We had a hypothesis and it turned out to be wrong. We assumed that big international resort chains — with corporate marketing budgets, in-house web teams, and millions of dollars at stake — would have fast, well-built websites, and that small independent Dominican hotels would be the ones struggling. So we measured it. We ran 36 hotel websites in Punta Cana, Bávaro, Bayahíbe and La Romana through Google's PageSpeed Insights and fetched each homepage directly to inspect what it actually contains. The results were worse than we expected across the board — and the pattern was the opposite of what we predicted.

What we did

Every performance figure here comes from Google's PageSpeed Insights API running Lighthouse against a simulated mid-tier phone on a throttled network — the same tool any developer can point at any site. Separately, we fetched each homepage with a mobile user-agent to check language versions, contact links, structured data, and booking setup. Nothing is estimated: every number is read from an actual response.

The sample is 36 hotels — 22 large resorts, 10 mid-size properties, 4 small independents — measured on 6 September 2026. Two honest limitations up front: these are lab scores, not real-user data, and we measured homepages only. Room and booking pages may differ.

The headline numbers

The median hotel scored 29 out of 100. The mean was 33. The best site scored 84; the worst real measurement was 16.

• 86% scored under 50 out of 100

• 53% scored under 30

• 97% took longer than 4 seconds to display their main content

• 83% weighed more than 5 MB

The median Largest Contentful Paint — the moment the main content actually appears — was 17.1 seconds. Google's guidance is 2.5 seconds. The median Dominican hotel site in our sample took nearly seven times that long to show a visitor what they came for. The slowest verified measurement was over 50 seconds.

The median page weighed 8.05 MB, with one property at 28.7 MB — a figure we re-measured to be certain, and which reproduced exactly. For context, that's a single homepage weighing roughly as much as a feature-length album download, delivered to someone on hotel Wi-Fi or a roaming data plan.

The finding that surprised us

Here is where our hypothesis collapsed:

• Large resorts | 22 hotels | mean 26.5 | median 27

• Mid-size | 10 hotels | mean 35.4 | median 33

• Small independents | 4 hotels | mean 64.0 | median 60

Small independent hotels scored more than twice as well as the large resort chains. Every one of the five best-performing sites was a small or mid-size property. Every one of the ten worst was a large or mid-size resort.

The large resorts were not merely mediocre — they clustered at the bottom. Twenty-two properties belonging to some of the biggest hotel groups operating in the Caribbean produced a mean score of 26.5, and not one of them cracked 50.

Why? Almost certainly because scale brings weight. Corporate resort sites carry booking widgets, chat systems, analytics stacks, personalization scripts, tag managers, and marketing pixels — each added by a different department, none ever removed. One large resort's homepage issued 244 separate network requests. A small hotel's site does one job with a fraction of the machinery, and it shows.

Where the weight comes from

Images accounted for a mean of 35% of total page weight — significant, but not the whole story, which surprised us too. The rest is code: the median site had 1.76 seconds of Total Blocking Time, meaning the browser sat frozen and unresponsive for nearly two seconds while processing JavaScript. The worst was over 14 seconds of blocking.

That distinction matters practically. Compressing images is the fix everyone reaches for first, and it helps — but on these sites the deeper problem is accumulated scripts. You cannot optimize your way out of 244 requests; that requires removing things, which is an organizational problem more than a technical one.

What they got right — and what we got wrong twice

We expected to find a bilingual gap. We didn't.

• 97% had a distinct English version

• 92% had a distinct Spanish version

• 92% had both

• 0% used a Google Translate widget

Not a single site in the sample relied on an auto-translate plugin. Whatever else is wrong with these websites, the tourism industry has genuinely solved multilingual publishing — 89% carry proper hreflang tags. That's a real credit to the sector, and it corrected a second assumption of ours.

Structured data was also widespread: 86% ship JSON-LD, most commonly Hotel and PostalAddress markup. 100% had a mobile viewport tag. None had a broken HTTPS setup.

One quiet finding worth flagging to anyone in the industry: 15 of the 36 sites still ship FAQPage structured data, which Google fully retired in 2026. It produces no search feature anymore. It's harmless, but it's maintenance debt sitting in production — and a decent proxy for how long it's been since anyone reviewed these sites.

The gaps that do exist

WhatsApp is largely missing. Only 25% of sites had any WhatsApp link — in a country where it's the dominant business communication channel. The split by size is telling: 3 of 4 small hotels had it, against 4 of 22 large resorts. The small operators understand their market; the corporate sites are built to an international template that doesn't account for how Dominicans and visitors here actually make contact.

Direct booking is weaker than expected. Only 33% offered booking on their own site. 44% sent visitors to Booking.com, Expedia or similar, and 22% offered no booking path at all from the homepage — just a contact form. Nearly half of these hotels are paying commission on bookings that arrived at their own front door.

Accessibility is mediocre. The mean accessibility score was 81, and 81% scored under 90 — contrast failures, unlabeled buttons, missing alt text. Ordinary problems, and ones that also affect how well search engines read a page.

The study behind the study

While building the sample we ran a separate check that produced the most striking result of the whole exercise. We took 20 small hotels, guesthouses and aparthotels from OpenStreetMap in the Punta Cana and Bayahíbe areas and asked a simpler question: does this business have a website at all?

Of the 16 we could positively identify, 12 had no website of their own — only listings on Booking.com, TripAdvisor, Expedia, or aggregator directories. One had a domain that returned an error and redirected to Facebook. Another's only web presence was a Facebook page. Three of the four that did have sites shared a single operator's domain.

That reframes the performance results entirely, and we should say so plainly: the small hotels that scored well are survivors. They're the minority that built a website at all. The broader small-hotel market isn't slow online — it's largely absent, renting its entire digital presence from platforms that charge commission and own the customer relationship.

What we'd tell a hotel to do about it

Three things, in order of return:

Find out where you stand. Run your own homepage through PageSpeed Insights on the mobile tab. It's free and takes a minute. If you score under 50, you are in the majority — which is not comfort, it's opportunity, because your competitors are equally slow.

Cut before you compress. Optimizing images matters, but the median site here lost nearly two seconds to JavaScript execution. Audit what scripts are loading and why. Most large sites carry tools nobody uses and nobody removed.

Own the booking and the conversation. If you're one of the 44% redirecting to an OTA, every commission you pay on a visitor who was already on your site is avoidable. And if you're one of the 75% of large resorts without a WhatsApp link, you're missing the channel this market actually uses.

Caveats we want on the record

PageSpeed lab scores vary between runs. We re-measured the four most extreme results and found real instability: one property scored 7 on the first run and 19 on two subsequent runs — so we've excluded that 7 from everything above and named no hotel among the poor performers, because a single run isn't a fair basis for a public claim about a specific business. Another site ranged from 20 to 50 across two runs. The two stable results we did verify — the 28.7 MB page weight, and the top score of 84 — reproduced exactly.

Run-to-run noise averages out across 36 sites, so the distributions and category comparisons here are sound. Individual hotel figures are not, which is why we've named only the best performer: Hotel Villa Iguana in Bayahíbe, which scored 84 and re-measured at 82 — a small independent hotel outperforming every international resort chain in the sample.

The sample is also modest at 36, weighted toward large resorts, and covers the southeast rather than the whole country. Anyone is welcome to re-run it; the tool is free and public, which is rather the point.

What this means

The Dominican hotel industry has solved bilingual publishing and largely adopted structured data. It has not solved speed, and the businesses with the most resources have the worst-performing websites in the sample. Meanwhile most small independent hotels don't have a website at all.

For a small or mid-size property, that's an unusually clear opening. You are competing for attention against resorts whose sites take 17 seconds to load, and speed drives both search rankings and bookings. A fast, genuinely bilingual site with WhatsApp connected and direct booking is not an exotic ambition. In this market it is a genuine competitive advantage, and the data says almost nobody has taken it.

At DR Web Studio that's what we build. If you'd like to know how your own site scores — and what specifically is slowing it down — contact us for a free consultation and we'll measure it and tell you honestly.