When Faire launched in 2017, its engineering team comprised just five people. Building a high-velocity team required attracting senior talent from the outset. Founding co-founder Marcelo Cortes spent over a year conducting prospective hiring meetings over coffee, sharing ongoing updates on milestones and releases to recruit targeted engineers. The early team drew on past experience building Cash App at Square. Applying those established patterns allowed the five engineers to construct a complex payment infrastructure—handling debit and credit card processing, partial refunds, asynchronous retries, voided transactions, and ACH payouts—in a couple of weeks without architectural missteps.
Trade Show Iteration and Early Product-Market Fit
During its initial phase under the original name Indigo Fair, the company relied on live retail events to validate its product. During three-to-four-day trade shows, executive leadership gathered immediate merchant feedback on-site while engineers deployed system changes nonstop. The team executed rapid live deployments to evaluate distinct customer incentives, including free shipping on first orders, buy-now-pay-later terms, free returns, and promotional credit on re-orders from the same brand. By continuously deploying software updates across these multi-day trade shows to measure customer response, Faire achieved strong product-market fit within six months of launching.
Establishing Foundation: Java, Kotlin, and Data at 10 Customers
Rather than postponing infrastructure decisions until scaling required major rewrites, Faire established architectural guidelines and data tracking at the start. The team began collecting user data and building data pipelines when the platform served only 10 customers. The core architecture replicated relational database events from MySQL to a Redshift data warehouse within minutes, making operational data accessible for developer queries. For backend services, engineers initially built using Java and MySQL, establishing code style guides, mandatory tests, and code reviews from day one. In 2019, the team transitioned its primary backend language to Kotlin, aligning code reviews with JetBrains conventions.
Scaling Through Metrics: The 1.6x Capacity Turnaround
Between 2017 and 2020, Faire’s engineering department expanded from five to over 100 engineers. Early operational metrics included tracking site uptime and continuous integration wait times to resolve developer bottlenecks. However, during the Faire Summer Market online trade show, surging site activity pushed the database capacity to its operational limits. To resolve database load without stalling feature development, engineering leadership created a metric tracking transaction run duration for every application database interaction. By mapping database transaction duration directly to individual engineering teams, leadership assigned each team a target to cut database usage by 20% over three months.
Six months after the summer performance issue, Faire ran its Winter Market online trade show. Armed with the database optimizations, the platform successfully processed 1.6x higher traffic than the summer event without reaching database capacity limits. By establishing data collection with just 10 customers, mandating code reviews from the first hire, and holding independent teams accountable to database usage metrics, Faire preserved engineering execution speed throughout a twentyfold organization expansion.
AI-generated editorial illustration; not a documentary photograph.
Source: How to maintain engineering velocity as you scale
SIGNAL TITANS
Leave a Reply