We looked for 2,447 hostels' email addresses. Half publish none.

We went looking for 2,447 email addresses

Nightfill keeps a catalogue of independent hostels and small hotels across Latin America and southern Europe, because we have to know who we are talking to before we talk to them. This morning it holds 2,447 properties across 54 named cities. Building it is a boring, mechanical job: read whatever OpenStreetMap carries for the property, then read the property's own website. Two places. The two a machine looks at first, and roughly the two a guest in a hurry looks at first.

Here is the finding worth your morning: 1,213 of those 2,447 properties, 49.6%, came back with no email address at all.

Say precisely what that means, because the honest version is more useful than the dramatic one. It does not prove those properties have no email. It proves the two most obvious public places came up empty. They may answer on WhatsApp in forty seconds. They may have an address buried three clicks inside a contact form. What they do not have is an address a stranger can find in ten seconds without asking somebody for it.

What the other half publishes

1,215 records carry a usable address. Sort them by what sits before the @ and the shape is consistent enough to be a rule:

• info@ - 579 properties, 47.7% of everyone who publishes anything
• reservas@ - 93, reservations@ - 39, booking@ - 22, bookings@ - 9
• contact@ - 24, contacto@ - 19, hola@ - 18, hello@ - 16
• admin@ - 6
• something else, usually a person's first name - 390

Add up every address that names the job the mail is for: reservas, reservations, booking, bookings. That is 163 out of 1,215, or 13.4%. Add up the generic front-desk addresses instead, info plus contact plus contacto plus hola plus hello, and you get 656, or 54%.

So slightly more than half of the properties that publish an address at all publish one that does not say what it is for.

Why the prefix is an operating decision, not a detail

An info@ inbox is a shared inbox, and a shared inbox is one nobody owns. A message reading "do you have two beds for the 14th and 15th?" lands there next to a laundry quote, a supplier invoice, a job application and four pieces of spam. It is a sale with an expiry date on it, queued behind the electricity bill.

Your competition on that enquiry is not the hostel down the street. It is the tab the guest already has open. Someone who cannot find your address in ten seconds opens Hostelworld or Booking, where the same bed of yours is one tap away and the answer is instant. That path works. It also charges. We took one of our own real eight bed-night Hostelworld bookings apart line by line earlier this month: three promotions that multiply rather than add, and an effective commission of 26.7%. That is the toll on the convenient path, paid on a booking you would have had anyway.

Which reframes the twenty minutes this post is about. A named inbox does not win you new guests. It stops you paying commission on the ones who were already trying to reach you.

The 93 that borrow somebody else's domain

Of the 1,215 published addresses, 93, or 7.7%, sit on a domain that is not the property's own website. In practice that is a free mailbox: gmail, hotmail, or the old address of whoever built the site in 2019.

It is not an artefact of where we looked. The rate holds either way we found it: 17 of 223 scraped from the property's own site (7.6%), 19 of 265 read off an OpenStreetMap tag (7.2%).

Two things follow, both measurable. The first is trust: a guest comparing your booking page against an OTA checkout reads that address, and a free mailbox on a property that does have its own website reads as a side project. The second is deliverability, which operators tend to discover late. A from-address on a domain you do not control is one you cannot authenticate, so your confirmations and payment links get filtered harder than the OTA's. We live under the same rules on the sending side: across our last 60 outbound messages our own hard-bounce rate ran at 6.7%, and the response was to stop sending and clean the list, not to send more. Deliverability is not a marketing problem. It is an inventory problem wearing a different hat.

Four checks, about twenty minutes

• The stranger test. Open a private browser window, search your property by name, and time yourself finding an email address. Over thirty seconds is a fail. You cannot run this logged in; you already know where everything is.
• Name the inbox for the job. reservas@ or bookings@ on your own domain, forwarded to wherever you actually read mail. It costs nothing, it takes one DNS record, and it tells the guest the message landed in the right place.
• Put it in three places. Your site footer, your Google Business Profile, and your OpenStreetMap entry. The third is not vanity. Of the addresses in our catalogue whose source we recorded, 265 of 488, or 54%, came from an OpenStreetMap tag rather than the property's own website. That tag is how machines, apps and aggregators read you, and most operators have never opened it.
• Time your own answer. Send an enquiry from an address nobody at the property recognises, at nine at night, on a Saturday. Write down when it gets answered. That number, not your ADR, decides how many enquiries become bookings.

What our own data says about yours

One more number, and this one is against us. The catalogue is messier than its totals make it sound. 1,194 of the 2,447 records carry no city label at all. "Oaxaca" and "Oaxaca de Juarez" are counted as two different cities; so are Medellin with the accent and without. Mexico City appears under seven different labels, mostly because somebody wrote the neighbourhood into the city field. The discovery run at 00:04 Mexico City time this morning checked eight cities, found 486 candidates and added 55 that were not already on the list. The run before it, at 12:04 the previous afternoon, failed on two of its eight cities, and we know that only because it says so.

We publish that because it is the same lesson pointed back at us. Nobody's operational data is clean, ours included. The difference between a list that works and one that quietly rots is not care at the moment of writing. It is a check that runs whether or not anyone is in the mood.

None of this moves your occupancy this week. A named inbox on your own domain, listed in the three places a stranger and a machine both look, plus an answer time you have actually measured, is not a growth tactic. It is the difference between an enquiry that reaches you and one that reaches an OTA instead, and the second arrives with a quarter of the rate already gone.

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