K
Kixie Kixie

Product Development
Fundamentals

Building teams through empathy and clear principles. From 1-hour standups to high-velocity squads.

Engineering Organization Transformed

When I joined Kixie, the engineering team was struggling with coordination. Twenty-three people would gather on Zoom every morning, cameras mostly off, for standups that stretched past an hour. Everyone was busy. Nothing was shipping.

These were talented engineers. They'd built a platform handling 100 million minutes of calls. But somewhere between a major outage and the risk-averse processes that followed, they needed better structure. They needed product leadership to help them focus.

Squad-Based Organization

Telephony Squad

System reliability and call features

Call quality & uptime
Power dialer features
Local presence dialing

CRM Squad

Integration stability and partnerships

Salesforce integration
Data synchronization
Partner API stability

Dashboard Squad

Reporting and account management

Sales analytics
User authentication
Account permissions

The Problem

Day 1: The Reality Check

My first standup at Kixie lasted 63 minutes. Twenty-three people on Zoom, all in the same office. Each person recited yesterday's work and today's plan. Nobody asked questions. Nobody offered help. When someone mentioned a blocker, the response was silence.

After the meeting, I checked Jira. The oldest ticket in the active sprint was 127 days old. It had been moved through six sprints. The description read "investigate performance issue." No acceptance criteria. No updates. Just a trail of sprint-to-sprint migrations.

The Symptoms

People arrived late to standups because they knew nothing meaningful would happen

Engineers worked on whatever seemed urgent, with no shared understanding of priorities

Customer issues bounced between teams because nobody owned specific domains

Product requirements lived in emails, Slack threads, and half-remembered conversations

Releases were unpredictable, with no consistent cadence or quality bar

Teams measured success by activity metrics rather than customer outcomes

The Production Incident

Three months before I arrived, a production incident highlighted critical gaps in our deployment process. The recovery took considerable time and impacted service availability. Leadership made it clear that reliability was now the top priority.

The team responded by extending QA cycles and batching releases. Deployments moved to weekends with the entire team on call. When I suggested more frequent, smaller releases, there was skepticism. One senior developer explained their perspective: "This is how we ensure stability."

The risk-averse processes had created deployment bottlenecks. Large releases actually increased risk. Extended cycles meant issues stayed in production longer. Weekend deployments led to fatigue. We needed to transform how we thought about releases.

The Root Cause

The engineers were talented. They cared about the product. They wanted to ship quality code. The system failed them. Without clear ownership, defined processes, or shared agreements, every day felt like swimming upstream. They needed structure that enabled autonomy, boundaries that created freedom, and most importantly, leadership that understood their challenges.

What We Built

Three focused squads. Clear ownership. Shared agreements. Real accountability.

1. Verticalized Squad Structure

We chose a streamlined approach: three autonomous squads, each owning a complete vertical slice of the product. This structure prioritized clear ownership and fast decision-making over complex matrix organizations. We eliminated inter-team dependencies to enable each squad to deliver value independently.

Telephony: Everything voice-related. Call quality, reliability, features.
CRM: All integrations. Salesforce, HubSpot, data sync.
Dashboard: Analytics, reporting, user management.

Product Strategy

Each squad had their own product roadmap, sprint cycles, and success metrics. This autonomy enabled faster feature delivery and clearer accountability. Product managers could make decisions quickly without cross-team coordination overhead.

The roadmap includes evolution. By 2026, we'll introduce cross-functional AI initiatives spanning all verticals. Voice AI will leverage CRM data. Predictive analytics will integrate telephony insights. But first, we established reliable delivery foundations.

When a customer reported a call quality issue, Telephony owned it completely. When Salesforce sync broke, CRM handled it alone. No confusion. No handoffs. No blocked tickets waiting for another team. Just clear ownership and accountability. The sophistication would come later. First, we needed to ship.

2. Staffing the Squads & Coaching

Squads without product leadership would fail. We needed three product managers, one for each squad. I hired two externally and made a bet on internal talent for the third.

The CRM Squad Story

For the CRM squad, I took a calculated risk. I promoted someone from our sales engineering team who had deep CRM expertise. He'd spent years working directly with customers on complex integrations, understood their pain points intimately, and had built relationships with our partner companies. Most importantly, he asked strategic questions: "Why are we building this?" not just "How?"

Every morning, 7:45 AM, we'd meet for coffee. Fifteen minutes. Not status updates, but product philosophy discussions. "A customer called support about this," I'd say. "What would you do?" Initially, his answers focused on implementation. Gradually, they became strategic. "We could fix the symptom," he'd say, "or we could redesign the flow to address the root cause."

Coaching Philosophy

Daily touchpoints: 15-minute morning check-ins to discuss decisions, not tasks
Learning by doing: Started with small product decisions, gradually increased scope
Safe failures: Created space to make reversible mistakes and learn from them
Customer exposure: Joined support calls to hear real user problems firsthand

Three months in, he ran his first sprint planning alone. I sat in the back, silent. He asked every question I would have asked. Six months in, the CRM squad had the highest velocity and customer satisfaction scores. The sales engineer who knew CRM setups inside and out had blossomed into a strategic product leader, presenting roadmaps to the board. The CEO pulled me aside afterward: "Where did you find him?" I smiled. "He was always here. We just gave him the right role."

3. Working Agreements

We created agreements that everyone understood and accepted. Written by the team, for the team.

📢 Communication

  • • 2-hour response SLA for team messages
  • • Default to public channels over DMs
  • • Emoji reactions for acknowledgment
  • • Update Slack status when unavailable

📅 Meetings

  • • Agenda required 24 hours in advance
  • • Start and end on time, always
  • • Decline if you can't contribute
  • • 15-minute standups maximum

🎯 Ownership

  • • If you commit, you deliver
  • • Speak up early if blocked
  • • No ticket, no work
  • • Update tickets daily

🤝 Respect

  • • Assume positive intent
  • • Let people finish speaking
  • • Feedback about work, not person
  • • Resolve conflicts directly

These agreements applied to everyone. Engineers, product managers, leadership. No exceptions. When someone violated an agreement, we addressed it immediately with empathy and clarity.

4. Sprint Process & Definition of Done

We established clear processes that made sense for our context. Every story had to meet specific criteria before entering a sprint.

Story Requirements

Clear business context explaining why this matters
Testable acceptance criteria
Dependencies and assumptions documented
Product owner approval before sprint planning

Definition of Done

Acceptance criteria met and validated
Product owner approval received
Customer-facing documentation updated
Staged for release with feature flags configured
Success metrics and tracking in place

5. Demo Culture & Feedback Loops

Every Friday at 3pm, we gathered for demos. Not status updates. Live demonstrations of working software.

Engineers showed what they built. Product gave feedback. Support shared customer reactions. Sales highlighted market needs. Everyone learned what others were working on and why it mattered.

These sessions built empathy across teams. Engineers understood customer pain. Product saw technical constraints. Support learned about upcoming features. The whole organization aligned around what we were building and why.

6. Modern Release Strategy with Feature Flags

The production incident had made the team risk-averse about deployments. But larger, less frequent releases actually increased risk. We needed to decouple deployment from release and make rollbacks instant.

Modern Release Process

We implemented a sophisticated release strategy that separated deployment from release. Code ships continuously when ready. Customer-facing features release bi-weekly with a controlled 5-day ramp. This approach dramatically reduced risk while increasing velocity.

Continuous deployment: Code pushed to production when ready
Bi-weekly releases: Customer features released every two weeks
5-day ramp: Gradual rollout from 1% to 100% of users
Instant rollback: One-click reversion in 30 seconds

The Turning Point

Week three with Harness. The new product manager on the CRM squad pushed his first major integration update. It broke for our largest enterprise customer. In the old world, this would have triggered panic calls, emergency meetings, and probably that promised "head rolling."

Instead: Click. Rolled back. 30 seconds. Customer never noticed. We fixed the edge case, deployed to 10 test accounts, verified the fix, then gradually rolled out to everyone. The new PM learned. The customer stayed happy. The CEO never knew.

Six months later, that skeptical senior developer came to me. "I was wrong," he said. "This is how we should ship code." Teams now deployed critical fixes during normal business hours with confidence.

The rip cord didn't just change our deployment process. It changed our relationship with failure. Mistakes became learning opportunities instead of career-ending events. Teams started taking calculated risks because the downside was limited to 30 seconds of rollback time.

My Role

Product Function

Established product management function at Kixie from the ground up

Squad Structure

Designed and implemented verticalized squad organization

Agile Transformation

Moved teams from ad-hoc to structured sprint methodology

Talent Development

Hired 3 PMs and coached an engineer's transition to product management for CRM squad

Cross-Team Alignment

Built bridges between engineering, support, and sales teams

Quality Standards

Defined sprint processes and quality standards for all teams

Working Agreements

Facilitated creation and adoption of team working agreements

Results

Measurable improvements in velocity, quality, and team satisfaction

How We Measured

Every month, we sent a five-question anonymous survey through Officevibe. Simple questions: "Do you understand your team's priorities?" (clarity), "Do you feel your work matters?" (purpose), "Would you recommend Kixie as a place to work?" (advocacy), "Do you have the tools to succeed?" (enablement), and "Do you trust your team?" (psychological safety). Each question scored 1-5. The average became our engagement score, providing clear signals about team health.

📊 Key Metrics Transformation

Before

Standup Duration 60+ min
Average Ticket Age 90+ days
Sprint Completion ~40%
Team Engagement 2.8/5

After

Standup Duration 15 min
Average Ticket Age 14 days
Sprint Completion 92%
Team Engagement 4.2/5
Continuous
Code Deployment
Push when ready, bi-weekly customer releases
30%
Reduction in Bugs
Customer-reported issues
3x
Feature Velocity
Stories completed per sprint
30 sec
Rollback Time
Previously hours of hotfixes
95%
Deployment Confidence
Teams shipping without fear
0
Weekend Emergencies
Since implementing feature flags

Management Philosophy

Empathy First

Every problem started with understanding. Why were people late to standups? Because they felt pointless. Why were tickets aging? Because nobody knew who owned them. Why was morale low? Because talented people felt trapped in dysfunction.

Clear Boundaries Create Freedom

When the Telephony squad knew exactly what they owned, they stopped waiting for permission. When working agreements defined response times, people stopped wondering if they were being ignored. Clarity enabled autonomy.

Process Serves People

We implemented Agile principles, but adapted them to our reality. Standups were 15 minutes because that's what worked. Sprints were two weeks because that matched our release cycle. The best process is the one your team actually follows.

Measure What Matters

We stopped counting lines of code and started measuring customer outcomes. Sprint velocity meant nothing if we weren't solving real problems. Team satisfaction scores mattered because engaged teams build better products.

Invest in People

The sales engineer I promoted to product management proved something important: talent exists everywhere in your organization. His customer expertise and CRM knowledge, combined with daily coaching sessions, transformed into strategic product leadership. Six months later, he was presenting roadmaps to the board. Investment in people compounds faster than any technology investment.

Frequently Asked Questions

How long did the transformation take?

Two weeks to restructure into squads. Three weeks to implement our modern release process. Three months for teams to fully adopt the new practices. Six months to achieve the 92% sprint completion rate. The structural changes happened quickly, but cultural transformation took time. Once teams experienced the benefits of smaller, safer releases, adoption accelerated.

What was the biggest challenge?

Building trust after the production incident. The team had become risk-averse and process-heavy. The turning point came with our first successful feature flag rollback - 30 seconds to revert, no customer impact. This proved that we could ship safely and frequently. As confidence grew, teams began taking ownership and innovating again.

How did you handle resistance to change?

We involved everyone in creating the solution. Engineers wrote the working agreements. Teams chose their own squad structures. People support what they help create. When someone resisted, we asked what would work better. Usually, they had valid concerns that improved our approach.

What metrics mattered most?

Sprint predictability told us if teams understood their capacity. Ticket age showed if we were solving the right problems. Team engagement surveys revealed if people felt heard and valued. Customer-reported bugs measured our quality. Every metric connected to either team health or customer success.

Did you use any specific frameworks?

We borrowed from Scrum, Kanban, and Shape Up, but created something uniquely ours. Spotify's squad model inspired our structure, but we adapted it for our size. Amazon's working backwards approach influenced our PRDs. The framework that works is the one that fits your context.

Why such simple squad boundaries?

The team needed clear ownership and fast decision-making. We eliminated inter-team dependencies so each squad could deliver value independently. No waiting, no blocked tickets, no coordination overhead. This structure was intentional. Once we established reliable delivery, we could evolve. By 2026, we'll introduce cross-functional AI initiatives requiring squad collaboration. Voice AI will leverage CRM data. Predictive analytics will span all verticals. But first, we needed to establish strong foundations for independent delivery.

Let's Build Teams That Ship

Transform your engineering organization with empathy, clear principles, and measurable results.