Bookings Won Back From OTAs: New Funnel Experimentations That Grew Bookings by 18%

Part of the Luxury Boat Tours series · one client, three lenses

You're reading the Funnel & CRO lens. Luxury Boat Tours was a single engagement we tell three ways — same client, three disciplines, plus the overview. Read the rest in any order: Overview, Performance Marketing, and Data & Analytics.

We worked with Luxury Boat Tours— a global, award-winning marine-tour operator running yacht cruises and guided snorkeling day-trips, with a category-leading product: a 4.9-star rating across 10,000+ reviews. The experience sold itself — but the growth engine didn't. Demand leaned heavily on the booking aggregators (the OTAs), an unhealthy dependence that quietly rented out margin the operator should have owned, while paid spend, especially on social, had burned budget with nothing to show for it and no one could read what any of it was doing.

Meet the operator.

Luxury Boat Tours runs premium island day-trips — yacht cruises, manta snorkeling, cliff-view excursions, all-inclusive — with a category-leading 4.9-star rating across 10,000+ reviews, a wall of global awards, and strong organic and direct demand. The product is excellent and the brand is loved; the gap was never the experience — it was a growth engine that leaned too hard on the booking aggregators.

The gap was the path to booking. Interest arrived, but the route from a landing page to a confirmed booking was scattered — bookings closed over separate messaging numbers, a loose website flow had no structured steps, and the full-margin direct channel was quietly losing share to the booking aggregators.

The challenge — demand arrived, the path lost it.

The operator never had a traffic problem. It had a path problem: every step between "interested" and "booked" either leaked or couldn't be measured.

1. No structured path from landing to booking. A scattered website flow gave a visitor no clear, stepped route to a confirmed booking — interest arrived and dispersed.
2. Booking lived across separate messaging numbers. Conversations closed over several disconnected channels, so the funnel had no single, readable spine from inquiry to confirmation.
3. The leak was concentrated, and invisible. Most people who started a booking never reached the checkout step — but without measurement, no one could see where the drop happened.
4. The aggregators were the easier path. When the direct route felt harder than booking through a marketplace, demand defaulted to the aggregators — a margin and a price haircut at once.
5. No before to measure an after against. With nothing structured or tracked, even a real improvement couldn't be proven — so the first job was building something measurable.

Our approach — a funnel for every offer.

Before the rebuild, the funnel ran one landing page for every buyer, the same offer and the same ask for everyone, bookings scattered across separate messaging numbers, and Meta ROAS at roughly 8× before our rebuild. After, it delivered a +25% conversion uplift: a journey per offer — shared, premium shared, private — with message → landing page → offer matched to the segment, WhatsApp conversation and nurture tailored per buyer, and Meta ROAS at 20× blended on the rebuilt account.

The table below is the high-level cut. The working build is roughly 10× more nuanced — each offer splits by source market, group size, and lead-time, with its own page, message, and conversation script.

1. Rebuild the path — from one landing page to a journey per offer.

Strategic reasoning

Before, every visitor met the same landing page and the same ask — a backpacker comparing a shared seat and a couple planning a private charter were funnelled into one route. But a budget shared-seat buyer and a premium private buyer are in completely different moments. We replaced the one-size path with a journey per offer: shared tours, premium shared tours, and private charters each get their own creative message, landing page, offer, WhatsApp conversation, and nurture material — so the route fits the buyer instead of averaging them. We also consolidated the scattered messaging numbers into one tracked flow, so for the first time the funnel had a readable spine to measure.

2. Match every offer to its moment — then test.

Strategic reasoning

With a funnel per offer, the message can finally fit the moment — the bucket-list pitch for the first-timer, the comfort upgrade for the tier-up, the bespoke day for the private buyer. And because there are now several funnels instead of one, we can actually test: page against page, offer against offer, conversation script against conversation script — and move budget to what converts. The discipline that matters most is restraint about where to test. The leak sat at one step — getting from a started booking to checkout — so that's where the testing concentrates, not the parts that were already working.

• A dedicated landing page per offer — message, proof, and ask matched to the intent.
• The offer matched to the moment — accessible, premium, or bespoke — never one ask for everyone.
• The WhatsApp conversation and nurture material tailored per segment, not a single script.
• Structured testing across offers and pages; budget follows the winners.
• Tests concentrated on the blocking step, where the funnel actually leaked.
• Booking direct made the easier, more reassuring choice than the aggregator alternative.

Why direct now wins against the aggregators.

The booking aggregators were the easy default — and every booking they took was margin the operator rented instead of owned, at a lower value on top. The funnel rebuild is what turns that around. Four things a marketplace listing can't match now sit on the direct path:

Together they make booking direct the more trusted, better-fitted choice — so demand stops defaulting to the aggregators and the full margin stays with the operator.

The shift, before → after.

What rebuilding this funnel taught us.

1. One path can't serve every buyer. — A shared-seat explorer and a private-charter buyer are in different moments. Give each its own offer, page, conversation, and nurture, and conversion stops leaking on the averages.
2. You can't fix a leak you can't see. — The first move wasn't a CRO test — it was making the path measurable. The blocking step was only obvious once the funnel had one readable spine.
3. Many funnels make testing possible. — With one funnel there's nothing to compare. With one per offer, CRO becomes real: page vs page, offer vs offer, budget to the winners — and tests concentrated on the step that actually leaks.
4. Every aggregator booking is margin you rent. — Defending the direct channel isn't a nice-to-have — it's the economics. Make direct the easier choice and the full margin stays with the operator.

The rest of the Luxury Boat Tours engagement.

Luxury Boat Tours was one engagement told three ways. Keep going — here's the rest of the set:

1. Overview — the full engagement, end to end.
2. Performance Marketing — a spray-and-pray social account became an intent-led engine.
3. Data & Analytics — scattered numbers tied into one tracked loop.