Every executive I've met believes you have to choose. Either you invest in customer experience — more agents, faster response times, better tooling — or you cut costs. The assumption is baked into how most organizations budget CX: it's a cost center, and cost centers get squeezed.
At Eventbrite, we proved that assumption wrong. In the same fiscal year, we reduced operating costs by 66% and improved Net Promoter Score by 10 points. Not sequentially. Simultaneously.
Here's what most leaders get wrong about that equation.
The false trade-off comes from measuring the wrong things
Most organizations measure CX efficiency by headcount and response time. They measure satisfaction by CSAT scores. These two sets of numbers almost never talk to each other — and that's the problem.
When you optimize for speed without understanding why customers are reaching out in the first place, you get faster answers to the wrong questions. You reduce handle time while ticket volume stays flat or grows. You hire less people, but your remaining agents burn out covering the same broken processes.
The real question is never "how do we answer faster?" It's "why are customers writing to us at all?"
Systems before symptoms
Every contact is data. A customer emailing about a failed ticket transfer is telling you something about your product flow. A cluster of contacts about refunds during a specific event type is telling you something about expectation misalignment upstream. A spike in repeat contacts from the same customer cohort is telling you your resolution process isn't actually resolving anything.
At Eventbrite, the first thing I did was build a taxonomy of contact reasons — not the vague categories that were already in the system, but the real ones. Why, specifically, is this person reaching out? What broke, and where in the journey did it break?
That taxonomy became our product roadmap. Every reduction in contact volume was a feature shipped, a flow fixed, a communication improved. We weren't answering faster — we were eliminating the need for customers to contact us in the first place.
What the 66% actually looked like
The cost reduction came from three places:
1. Contact deflection through product fixes. Once we could show the product team that a specific user flow was generating 12% of our inbound volume, we had a business case that wasn't about CX at all — it was about engineering efficiency. Fixing the flow cost far less than staffing to handle the contacts it generated.
2. Automation applied correctly. We didn't deploy AI chatbots to answer everything. We identified the narrow slice of contacts — truly simple, high-volume, zero-judgment-required — where automation would reduce friction rather than add it. For everything else, we made sure a human picked it up fast.
3. Right-sizing the team to the real volume. After deflection and automation, the contact volume that remained was genuinely complex. A smaller, better-trained team handled it with higher quality and higher engagement scores than our larger team had before.
Why NPS went up while costs went down
When you fix the upstream problems that generate contacts, the customers who do reach you are genuinely stuck — and your team can focus entirely on solving hard problems instead of processing routine ones. Agents who handle interesting, complex cases stay engaged. Engaged agents give better service. Better service drives loyalty.
The math isn't magic. It's just the result of asking the right question first: what is this contact telling us about our product?
If you're looking at your CX budget and wondering how to do more with less, start there. The answer is almost never hiring fewer people. It's understanding why people are reaching out — and fixing that.