When Leadership's Best Guess Meets What Guests Actually Do

How Resorts World Las Vegas Found $135,000 a Month Hiding in a Rate Plan Nobody Wanted

Margo O'Neill, director of revenue systems at Resorts World Las Vegas, almost lost an argument she didn't know she needed to win.

Leadership wanted to cut a rate plan called Signature Savings. It's the full retail rate, no discount, no food and beverage credit. On paper, it looked like clutter sitting next to offers that were actually designed to convert. The assumption was simple: guests are price-sensitive, and nobody books at full price when a discount is right next to it.

O'Neill wasn't so sure. So she pulled up Pendo's session replay tool and started watching. She went through 150 confirmed bookings. Eighty-five percent of them booked the full retail rate. Most didn't even glance at the discounted offer next to it. They knew their dates, clicked, and booked.

When she ran the numbers, that rate plan was worth roughly $135,000 a month in room revenue alone. Cutting it would have meant cutting that.

"It's very hard to argue when you have very concrete information and behavior," O'Neill said.

The finding came out of a pilot Resorts World Las Vegas ran with Balboa, a consulting firm that works with companies on Pendo implementations and customer data strategy. Keith Wagner, Balboa's co-founder and CTO, worked with O'Neill on the project and hosted a session this week walking through what they found. What started as a three-day trial run stretched into four months, largely because the data kept surfacing things nobody had gone looking for.

That's the pattern worth paying attention to here. This wasn't a story about replacing intuition with dashboards. It was about intuition being wrong in expensive ways, and data catching it in time.

The Signature Savings case wasn't the only one. Ahead of a summer campaign, the team noticed that a new luxury package wasn't booking as expected, despite driving the most traffic to the booking engine. The obvious explanation was price: it carried an $800 value bump, priced at $200 to $300 more than a standard room. Before the pilot, O'Neill said she probably would have made that same call herself.

Instead, she watched session replays and found something else. Guests were clicking into April and May dates that simply weren't available yet, since the offer's stay dates didn't start until after Memorial Day. The fix wasn't a lower price. It was moving the start date up. The rest of the summer offers took off once that was corrected, and the team held its rate instead of discounting reactively later in the season, a scenario O'Neill said would likely have taken another month to diagnose without the data.

The most expensive discovery, though, had nothing to do with pricing. The booking engine's checkout page had developed a severe lag. Multiple vendors got on calls to debug it and mostly pointed fingers at each other. Wagner eventually traced it to a page health feature in Pendo and found the culprit: a Reddit ad pixel that had no business being there. Once it was removed, the lag dropped by roughly 95%. O'Neill compared three weeks of bookings before the fix to three weeks after and estimated the impact at $500,000 to over $1 million in recovered revenue and conversion, depending on how the math was framed.

None of this came easily. Ninety-five percent of the site's visitors are anonymous, which makes tracking behavior across sessions and devices genuinely hard without a login to anchor to. O'Neill's team has been working to convert more of that anonymous traffic into known guests through the rewards program, partly because Resorts World runs three distinct brands under one roof, each attracting a different kind of traveler, and generic, one-size-fits-all promotions were leaving money on the table.

That last point may be the more durable lesson for anyone running a digital product with more than one customer segment. O'Neill's team started building personas based on behavior, not just past purchases: what a Conrad guest lingers on, what a Hilton guest actually books, what nobody in the Crockford's tier even looks at twice. Purchase history tells you what has already happened. Behavioral data, done well, gets closer to what someone is about to do.

O'Neill's advice to teams hesitant to invest the time: it's a real investment, and it won't feel efficient at first. But it tends to pay for itself, and often in places nobody was looking.

"You never know what you're going to find," she said. "It's not that data will just validate the assumptions you had. Sometimes something completely different crops up that's more meaningful than what you were trying to sift out in the first place."