Share
HOSTING RENEWAL - BILLING PHISHING

"Your domain expires in 3 days." It came from a real publisher's server.

This one landed in our own support inbox on 10 October, wearing a hosting company's branding, with a Manage subscriptions button that went to a payment page on a small Dutch publisher's website. The publisher's own website never served that page. Here is the DNS comparison that shows it, what that comparison genuinely proves and what it cannot, and why it decided exactly what we were allowed to block.

SafeBrowz Threat Research Security Research · · 12 min read

Verdict: phishing, and the sender domain is the giveaway

A renewal email that threatens to drop your domain in days and sends you to a payment page somewhere other than your provider's own site is a scam, however correct the branding looks. The one we received displayed the sender name "HostingerBilling", which is a display name anyone can type, and its button led to customerservice[.]cloudpressdelft[.]nl, a subdomain of an unrelated Dutch publisher. Hostinger's own domain, hostinger.com, publishes a DMARC policy of reject including subdomains, so most major mailbox providers will reject mail forged from it, though enforcement is each receiver's choice rather than something the brand can guarantee. These campaigns use an unrelated domain and dress up the display name instead, which costs nothing and works regardless of what the real brand publishes. Do not pay from the email. Open a new tab, sign in to your provider's control panel yourself, hpanel.hostinger.com for Hostinger, and look at your renewal date there. If you want to know what a link in a renewal email leads to before you click, paste it into a scam checker, because this one sat on a domain registered in 2019 and had no newness for you to notice.

Not sure about a renewal link? SafeBrowz checks it before the page can load. Add to Chrome, free Get the free Android app or scan a URL now →

The email

Display name "HostingerBilling". Subject line: "Your domain name will expire in 3 days, please renew". Full brand styling, the right logo, a tidy layout, and a single prominent button labelled Manage subscriptions. It arrived at our support address, which is published on our site, so there was no mystery about how it found us.

The pressure is well chosen. Losing a domain is not like missing a software renewal: your site goes dark, your email stops, and anyone who has ever had to recover a lapsed domain knows it is a bad afternoon. Three days is short enough to panic and long enough to feel real. A reader who manages one website and has not thought about its renewal date in a year is the exact target.

The button did not go to the hosting company. It went to a payment form on a subdomain of a small publisher in Delft, a genuine imprint that says it has been active since 2002, on a domain registered in 2019, publishing books on arts and sciences. That is the part worth slowing down on, because it is both why the email was convincing and why the response required care.

The DNS comparison, and what it does and does not prove

When a phishing page lives on a subdomain of a real company's domain, the useful first question is narrow: is the company's own site serving that page, or has one name been pointed somewhere else? You can answer that one from outside, and you do not need to take anyone's word for it.

The easiest way is a web-based DNS lookup, which needs nothing installed and works the same on any machine: search for a DNS lookup tool, enter the main domain, then enter the suspicious host, and compare what comes back. If you already live in a terminal, dig +short example.com does the same job on macOS and on Linux where the dnsutils or bind-utils package is installed, and on Windows it is nslookup example.com with the name on the same line, because typing nslookup on its own drops you into an interactive prompt you then have to exit.

Here is what we actually got, on 10 October and again on 14 October. The main domain and its www both answer with the name website-rendering.jouwweb.nl, which resolves to 35.204.150.5. The suspicious host answers with 167.17.76.77, which belongs to a network called Redoubt Networks in the United States. Two unrelated destinations.

Now the part that matters more than the numbers. Look up who owns 35.204.150.5 and you are told Google LLC, which is true and almost useless: the publisher has no cloud of their own. Their site runs on JouwWeb, a Dutch website builder that serves its customers from shared addresses on Google's cloud. So the IP's registered owner is two companies away from the business whose name is on the domain. That is the honest limit of "search the address": it tells you where something is hosted, not who controls it.

What did convince us was the shape of the zone rather than the ownership of the IP. Every other name under that domain, including one we invented on the spot, resolves to the website builder. The suspicious host is an explicit exception written over that default, pointing at an unrelated network, and it is signed by the domain's own DNSSEC key with a signature valid from 1 October, more than a week before the email arrived. So the record was written into the authoritative zone by somebody with access to it, not injected at a resolver along the way. That is the accurate description, and it is narrower than "DNS hijack", which people often read as something happening in transit.

So here is the honest version of the rule. A different destination tells you the main site is not serving the page, which is worth knowing and is what kept this publisher's own website out of our block list. It does not by itself tell you whether that subdomain is attacker-controlled, legitimately delegated to a third party, or a stale record left pointing at an address someone else later took over. A customerservice subdomain on a different network is exactly what an outsourced help desk looks like. The question that closes the gap is whether the subdomain has any business reason to exist: a Delft imprint publishing books on art has no reason to run a hosting company's billing page, and that, combined with the unsolicited branded email and a card form, is what made this attacker-controlled rather than merely third-party.

The reverse reading is weaker still, so do not rely on it. Phishing is hosted on the big clouds all the time, and this publisher's own site sits on one of them, so "both on the same provider" would prove very little. Behind a CDN both would show the CDN's addresses and the comparison tells you nothing at all.

Why we blocked one subdomain and not the domain

This decided what we were allowed to do. A block list entry on the apex would have marked a real publisher's entire website as dangerous, to every user we have, indefinitely, over something they did not do and probably did not know about. So the entry covers customerservice[.]cloudpressdelft[.]nl and nothing else. Their own site, their www, and any other subdomain are untouched, and we verified that boundary held after adding it rather than assuming it.

We are writing that reasoning down because it is a rule rather than a judgement call, and because the opposite mistake is common and expensive. Blocking the parent domain is the lazier action and it looks more thorough. It is also how a scanner ends up telling thousands of people that a small business in Delft is a threat.

The same restraint applies to a free host or a platform. A phishing page on a subdomain of a large hosting platform, a site builder or a shortener never justifies blocking the platform, and a page on a path under a document service never justifies blocking the service. The narrowest entry that stops the actual page is the correct one.

Where we failed on this one

Before we added that entry, our own scanner returned safe with a high score for the phishing subdomain. That is a real miss and it is worth explaining rather than skipping, because the reason generalises.

We checked our own telemetry rather than guessing, and the cause is more specific than "it inherited a good reputation". It is not reputation at all: our reputation check is deliberately apex-only, so a subdomain can never inherit a parent's standing, and this domain is not on that list anyway. What it inherited was age. When we ask how old a host is, the lookup walks up the name until a registry answers, so the phishing subdomain was scored with its parent's registration date and came back as 2,705 days old.

That one inherited number disabled two things at once. We have a guard built for exactly this situation, a page with nothing readable on it, which raises a caution when the domain is also new; it only applies under 30 days, and 2,705 is not under 30, so it never ran. And the page itself was already dark by the time we looked, 555 bytes of nothing, so the content layer had no card form, no login and no brand to judge. Two layers stood down, for the same reason, and that is the honest account: not that we were fooled by a convincing page, but that an aged parent domain made the subdomain look established and there was nothing left on it to read.

This is why an attacker prefers a subdomain on an established domain to a fresh registration of their own, and why a "was this registered recently" check is a weak defence. It is the same inheritance behind a job scam arriving from a genuine university address.

What catches this class, when the page is still up, is the page rather than the domain: a form collecting card details while wearing a hosting brand that does not own the domain under it. That is a content judgement, and it is the layer we lean on when a domain's age says nothing useful. It is also worth saying that no domain registrar or shared-hosting brand is in our brand database, Hostinger included, so there was no Hostinger keyword here for a domain-level check to match even in principle.

What a real renewal notice looks like

Two things are worth knowing about the real thing, and one of them is a useful general rule.

First, the sender name in your mail app is decorative. "HostingerBilling" is a display name, and anyone can set theirs to anything; it is not an address and it proves nothing. Tap or click it to expand the actual address and read what comes after the @. Hostinger's domain, hostinger.com, has been registered since 2002 through the company's own registrar business, and it publishes a DMARC policy of reject that explicitly covers subdomains too, so most major mailbox providers will reject a message forged from it, though enforcement is each receiver's choice rather than something the brand can guarantee. These campaigns use an unrelated domain and dress up the display name instead, which costs them nothing and works regardless of what the real brand publishes.

Second, a renewal is something you can always check from the other direction, and that is the move that makes the email irrelevant. Your domain's expiry date lives in your provider's control panel, at hpanel.hostinger.com for Hostinger, and it is also public for most domains. If you are not signed in you will be bounced to auth.hostinger.com to log in; that is also Hostinger's. For .com and most other gTLDs a WHOIS lookup on your own domain also shows the expiry without you logging in anywhere, though many country domains including .nl and .de do not publish one, so use the panel for those. If the panel says you are paid up until next August, an email claiming three days is answered.

Flag a renewal link before you pay an invoice

SafeBrowz is a browser extension and Android app that reads a page before you interact with it, in three layers. Layer 1 runs in the extension itself: URL pattern and typo-squat checks that run without a network call. Layer 2 is server-side, and it is where our 550+ brand database, our own block list and the external threat feeds live; this exact subdomain sits there now, as an entry for that one host rather than for the domain. Layer 3 is AI content analysis in over 100 languages, which reads what the page is asking for and flags a card form wearing a brand that does not match the domain under it. Worth being straight about which of those could have helped here: the host carried no brand name to match, so the page itself was the only thing left to judge.

Detection comes from threat-intelligence research and our own brand database, not from user browsing data. SafeBrowz does not store per-user browsing history.

Honest scope: we did not catch this one. It returned safe from our scanner until we read the email, ran the DNS comparison above and added the entry by hand, and the reason was an inherited registration age, not an inherited reputation. A browser extension also cannot see your inbox, so the display-name check and the expanded sender address are yours to do, and no scanner replaces them. What the product is actually for here is the moment after the click, when a page asks for a card.

The check, in order

  1. Expand the sender address. Not the display name. If it is not your provider's own domain, you are done.
  2. Do not use the button. If you want to know what it leads to, paste the link into a checker rather than opening it.
  3. Go in the other direction. New tab, your provider's control panel typed by hand, and read the renewal date there. For a .com you can also read the expiry straight from a WHOIS lookup, though many country domains such as .nl do not publish one.
  4. If the host looks like a real company's subdomain, compare the addresses. Resolve the main domain and the suspicious host. Different destinations mean the main site is not serving the page, which is worth telling the owner. It does not on its own tell you whether that subdomain is attacker-controlled, outsourced to a supplier, or just stale.
  5. Pay only from inside the panel. Never from a link in a notification, even a genuine-looking one.

Red flags in a renewal or billing email

  • A countdown on something you can verify yourself. Expiry dates are not secret. Three days is pressure, not information.
  • The payment page is not on your provider's domain. Read the host in the address bar, not the branding on the page.
  • A display name doing the work an address should do. "HostingerBilling" is typed text. The address after the @ is the fact.
  • A card form for a renewal you normally pay from the panel. Billing lives in your account, not in your inbox.
  • A host that is a brand-ish subdomain of something unrelated. A support or customerservice subdomain on a domain that has nothing to do with your provider is the shape to remember.

Chrome Add to Chrome Firefox Add to Firefox Edge Add to Edge Google Play Get it on Google Play

🛡 LIVE CHECK

Not sure about a renewal link?

Paste it here before you enter a card. Our 3-layer engine (Local + APIs + AI) reads the host, the lists, and what the page is actually asking for. The red subdomain above is live on our block list, so you can click it and see the verdict.

Full scan with deep AI analysis → · No URL is logged to your identity.

Frequently asked questions

Is the "your domain will expire in 3 days" email real?

Treat it as a scam until you have checked from the other direction. A genuine provider does show you a renewal notice, but the date is never a secret and you never have to take the email's word for it: sign in to your control panel yourself, at hpanel.hostinger.com for Hostinger, and read the expiry date there. For .com and most other gTLDs a WHOIS lookup on your own domain also shows the expiry without you logging in anywhere, though many country domains including .nl and .de do not publish one, so use the panel for those. The one we received displayed the sender name HostingerBilling, which is typed text anyone can set, and its payment button led to a subdomain of an unrelated Dutch publisher. If the payment page is not on your provider's own domain, stop.

The payment link was on a real company's website. How is that possible?

Because only one subdomain record was pointed at the attacker's server, not the company's whole site. You can check part of that yourself: resolve the main domain and the suspicious host separately and compare where each one points. In this case the main domain and its www both resolved to the website builder their site runs on, served from a shared address on Google's cloud, while the phishing subdomain resolved to an address belonging to Redoubt Networks in the United States. Two unrelated destinations, which shows the company's real site was not serving the page. On its own that does not prove who controls the subdomain, since an outsourced supplier or a stale record would also point elsewhere; what settled it here was that a Delft book imprint has no business reason to host a hosting company's billing page.

Should the whole domain be blocked then?

No, and this is the part that matters if you report it. Blocking the parent domain would mark a genuine business's entire website as dangerous over something they did not do, so the correct action is an entry for the exact attacker-controlled host and nothing else. We blocked that one subdomain and left the apex, the www and every other subdomain untouched, then verified the boundary held rather than assuming it. The same restraint applies to a phishing page on a hosting platform, a site builder or a document service: the platform is not the threat.

Why did your scanner say the phishing page was safe at first?

Not because of reputation: our reputation check only ever applies to a domain itself, never to a subdomain beneath it, and this domain is not on that list anyway. What the subdomain inherited was age. The publisher's domain was seven years old with a clean history and a real site on it, so the phishing subdomain arrived pre-trusted without having earned any of that. That inheritance is exactly why attackers prefer a subdomain on an aged domain to a fresh registration, and it is why a "was this registered recently" check is a weak defence on its own. What should catch it is the page content, a card form wearing a brand that does not own the domain under it, which is the layer we rely on when a domain's age says nothing useful.

I already paid on one of these pages. What should I do?

Call your card issuer now, using the number on the back of the card, and tell them the details were entered on a phishing page so they can block it and reissue. Do that first, because speed decides whether a charge can be reversed. Then sign in to your real hosting account directly and check the renewal status, the billing contact and any recently added users or API keys, and change the account password if you entered it anywhere. Watch for a small test charge before a large one. Report it to the FTC at reportfraud.ftc.gov, or, in the UK, Report Fraud at reportfraud.police.uk, which replaced Action Fraud in December 2025, and forward the email to your provider's abuse address so they can pursue the takedown.

Follow on Google