Picture this: your engineering team—the people building your actual product—are spending their days on customer support calls. Your most expensive resources are drowning in a weekly meeting that looks like a fire drill, with everyone shouting requests across a room. We'd gone from 10 to 100 employees, and our communication had turned into a spider web of chaos where nobody had context for what they were actually working on.
Quick Summary
I built a centralized ticketing system that transformed Lyft's internal request management from a chaotic web of one-to-one relationships into a structured, transparent operation that freed up multiple engineering FTEs while actually improving service speed.
The Problem
When you map out how 50+ sales reps were communicating with engineering, analytics, and product teams, it looked like one of those insane LinkedIn network graphs—every person connected to every other person with no rhyme or reason. Sales reps were pinging whoever they knew personally. Engineers were ending up on customer calls. The weekly escalation meeting was absolute chaos—a laundry list of "I need this report" and "when will this bug be fixed" with zero prioritization or context.
The real issue wasn't just the noise—it was that our most valuable people were doing the wrong work. Engineers were handling customer support tickets that our customer success team wasn't technical enough to address. Everyone was working, but nobody knew if it was high-value work.
My Approach
I stepped back and saw this as a routing and triage problem, not a communication problem. The system needed a central nervous system—one place where all requests flowed through so they could be properly categorized, prioritized, and routed to the right people with the right context.
I knew this would be politically difficult. Nobody likes having their direct access cut off. But I also knew that if I could create an "RSS feed of the business," I could turn chaos into signal.
What I Built
I forced every request from sales and customer success to flow through a single ticketing system (accessible via Slack or email). My analytics team became the central routing hub—we'd triage every request that came in.
The magic was in the layered response system:
- Immediate answers: If we'd already done it, instant response
- Quick tasks: Simple requests got assigned to analysts
- Engineering requests: We'd do the legwork first, then package up structured, digestible requests for engineers with full context
- Not-gonna-happens: We'd flat-out decline and explain why
I had to enforce this religiously. When team members got direct requests, they'd forward them to me and I'd reply back. Nobody was above the system—including me.
The Impact
The transformation was dramatic. That weekly chaos meeting became quiet—fewer people, higher-level strategic discussions about roadmap priorities instead of fire drills. I'll never forget walking into one of those meetings months later and getting what amounted to a standing ovation from the engineering team.
We shielded product and engineering from 80% of requests immediately. Sales reps and customer success got faster responses because we could instantly serve up work we'd already done. I could give sales managers heat maps of their team's request patterns—real data instead of just complaints from one-on-ones.
The Takeaway
Sometimes the most valuable thing you can build isn't code or dashboards—it's a system that puts the right work in front of the right people at the right time. The best optimization isn't always making things faster; sometimes it's making sure you're doing the right things in the first place.