Fyxer's growth engineering team conducted 360 experiments in a year, scaling ARR from $1M to $35M. AI helped automate the experimentation loop to achieve this dramatic growth.

Its explosive growth was driven by a company-wide culture of experimentation, resulting in 541 A/B tests run within 12 months (averaging more than two per working day).

The success relied on two pillars: the right mindset and an AI-first execution loop.

1. The Growth Engineering Mindset

  • Learning Over Intuition: The team treats every product change as a hypothesis to validate rather than just a feature to ship. Because Fyxer’s win rate was 25% (meaning 75% of ideas failed), A/B testing acted as a critical shield to prevent harmful changes from rolling out to all users.

  • Testing Early and Boldly: Kameron rejects the idea that startups are "too small" to A/B test. While small companies can't easily detect minor 5% lifts, they can and should test for massive 20% to 30% shifts in core areas like pricing models, usage limits, and onboarding.

  • Radical Transparency: To embed this culture, the team created a public Slack channel where every single win, loss, and experiment result was shared openly with the entire company.

2. The AI-First Tech Stack

To run hundreds of tests without scaling headcount, the team used AI to compress the experimentation lifecycle from weeks to hours:

  • Development & Testing: They used Cursor (specifically its desktop mode with virtual environments) to build and visually review changes without developers needing to pull down code locally. For simpler backend or feature flag changes, they relied on tools like Claude Opus, Codex, and Tembo to complete the implementation in one go.

  • Data Analysis: They utilized Dot, an AI data analyst connected to their BigQuery warehouse. This allowed non-technical team members to ask complex data questions directly in Slack and receive answers in seconds.

  • Management: The entire experimentation program was managed using GrowthBook, an open-source feature flagging and A/B testing platform.

That leads to several highly effective, counterintuitive breakthroughs:

  • Free-to-Paid Conversion: Boosted conversion from 5% to 35% by adding an optional upfront credit card gate before the free trial, drawing design inspiration from Canva's clear checkout timeline.

  • Annual Plan Optimization: Shifted 50% of subscribers to annual plans (a 2.3x increase) simply by defaulting to yearly billing, increasing the discount to 25%, and displaying the effective monthly price.

  • Trial Conversions: Increased the trial start rate for personal email users by 65% by segmenting trial lengths based on signup type.

Fyxer is expanding its growth engineering team from 6 to 13 people with the goal of running 1,000 experiments this year, targeting an ambitious ARR goal of $100M to $150M.