Ninety-six of 1,266 properties publish a contact address that is not at their own domain

Ninety-six properties publish an address they do not own

We keep a working list of independent hostels and small hotels across Latin America, Spain and Portugal. It stands at 2,610 properties today. Of those, 1,266 publish a contact email address we can read off their own website or their map listing. The other 1,321 publish none at all, which is a different problem and one we have written about separately.

Inside those 1,266, we compare two things on every record: the registrable domain of the property's website, and the domain of the email address the property publishes. Ninety-six of them disagree. 1,170 match. There is no third bucket here: every record in this count has both a website and an address, so none of the 1,266 is a "we could not tell".

Ninety-six out of 1,266 is 7.6 percent. Roughly one property in thirteen has a website at its own domain and a contact address at somebody else's.

Why I believe the number instead of blaming our scraper

The first thing I assumed was that this was our fault. A scraper that reads a contact page can easily pick up the web designer's address, or the booking widget vendor's, and file it as the hostel's. If that were happening, 96 would be a bug report about us, not a finding about the market.

Our records carry how each address was found. 237 were scraped from the property's own website. 304 came from an OpenStreetMap contact tag, filled in by the property or by a mapper. For 725 the provenance was never stamped, which is its own honest answer: we do not know.

The 96 mismatches split 17, 22 and 57 across those three groups. As rates that is 7.2 percent of the scraped, 7.2 percent of the map-tagged, and 7.9 percent of the unstamped. If our scraper were the problem, the scraped group would be the worst by a wide margin. It is the joint best, to a tenth of a point. The mismatch travels with the property, not with the route we took to find it.

Two smaller counts from the same pass. Four of the 1,266 publish an address that another property on our list also publishes: one mailbox, two businesses. And three would be refused outright by our own send filter today, because the address they publish sits at another company's domain entirely.

What a borrowed address actually costs you

None of this shows up on a P&L line called "email". It shows up in three places instead.

You cannot hand it over. A reservations address at a free provider gets created by whoever was at the desk the week it was needed. The thread history with every returning guest, every OTA account recovery link, every supplier and every tax invoice lives behind one password held by one person. When that person leaves, and they always leave, your handover is a password or it is nothing. A mailbox at your own domain is reassigned by an administrator in about ninety seconds and no guest notices anything happened.

You cannot route it. At your own domain, reservas@, grupos@ and facturas@ can land in three different inboxes, or all in one, and you can change your mind next month without reprinting anything. With one free-provider address, everything arrives in the same place and the person who reads it is whoever is holding the phone.

It reads as a smaller operation than you are. This one I cannot measure and I am not going to pretend otherwise. What I can describe is what the guest is comparing it against. Someone on your site deciding whether to email you or book the same bed on an OTA is weighing your contact line against a written cancellation policy and a card already on file. Your address is not what decides it. It is a free half point you are handing over in a comparison where the structure is already against you.

The arithmetic on the other side of the decision

The only reason any of this is worth twenty minutes is what a direct booking is worth next to the same booking through a channel.

We measure that at our own hostel in Oaxaca. On one real eight bed-night Hostelworld booking, three promotions stacked multiplicatively - length of stay, mobile and app - and multiplied out to exactly 0.729 of the rate. Once that was done, the effective commission on that booking came to 26.7 percent. Booking.com's worst case is encoded in our engine at 20 percent. For scale, the competitor floor our engine probes daily in Oaxaca has run between 168 and 206 pesos for a dorm bed three days out.

Take a bed you sell at USD 14. At 26.7 percent, the channel takes USD 3.74 of it. If a working contact address and a reply inside the hour move thirty bed nights a month from that channel into your own inbox, you keep about USD 112 that month. That is not new revenue. It is the same guest, the same bed, the same night, arriving by a different route.

Your rate, your channel mix and your reply speed are all different from mine, so run that line with your own numbers before you believe it. The point is only that the gap is a fifth to a quarter of the rate, and the address is the cheapest place in the whole chain to stop leaking it.

The twenty-minute version

• You already have the domain. Every one of the 96 has a live website, which is the only reason we could compare in the first place. Nothing needs to be bought.
• Create one mailbox on it. reservas@, hola@, hello@ - the word matters far less than the domain. Avoid info@: 601 of the 1,266 already use it, and a mailbox named after nothing in particular tends to be read by nobody in particular.
• Forward it to the mailbox you already read. You do not have to change a single habit. You are only changing what the guest sees and what you can take back later.
• Put it everywhere, not just on the contact page. Footer, OTA profiles, your map listing, the Instagram bio, the booking confirmation.
• Leave the old address forwarding for a year. Guests from two seasons ago will still write to it, and so will your accountant.

What I did not find

I cannot tell you these 96 properties take fewer direct bookings than the 1,170 whose addresses match. We do not have their booking data, nobody who does is publishing it, and I am not going to build a causal claim out of a domain string. That study does not exist in this dataset.

What the number does show is a risk that 96 operators are carrying without having priced it. On the day the receptionist leaves, the only address a guest can reach the property at leaves with them, along with every conversation ever had through it. Removing that risk costs an afternoon and no money. Most of them have not removed it, for the ordinary reason that nothing breaks on the day you do not.

If this sounds like your property, see the three Nightfill tiers and start a free 14-day pilot on your own data: view pricing.