Skip to main content

Fast website, findable content

A guest usually plans a holiday on a phone, with a handful of tabs open and little patience per page. When a page loads too slowly, one of the other tabs wins. And before that comes another hurdle: the guest has to find the page at all, through a search engine or AI assistant that can only recommend what it can read. Two conditions, then, neither visible on the page itself. Together they decide whether the photos, the description and the reviews are ever seen.

The wait before the first impression

A page that takes too long to appear loses the guest before the first photo has loaded. That guest rarely returns to find out what was there: the next tab was already open. On the road, on hotel wifi or a patchy train connection, the margin is smaller still.

Speed, in that light, is not a technical detail. It decides whether the visitor is still there when the content arrives. The most carefully chosen photos and the most honest description do nothing for a guest who has already closed the tab.

Search engines look at speed too

The same loading speed counts again, this time at Google. On a phone, that speed feeds into the ranking, not as the main factor but as the difference between pages that would otherwise be close together. For accommodations that appear alongside similar ones, that closeness is the norm rather than the exception.

This is why speed deserves attention even when a site looks fine on the operator's own screen. Testing usually happens on a fast connection and a device that has loaded the page before. The first-time guest on a phone, far from home, has none of those advantages. Visits from exactly that kind of guest count in the measurements Google partly bases its ranking on.

Seen by a person, invisible to a machine

Speed decides whether a page is seen. Whether it can be read is a separate question, and the answer depends on the kind of reader. A page can look complete to a visitor while a tool that reads only what the server sends finds little of it. The visitor's browser does the assembling, so the visitor sees everything.

Google reads pages in two passes: first the text the server sends, then once more with the scripts running. For most sites the second pass succeeds, with some delay. When the content only arrives in the page after scripts or external API calls have run, that second pass can stall. AI crawlers usually work differently and do not run scripts, so they depend on what the server delivers in one go. A page that is only complete after its scripts have run is largely unreadable to such a crawler.

What an AI crawler reads. Left, a page built by scripts, which stays empty. Right, a page that comes complete from the server, with photo, text and buttons.

Not just Google anymore

For a long time, being readable mattered mainly for Google. It now matters for a wider set of tools. Guests increasingly ask an AI assistant for a recommendation, in full sentences. The assistant bases its answer on what it has been able to gather about the subject, sometimes from pages it fetches on the spot, sometimes from material collected earlier in its training data. Which accommodations surface in that answer depends on much more than how a single page is built: reputation, mentions elsewhere on the web, partnerships that link back to the site. But without a readable page, that material lacks what only the operator can tell, such as the location, the number of bedrooms and the amenities. That is exactly what a guest asks about.

The shift towards AI assistants does not change the underlying work, but it raises the stakes. A page that is fast and readable was always better for guests and for Google. Now it is also the precondition for being counted at all by the tools a guest consults.

The groundwork first

Speed and readability are the groundwork beneath the photos and the reviews. They are not what convinces a guest, but they decide whether the things that do convince are ever reached. The work is largely invisible, which is also why it is easy to leave undone.

For accommodation pages, BonBooking takes this work off the operator's hands. The server delivers each of these pages in one go, readable by search engines and AI tools. What is not needed for reading, such as the map and the booking widget, loads afterwards, so what matters is on screen first. The content stays the operator's: the photos, the descriptions, the prices. The technical structure beneath them is the platform's responsibility. The examples show what websites on BonBooking look like in practice.

Being read is the starting point. The next step is understanding: letting a tool know that this is an accommodation, that this is the number of guests, that this is a review scored 9.2. How that meaning is made explicit, with structured data in the form of JSON-LD, is the subject of the follow-up: From being read to being understood.