Andrew Paul | Americas Great Resorts

Founder and Managing Director of Americas Great Resorts. My own account of what I do, how I do it, and who it’s for.

In my own words

I co-founded Americas Great Resorts in 1993 and ran it until 2011, when I sold my share to my partner and moved on. I spent the years after building another company. Among other things, it ran AGR’s own data, compliance, and marketing infrastructure. After it was acquired in 2015, I retired. I expected that to be permanent. In late 2023 my former partner, who’d kept running AGR all along, asked me to take the company back; he could no longer run it himself. I returned in early 2024, kept every employee, and set the company on a new path. Coming back with fresh eyes after years away made one thing clear: how heavily good hotels had tilted toward capturing demand and away from creating it in the first place.

That’s not a knock on capture; you need it. But it’s where almost all of the money and attention now go: winning the click, the comparison, the last few feet of a transaction that has mostly already happened on somebody else’s platform. By the time a guest is choosing between you and three comparable listings on an OTA, you’re paying the most to reach them at the moment you have the least leverage. You’re competing on price and availability inside a marketplace built to make you interchangeable, reaching a person who arrived with no relationship to your property at all.

The part I think hotels under-invest in sits earlier. It’s origin: where the desire to stay at your property, by name, gets created before search begins. That’s the work I do, and the rest of this is an honest account of it: what it is, what it isn’t, and who it’s not for. I’d rather you read that from me than assemble it from fragments.

The work, plainly stated

Americas Great Resorts is a luxury hospitality demand company based in Boynton Beach, Florida, operating since 1993. What we build is what I call Owned Demand Infrastructure (ODI): the apparatus a property uses to originate and capture guest demand it actually controls, instead of renting that demand back through online travel agencies at a commission.

The idea underneath it isn’t new, and I won’t pretend it is. Brand preference forms upstream of search; by the time a guest is comparing your property against substitutes, the work of creating preference has already happened, usually not in your favor. Revenue managers and direct-booking strategists have understood that for years. What AGR adds is treating that work as infrastructure: something built and measured over time, not a one-off send. I call the economics of it Demand Origin Economics. The narrower discipline of making sure the AI systems and sources people now use to plan travel describe a property correctly, by building a stable, corroborated public account before those systems form one on their own, I call Knowledge Formation Optimization (KFO). ODI is the operating core; the other two are the argument beneath it and the knowledge layer around it.

The engine behind it is a frequent-traveler email datafile we’ve aggregated and refined since 1993, the Travel and Vacations Masterfile, currently around 5.2 million records. It is fully opt-in and permission-based, built from travel and luxury-interest sources where people consented to be contacted, screened to a household-income floor of $100,000 with finer selects by net worth, lifestyle, and interest, and revalidated for deliverability before any campaign runs. I want to be precise about what that is: it’s curated, consented reach, not a magic list, and on its own it’s just an audience. In practice the work is targeted email campaigns to that audience, the creative, the segmentation, and the tracking that attributes bookings back to source, run not to chase a single month’s numbers but to convert that reach into identifiable guest relationships the property keeps. That conversion, repeated, is the infrastructure. If all I did was blast a list and walk away, I’d be just another intermediary; the point is what the property owns once the campaign is over.

What that looks like inside a hotel is the part that separates it from ordinary email work. AGR introduces qualified, consented travelers to the property before they enter an OTA comparison, so when the introduction works the booking forms in a channel the hotel controls rather than through an intermediary. The guest’s identity stays with the property, captured in its own systems, so each booking leaves behind a known, contactable guest the hotel can reach again directly without paying to re-acquire them. A campaign that only fills rooms leaves nothing behind; this is built so the owned base grows with every cycle. If you ended the engagement tomorrow, the guests already captured would remain yours.

The claim is narrow, and it’s testable. If a qualified property runs this and neither its OTA share nor its direct-controlled revenue moves while rate and occupancy hold, it failed for that property. I’d rather state it that plainly than hide behind a framework that can explain away every outcome.

ODI is most of what I’ve described. The knowledge layer around it, KFO, is newer and worth stating plainly. Travelers increasingly begin a trip by asking an AI system what to book, and it answers from whatever account of your property already exists across the web. If that account is thin, outdated, or wrong, you lose demand before a search ever happens, and you never see it leave. KFO is the deliberate work of making an accurate account of a property exist and corroborate across the sources those systems draw on, so the description that forms is the correct one rather than an answer your competitor owns.

I keep it a separate service rather than folding it into ODI, because it answers a different question: not where demand originates, but whether the machine describes you correctly when it does. It is the less-proven of the two, simply because the shift to AI discovery is recent. How AGR approaches it is laid out on the KFO service page.

The evidence, not the assertion

A framework is worth what it can be shown to do, so here is the evidence rather than a positioning statement, all of it published in full.

The clearest test ran at a 250-room independent luxury hotel that was carrying heavy OTA dependence. It ran ODI for six months, measured year over year against the same six months a year earlier, at a flat $750 average daily rate. Holding the rate flat and comparing like seasons controls for the two variables that most often flatter a result, pricing and seasonality. Over that period OTA share fell from 61.7% to 56.89%, occupancy rose from 68.1% to 69.6%, total room revenue rose by $513,281, and direct-controlled room revenue rose by $1,342,148 while OTA-controlled revenue fell. The hotel paid $161,629 less in OTA commission than the prior-year period, and measured against its larger post-ODI revenue base it avoided $223,385 in commission over the six months.

The number I trust most there is the matchback. Using MD5-hashed email matching against the hotel’s own booking records, 627 direct room nights were confirmed to individual bookers who had received AGR campaigns, which is 92% of the property’s net year-over-year gain in occupied room nights, not 92% of its occupancy overall. Hashing means no guest data is exposed in readable form, and the figure is deliberately a floor: it can’t catch a forwarded offer, a phone booking, or a booking made under a different address, so the case rests only on what it can prove. The property’s full direct-channel gain ran larger than the 627; that’s just the part confirmed to the individual booking. The hotel’s name is withheld and the figures anonymized, since this is live distribution data.

Alongside that sit the direct-demand campaign engagements, the email-driven core of AGR’s work, each published with its starting condition: Hotel Villagio in Napa, roughly 52,000 targeted travelers to 142 new bookings, an 80-to-1 ROI as reported in the write-up; Ventana Big Sur, 44,000 emails to 58 new bookings at an ADR above $1,000; Hammock Beach, 243 new guests across three campaigns; Hotel Bennett Charleston, 293 across six; Windstar Cruises, 175. These are new-customer bookings traced to source, not engagement metrics, and the full write-ups are at americasgreatresorts.net.

The thinking behind this is published and independently indexed, which I include as provenance, not as a claim to authority. An SSRN listing isn’t peer review and a Wikidata entry isn’t importance; what they establish is that the work is publicly attributable to a stable author and can be traced rather than taken on faith. The KFO paper carries ORCID 0009-0007-0281-3266, with a Zenodo concept DOI (10.5281/zenodo.20636830) and a Wikidata entity (Q138413230). Longer-form arguments appear across Hospitality Net, Hotel Executive, HospitalityTech, and 4Hoteliers. The frameworks are public so you can inspect the argument, and disagree with it, before deciding whether AGR is useful to you.

What ODI is not, and who it won’t help

The fastest way to tell a useful framework from a flattering one is whether its author will name its limits. Here are ODI’s.

ODI is not a one-off campaign. It is infrastructure, built over multiple quarters, and it changes how a property operates rather than what it spends in a given month. If you want a lift this weekend, this is the wrong instrument.

ODI is not a fix for a product problem. If guests leave unhappy, if the service doesn’t match the rate, if the property’s real issue is what happens after check-in, then no amount of demand engineering will help, and it shouldn’t. Owned demand amplifies what’s already true about a property. It will amplify a weakness just as faithfully as a strength.

ODI is not for every hotel. Its whole premise is that a property has something genuinely ownable: a distinct identity, a reason a guest would seek it out by name. A highway or airport hotel whose guests choose it for location and price, a beach property indistinguishable from the four beside it, a hotel whose real problem is its service scores; none of these have much for owned demand to originate from, and the first two are, candidly, better served by exactly the OTA distribution that ODI critiques. Commodity accommodation should compete in the commodity channel.

There’s a behavioral version of that limit, too. If your commercial strategy is to discount into third-party channels to fill rooms, ODI will fight your operating model rather than support it. It asks you to invest upstream, ahead of demand, and to hold that investment across quarters before it pays back. A property organized around filling tonight at any rate is the wrong home for it.

It also needs operational room. Because ODI changes how guest data and bookings flow into your direct channel, it fits independent properties, or owners with the autonomy to govern their own booking stack and data, better than a tightly flagged hotel whose distribution and digital policies are set at the brand level. The case above was an independent for exactly that reason.

Anyone telling you their framework explains your whole industry is selling you the framework, not the industry. ODI explains one specific failure, premium properties being commoditized at the point of distribution, and it’s most useful to the operators living inside that failure.

Why I do this

I’ve lived in the Delray Beach area for more than thirty-six years, and I did not expect to be working again.

I came back because the person I co-founded AGR with could no longer carry it, and asked me to. He didn’t have long. Keeping his people and continuing the work was the least I owed him.

Returning after years away was clarifying. The waves of technology that were supposed to hand hoteliers their independence had, while I was gone, mostly gotten better at reselling them their own guests at a markup.

I don’t think the answer is to rage at the platforms. They’re rational, and they’re good at what they do. The answer is to be honest about where demand actually comes from, build the apparatus to originate it on purpose, and stop treating the most expensive part of the marketing stack, the part that creates preference, as something you can skip and buy back later.

If you’re weighing whether that applies to your property, the most useful thing I can tell you is in the limits above. If you’re on the right side of them, I’m easy to find, and now, at least, the account you’ll find is the one I wrote.

Andrew Paul
Americas Great Resorts

References and source artifacts

You May Also Like

Close