This is the story of how a scuba diving centre's booth became a live product experiment — where an AI-informed pricing model, a gamified loyalty card, and a handful of shamelessly funny memes turned foot traffic into $20,000+ in three days, and a program that's still running today.
Diving trips in Indonesia run IDR 500,000–975,000 a session — a luxury purchase, not a habit. Most operators win a customer once and never see them again. Retention was the real product problem, not acquisition.
Asia's No.1 diving expo, 56th edition, three days, thousands of serious divers walking the floor of the Shanghai World Expo Exhibition Centre. A booth is a funnel with a 72-hour clock — every decision had to compound fast.
Instead of treating the booth as a one-off sales push, I ran it like a product launch: a roadmap, a pricing model, a loyalty mechanic, and a distribution channel — each one measurable, each one meant to outlive the show.
Defined and executed a cross-functional product roadmap that combined AI-driven personalisation with machine-learning-assisted pricing — reading demand signals in real time and adjusting flash-sale thresholds instead of guessing at a single static discount.
Most loyalty programs ask customers to earn trust before they see any benefit. This one flipped it: the reward card went out at the first meeting, with the benefit visible up front — a bet that reducing hesitation early would outperform reducing it later.
"People don't want to buy a quarter-inch drill. They want a quarter-inch hole." The loyalty card wasn't a punch-card afterthought — it was designed around Customer Lifetime Value as the north star, not the single transaction. Because diving isn't a habit customers form on their own, the card had to manufacture the habit: one stamp per invoice, rewards communicated from the very first conversation, so people could see exactly what buying more, and diving more, would earn them.
It was handed over at the first meeting — before hesitation had a chance to set in — as a quiet signal that ScubaGo was building a relationship, not closing a one-time sale.
Social-driven Flash Sale and Early Bird channels went live on Xiaohongshu ahead of and during the show — pricing informed by the same data and ML models behind System A. Instead of polished ad copy, the creative leaned into the platform's native language: diver-brained meme formats, a "3 dives for 675 RMB" hook repeated until it became a running joke on the timeline, and a very sad, very persuasive cat.
It worked because it didn't read like a booth flyer. It read like something a diver would send their dive buddy.
The loyalty programme handed out on the Shanghai floor in March 2025 wasn't a show promotion — it's the system ScubaGo still runs on. That's the actual test of a product roadmap: not whether it moved a number for three days, but whether it was worth keeping after the booth came down.
Roadmap defined → executed → still in production