Let's Ferry.

Let's Ferry is one of the largest online platforms for ferry ticket bookings, combining reliable service, speed, and modern functionality. It serves the biggest ports in the Mediterranean, with expansion to international markets in progress — with the stated goal of becoming one of the leading global players in the category. Today it operates as one of the most important business units within the Gelasakis Group of Companies, taking bookings from 147 countries.

Since the start of the partnership, Let's Ferry revenue has grown 1,036%. By 2019 the platform had captured a significant share of the online ferry-booking market. The challenge at that point shifted: with strong market position and volume in place, the next priority became maximizing profit while retaining the majority of revenue — moving from pure scale optimization to efficiency optimization without sacrificing the top line the team had worked years to build.

That pivot is harder than it sounds. Scaling revenue is an additive problem: more queries, more audiences, more creative. Maximizing profit at the top of the market is a subtractive problem: find every euro of spend that is not contributing to profitable margin and remove it, without simultaneously killing the broader-funnel activity that sustains the volume. The optimization loop has to run faster than human analysts can move — the traffic patterns, CPC dynamics, and conversion economics in the travel category change hourly, especially during seasonal peaks.

We split the solution into two complementary pillars. The first pillar is real-time account analysis. A data layer pulls Google Ads performance, Meta ad data, and first-party booking data into a warehouse continuously — not on a daily refresh but close to real time. On top of the warehouse, an analysis engine identifies, at any given moment, the single intervention most likely to deliver the largest profitability gain. That might be pausing a poorly converting query cluster, adjusting a bid strategy in a specific audience, or shifting budget across route segments.

The second pillar is automation of those optimizations. Many of the interventions the analysis layer recommends are time-consuming to execute manually — not cognitively hard, just repetitive and urgent. Those got automated. Scripts and workflows apply the recommended actions 24/7, with guardrails for risk-sensitive changes (anything that moves significant budget still requires explicit approval). The combination means the account is always optimizing, including overnight and during seasonal peaks when manual reaction times would lag the market.

The two pillars reinforce each other. Real-time analysis surfaces what matters faster than weekly reviews can. Automations execute on what matters faster than an account manager can click through the UI. The net effect is an account that converges toward maximum profit continuously — not in bursts after each planning session, but every hour of every day, including the ones that historically went unmanaged.