Starting with the Problem
We did not start with a product idea. We started with a problem we experienced firsthand: managing marketing for multiple clients was overwhelming, and every existing tool was either too generic or required too much manual input. The idea for Roxy Agent emerged from that frustration.
Key Lessons Learned
Ship Early, Iterate Fast
Our first version was embarrassingly basic. Three agents instead of seven, no integrations, and a UI that only a developer could love. But it worked. And more importantly, it let us get real user feedback within weeks instead of months.
Listen to What Users Do, Not Just What They Say
Users told us they wanted more customization options. But our analytics showed they barely used the customization features we already had. What they actually needed was better default outputs. We invested in improving our AI models instead of building a settings panel nobody would use.
Pricing Is a Feature
We changed our pricing model three times before finding what worked. The winning formula was simple: charge per client managed, include all features, and offer a generous trial. Agencies understood this model instantly because it mapped to how they already think about their business.
What Is Next
We are building in public wherever we can. Each major update gets a detailed changelog, and our roadmap is shaped by the agencies using the platform every day. The product is never finished. It is always getting better.
Alongside Roxy Agent, our team builds custom e-commerce solutions for clients on Shopify, WordPress, and BigCommerce. The lessons from building a SaaS product (ship early, measure everything, iterate based on data) are the same principles we apply to every store we develop.