with them to troubleshoot technical issues live. The answer, overwhelmingly, should be “no.”

It’s not about being unwilling to help. Technical issues simply require space in which you can dig in—to figure out how to replicate the issue, to consider root causes, and to apply whatever fix is necessary. Troubleshooting technical issues is a form of sleuthing; and it’s much harder to do when someone else is watching on and jabbering in your ear as you try to think. It’s a waste of time for both parties. 

Troubleshooting the issue independently and then sending the customer a video outlining what the issue was and how it was resolved is almost always preferable. This gives you the opportunity to both solve the issue more quickly, but also delight the customer by showing them what the issue was and how it was solved.

Customers absolutely love personalized video responses—I used to use Wistia Soapbox for these and now use Tella.

Detailed tickets = Detailed responses

Customers often fail to recognize how much impact they have on the quality of the customer service they receive. I can’t tell you how many generic “Outseta isn’t working” tickets we’ve received over the years—which immediately kicks off a back and forth interaction that’s nothing more than information gathering on the issue the customer is actually experiencing.

On the flip side you have customers who will provide a very detailed breakdown of the issue at hand, potentially including screenshots, videos, and steps to reproduce the problem. 

In any sort of tech support related context, the person tasked with solving the issue almost always first has to replicate the issue to diagnose what’s causing it—so incoming tickets with this level of detail start out 5 steps ahead. 

It’s a “help me, help you” scenario— if taken to heart customers will be surprised how many “one response resolutions” to problems they receive.

The battle between mowing down tickets and providing the best answer possible

In terms of the day-to-day work of responding to customer service questions, even 8+ years into this I still find myself fighting the inner battle between working through support tickets quickly versus providing the best answer possible to any given question.

Look, for most people (including me) responding to customer service queries probably isn’t their favorite part of the job. And when tickets pile up, it’s natural to treat tickets like a to-do list and try to work through them as quickly as possible.

But without fail, the extent to which you take shortcuts will come back to bite you. If your responses are unclear or don’t walk the user to a resolution as clearly as possible, the odds of them coming back with subsequent questions go up dramatically.

Yes, you need to be efficient and it’s tough to treat every support ticket like it’s your own personal thesis on the problem at hand. But to the extent that you can rewire your brain to “How can I provide the best possible answer?” you’ll save yourself time—even if the initial response takes you 50% longer to draft.

More scale = More problems

I see many Founders become frustrated with the amount of time that they are spending on support—and only then do they make significant investments in their ability to scale support expecting that “the problem” will just go away. That might even feel like it’s true if you genuinely make some progress—and your business remains relatively stagnant in terms of growth.

In the context of a growing business, it’s going to feel like the number of issues is growing—even as you improve in your ability to resolve customer issues. It’s just math, so you’re best off accepting this.

For example, let’s say for every 1,000 customers that you have 1 customer is experiencing a truly red-alert, critical, meltdown level issue at any given time. If you’re at 100, or 200, or 500 customers you’ll have plenty of days without any such emergencies.

But if you scale to 10,000+ customers, you’re going to have 10+ customers having such issues concurrently. This will happen—it’s best you prepare for it.

URGENT = Almost certainly not urgent

This one is more just a funny observation, but it’s uncanny how often it’s true.

Nearly all incoming support tickets that say “URGENT” in the subject line are in fact, not that urgent. This is not to discount any customer’s feelings—it’s just a very clear pattern that emerges when looking at the actual issues at hand when tickets come in with this word in the subject line.

This is a much more direct indication that the customer is distressed than the gravity of the issue at hand. 

There’s such a thing as too responsive

An article by Arvid Kahl changed my mind on this one—in the article, Arvid speaks to the early days of his SaaS product FeedbackPanda, when he prided himself on responding to incoming support tickets or chat conversations often within minutes. 

Arvid’s point is that while this feels good—and is in fact an opportunity to “wow” customers—this level or responsiveness can actually come to work against the best interests of customers. When you become this dedicated to dropping everything to solve each individual issue that presents itself in your support queue, you’ll eventually find yourself fully devoted to this queue at the expense of time when you could be working on product improvements that benefit the entirety of your customer base.

The point is not that being quick to respond is a bad thing—of course it’s a good thing, and customers appreciate it. But you can go overboard, too. 

Like Arvid, I’ve often been guilty of trying to be too responsive. And as I’ve tried to combat that and make more time for deep work that affects the customer base at large, I’ve found an interesting second-order effect. By being even a little bit less responsive, I’ve seen:

• Customers are more likely to solve their own issues
• Fewer support tickets are submitted

I can’t tell you how many tickets we see, where if we don’t respond in even 1-2 hours the customer pops back up in our inbox saying “You can close this ticket, I figured it out.”

Beyond that, there’s absolutely a persona of customers who submits a request for help the moment they have any sort of question. Responding in minutes encourages this behavior. By being even slightly less responsive, these customers will often try to self-serve or at least try before they reach out.

Regardless of whether your response time is one minute or one day, acknowledging that a submitted ticket has been received and setting expectations in terms of when the issue should be addressed is a win-win.

You can’t make every crisis your own

This one probably sounds not very customer friendly—after all, customer service is about solving your customers’ problems, right? But hear me out—this one is about maintaining your sanity.

If you work in customer service long enough, you’ll experience customers that try to make their own crisis your problem. In the context of Outseta specifically, this is often a support ticket that comes through on a Sunday morning saying something like:

“Hey, I signed up for Outseta and have to launch a very complex SaaS product for my client by tomorrow morning. I haven’t gotten started yet but the deadline is tomorrow. I need urgent help—where are you?”

You simply need to recognize when a customer is putting their own self-inflicted problems on you. This scenario is very, very different from one where you or your product are directly responsible for causing the issue that the customer is experiencing. 

If you caused the problem, the crisis is yours to assume. If the customer’s lack of responsibility is to blame—you can’t make every crisis your own.

People want to talk to people

To the extent that you can make customer service self-service, you should. This can be in the form of great documentation, tutorial videos, FAQ’s, or using AI tools to answer customer questions. Do whatever you can.

But when these methods break down and customers need real help—when they need someone to truly listen to them and help them get unstuck—people overwhelmingly want to talk to people.

Great customer service delivered by a human is appreciated by everyone—and it’s increasingly a huge differentiator.

Fewer support channels is good

One of the flags we planted in the ground early on at Outseta that I think we got very right is we chose to focus on delivering support via one channel—support tickets.

If our customers want to self-serve we invest a lot of time and effort into our documentation—if they still need more personal help, support tickets is how we serve them. This streamlines our business a lot and helps us serve our customers better.

I think it’s very common that companies feel the need to be everywhere—responding to issues on social media, via live chat tools, tickets, their own online community, and building automated AI tools too. This approach is great in theory—be wherever the customers are—and I think generally makes sense for larger or heavily funded companies. 

But in the context of running a small company, simplicity wins. If you spread yourself across too many channels, it soon becomes a logistical nightmare and you’re spread too thin everywhere.

AI’s evolving role in customer service

I can’t write this article without at least touching on AI’s growing role in customer service. I’ll start by saying there are lots of companies whose customer service tends to be very basic stuff—“Please tell me how to cancel my account” or “How can I update my email address.” But then there are companies whose support volume tends to include far more wide-ranging, specific, and technical questions.

If you fall into the first bucket, great—AI centric customer service tools may very well represent all that you need. If you fall into the second bucket, I am highly skeptical that AI tools are truly going to help you solve customer problems well. You can certainly still employ them to deflect as many of the relatively simple questions as you possibly can, but don’t go overboard with this approach. When people want to talk to a person, let them talk to a person. 

I’m firmly on the “have AI tools quickly escalate issues they can’t easily solve to a human” team.

This comes from my own experience as a customer. Back in 2018, I wrote about automated “chat bots” being the #1 source of the lousy customer service interactions that I had. In 2025, that has absolutely shifted to AI tools. It’s absolutely maddening when you can’t get a response that’s accurate or a person to actually speak to you. Nothing sours my opinion of a company more than an endless loop AI-generated responses and the inability to reach a human.

I’ve deliberately spoken to a lot of SaaS companies who have made big claims about the impact that AI tools have had on their customer service teams. “AI is now responding to 60% of our customer issues!” is a common response. But then when I test their support myself, the experience is terrible—so what gives?

These companies aren’t wrong that AI *is* responding to more of their support queries than ever before. And yes, they are all tracking how often AI is actually resolving issues. But there seems to be a very muddy middle that’s all but unaccounted for—the cycles of customer frustration where users are wrestling with AI tools and having a nightmarish experience.

SaaS companies are better at tracking nice clean metrics like “resolutions” than they are things like customers abandoning a customer service interaction altogether or the amount of time customers spend feeling like they want to throw their computer out the window.

At the core of this tension is a conflict in terms of who these tools actually serve—the company, or the customer. From the perspective of the company, AI tools deflect ticket volume and allow them to operate more efficiently, with less staff—all positive outcomes. But assessed through the eyes of a customer, time wasted and frustration resulting from AI service loops comes at a serious expense. Yes, most people would prefer to self-serve but when they need help they are typically looking for something much closer to “service” than a robot with wrong answers and endless endurance.

‍

That's a wrap. I'd like to add a personal "thank you" to all of the customers that I chat with every day in our support queue. I so appreciate all of you who choose to be pleasant, who document their issues as clearly as possible, and who are patient with me and our team as we work to solve you problems. It's a pleasure to serve you.