How we engineered a funnel that cut a luxury villa developer's cost per lead by 70%.

Intro

We worked with Palma Islands— a luxury Bali villa developer and full-service real-estate firm with a strong product story and a Central European investor pipeline built entirely on referral and word of mouth. What they didn't have was a way to win cold international buyers at scale: their website was built for warm, referred visitors, not for the skeptical foreign investor who asks “is this even legal for me?” before anything else, and there was no paid acquisition system at all.

This page is the Performance Marketing lens: how cold spend across scattered international markets was separated, concentrated, and scaled onto the buyers who actually convert. The conversion path it fed lives in Funnel & CRO; the message that earned the click lives in Performance Creative.

Meet Palma Islands.

Palma Islands is a Bali property developer and full-service real-estate firm working across the Berawa-to-Canggu corridor — villa sales, rental, land, build-to-order, ownership setup, and management. The product is real and delivered: dozens of completed properties, buyers from a dozen-plus countries, and genuine return records. For years, the pipeline filled itself through referral and reputation.

The gap opened the moment they wanted to grow beyond word of mouth. Bali property has moved from boom to maturation — supply is up, occupancy has compressed, and the qualified investor now goes to whoever can be founded and trusted online. Palma had the trust signals (delivery history, real yield, legal know-how) but none of them were visible to cold paid traffic, the contact path dead-ended in a live chat, and there was no landing page, no lead capture, and no nurture. They didn't need a vendor to run ads; they needed someone to build the acquisition engine from zero.

The challenge — a path built for warm referrals, leaking cold buyers.

The website assumed every visitor arrived already trusting Palma. Paid traffic doesn't. The result was a conversion path that leaked at every stage — and a read that proved it: only about 3% of the people who clicked to message actually did.

1. One page for every buyer. A single scrolling catalogue served a return-led investor, a second-home buyer, and someone relocating to Bali — none of them well, because none of them wanted the same thing.
2. It answered the wrong questions. The page assumed the trust a referral already has. The cold foreign buyer's first fear — can a foreigner even own this? — went unaddressed entirely.
3. The contact path dead-ended. Every lead required a live-chat conversation, and only about 3% of the people who clicked the message button actually sent one. The funnel was losing all but a sliver at that single step.
4. No lead capture, no nurture. The contact form went nowhere structured — no automation, no follow-up — so a lead that wasn't ready to buy on day one had no path back.
5. No way to convert interest without a real-time chat. No downloadable asset, no gated magnet, no dedicated destination — so genuine interest that wasn't ready to message was simply lost.

Our approach — read first, then build a funnel per buyer.

1. Read the funnel before touching it.

Strategic reasoning

Most agencies inherit an account and start changing pages. We didn't touch a thing in Week 1 until the funnel was readable. The stage-by-stage diagnostic surfaced the real leak — only about 3% of people who clicked to message actually did — which told us the problem wasn't the top of the funnel at all. It was the moment a cold buyer was asked to commit to a live conversation before the page had earned it.

• Stage-by-stage funnel diagnostic across impression, click, landing-page view, and lead.
• The blocking step identified: the jump from interest to a live-chat message.
• No page or creative changed until the read was clear.

2. Stack five fixes in one week — not one slow test.

Strategic reasoning

The landing-page-to-lead stage was leaking from every direction at once — the page, the message prompt, the autoreply, the business profile. Single A/B tests would take a month to move it. Five changes hitting the same stage in the same week moved it in seven days: we rewrote the message pre-fill and autoreply for less friction, restructured the page into curated collections, split the audience, and rewrote the business profile around credibility signals.

• Message pre-fill and autoreply rewritten for lower friction, higher intent.
• Landing page restructured into curated collections instead of one endless scroll.
• Business profile rewritten around delivery history and credibility.
• Result: a 3.4× lift in landing-page-to-lead conversion in seven days.

The table below is the high-level cut. The working build is roughly 10× more nuanced — each persona page splits by price tier, lead-time, and market, with its own collection, magnet, and nurture cadence.

3. Build the path per buyer — page, magnet, and nurture.

Strategic reasoning

A return-led investor, a second-home buyer, and someone relocating to Bali are three different people with three different fears. So we stopped sending them to the same place. Each buyer now lands on a page built for their intent — split again by price tier into curated collections — with its own lead magnet and its own follow-up. When the page matches the buyer, they stay: average session duration rose over 350% against the old one-size page.

4. Bypass the leak entirely with the magnet.

Strategic reasoning

Some of the leakiest stages didn't need to exist. Rather than keep optimizing the live-chat step, we built a gated lead magnet for each persona and captured it on a Meta Instant Form — bypassing the website and the chat altogether. The magnet did two jobs: it made the form cheap, because it gave a buyer a real reason to share contact details, and it warmed the follow-up, because the conversation now opened with the exact asset that buyer wanted. That route ran roughly 90% cheaper than the funnel we'd been optimizing.

• A gated lead magnet per persona, captured on a Meta Instant Form.
• The website and live-chat steps bypassed for the buyers who weren't ready to message.
• The magnet doubling as the opener for the tailored nurture sequence.

From cold impression to booked site visit.

The rebuild mapped to the full investor journey — and the Instant Form track was engineered to collapse the two stages that leaked most, landing-page engagement and the live-chat ask, into a single in-platform capture.

Cold impression → booked site visit

• Aware: First brand impression · Meta Ads, cold international audiences
• Interest: Researches Bali villas · dedicated landing page, built per buyer
• Consider: Evaluates the options · ownership legality, delivery proof, documented returns
• Engage: Enters direct conversation · WhatsApp flow, Instant Form leads
• Convert: Site visit, deposit, sale · 3–5 qualified leads/day, CPL −70%

The shift, before → after.

What rebuilding this funnel taught us.

1. Read the funnel before you spend. — The leak wasn't where anyone assumed. Only about 3% of message-button clicks became a message — and you only find that by diagnosing the path before changing it.
2. Stack optimizations, don't isolate them. — Five changes hitting the same leaking stage in one week moved it 3.4×. Five sequential A/B tests would have taken five months.
3. One page can't serve three buyers. — An investor, a second-home buyer, and a mover want different things. Give each a page, a magnet, and a nurture built for them, and they stay — session duration up over 350%.
4. The lead magnet is the conversion accelerator. — It made the form cheap and the follow-up warm. When the leakiest stage fights back, sometimes the answer is to bypass it, not optimize it.

The rest of the engagement.

This is one of three lenses on the same build. Funnel & CRO covers the conversion path this spend fed, and Performance Creative covers the message that earned the click