How I Audit Hotel Websites

What counts as a hotel

I only look at properties that compete with each other: real hotels that match the target market for their location and are professionally run. Apartments, holiday rentals, motels, and budget or economy lodging serve a different guest with a different booking process, so they’re not included.

In a typical city, a search for lodging returns far more results than there are real hotels, often two or three times as many once apartments and private rentals are counted. I filter those out and keep the properties that compete for the same guest. In smaller destinations, where excellent hotels are often very small, I judge by the quality and positioning of the property rather than applying a fixed room count.

I also tend to exclude chain hotels whose website lives on a sub-domain of a parent chain domain, rather than on their own dedicated domain. These properties can’t control or improve their own website, since it’s built and managed centrally by the chain. Measuring it tells you about the chain’s template, not the individual hotel.

How I find the properties

I start from Google Maps data for the target area, then filter down to the qualifying hotels described above. This gives me a complete, reproducible starting set instead of a hand-picked one. The goal is to measure every qualifying hotel in a location.

What I measure

Each qualifying hotel’s website is assessed on the same set of metrics. Plain-language definitions of every term, and why each one matters for bookings, are in the metrics glossary.

• Platform (CMS): what the site is built on, such as WordPress, TYPO3, Wix, Squarespace, or a custom system.
• Core Web Vitals: Google’s real-user assessment of loading speed, responsiveness, and visual stability on mobile.
• PageSpeed Insights: lab performance scores for mobile and desktop, used as a diagnostic alongside the real-user data.
• Booking path: whether a guest can book direct through a booking engine, is sent to a contact form, or is pushed out to an online travel agency.
• Structured data (JSON-LD schema): whether the site helps search engines understand what it is.
• Reputation: Google rating and review count, recorded for context alongside the website’s technical state.
• Technical basics: HTTPS, index status, and other foundational signals that affect how a site is found and trusted.

How I measure performance

Performance comes from two sources that answer different questions. Core Web Vitals come from the Chrome UX Report, which reflects how real visitors actually experience the site. This data needs enough real-user traffic to exist, so sites below that threshold are marked “no field data” rather than pass or fail, and that absence is itself a rough sign of low traffic.

PageSpeed Insights provides lab scores for every site regardless of traffic. Where a score is reported, it’s the median of several runs to reduce the natural variation between tests. A site can score well in the lab and still fail in the field, or the reverse. That’s why I report both.

Everything else, I verify by reviewing each live website by hand.

How I grade photography

One of the things I assess on each website is the photography, on a simple scale from very good to very bad. This is the one part of my study that’s openly subjective. Photography is my other trade, so treat it as my professional opinion, not a measured score.

The grade is about how the photos are used, not just how good they are. A hotel can have beautifully shot rooms, but if a spacious 40 m2 room gets a single photo, with no gallery and no way to enlarge it, that’s a missed opportunity, and it won’t earn a high grade from me.

If you want to see what I think strong hotel photography looks like, that’s something I do myself. You can read more about it on my hotel photography page.

How I keep it fair

Every figure reflects the live state of each website on the measurement date, which is stated on each audit. Websites change, and a site measured as slow today may well be improved tomorrow. That’s a good outcome, and it’s why the date matters.

I report measurable facts, like a specific performance score. Where I highlight the sites with the most room to improve, I frame it as opportunity, and I keep the underlying data so any figure can be checked. Where platform or plugin versions are detectable, they’re only ever reported in aggregate across a city, never published for an individual property.

Questions

If anything here is unclear, or you would like your own property assessed, get in touch.