To detect parked domains and dead sites in a lead list, run three checks on each website: DNS (does the domain resolve, and to whom?), HTTP (what status code, which redirects, what final URL?) and content (does the page read like a business or like a parking, expiry or hosting default page?). Record the outcome for each site rather than deleting it, because a business that has lost its website is junk for email extraction but often a strong lead for web designers.
This is part of the scraping troubleshooting guide. Dead sites are not blocks, but they fail in similar ways and are easy to mistake for them.
Signals table
| State | DNS | HTTP | Content | Treat as |
|---|---|---|---|---|
| Live business site | Resolves | 200 | Business name, services, contact details | Crawl normally |
| Parked by registrar or parking service | Resolves, often to parking name servers | 200 | "This domain may be for sale", ads, few words, generic template | Parked |
| Listed for sale | Resolves | Redirect to a domain marketplace | Price or "make an offer" | Parked (for sale) |
| Expired | Resolves to registrar, or stops resolving | 200 or no response | Renewal or expiry notice | Expired |
| DNS dead | NXDOMAIN or SERVFAIL | None | None | Dead |
| Server down | Resolves | Connection refused, timeout or 5xx | None | Dead or temporarily down |
| Hosting default page | Resolves | 200 | "Welcome to nginx", "Apache2 Default Page", control panel placeholder | No site |
| Hosting suspended | Resolves | 200, 403 or 503 | "Account suspended" | No site |
| Builder placeholder | Resolves | 200 | "Coming soon", "Under construction", builder branding | No site |
| Homepage missing | Resolves | 404 or 410 | Error page | Dead |
| Redirect to another business | Resolves | 301 or 308 to a different registrable domain | A different company | Moved or sold; check |
| Redirect to a social page | Resolves | 301 or 302 to a social platform | Profile page | No own site |
| Certificate error | Resolves | TLS fails before HTTP | Browser warning | Broken; often lapsed hosting |
| Bot challenge | Resolves | 403, 429 or 503 with a challenge page | "Checking your browser" | Unknown, not dead |
The last row matters most. A challenge page from a security service means the site is alive and protecting itself. Treat it as unknown and do not retry aggressively; see web scraping blocked: troubleshooting.
DNS checks
- A and AAAA records. No address at all (
NXDOMAIN) means the domain does not resolve: lapsed, never set up, or mistyped. - NS records. Name servers belonging to a parking service or a registrar's default set are a strong parking signal, especially if the site itself is thin.
- MX records. Record them separately. A domain can have a dead website and working email, or the reverse. A domain that publishes a "null MX" record is saying it accepts no email at all; the free MX lookup checks a whole list of domains.
HTTP checks
Status codes are defined in RFC 9110. For dead site detection:
- 200 OK says nothing about whether the site is real. Parking pages, default pages and placeholders all return 200.
- 404 Not Found on the homepage means the server is up but the site is gone.
- 410 Gone indicates access is "no longer available" and that this is likely permanent.
- 5xx errors and 503 Service Unavailable can be temporary. Check again later before calling the site dead.
- 403 Forbidden means the server "understood the request but refuses to fulfill it". On its own it is a block, not proof of a dead site.
Use a short timeout (around 10 to 15 seconds) and treat a timeout as "unknown, retry later" the first time.
Redirect chains
Follow redirects and record every hop, not just the final page.
- Record each status code and
Location. RFC 9110 defines 301 and 308 as permanent ("assigned a new permanent URI") and 302 and 307 as temporary. - Compare the registrable domain of the final URL with the original.
brightsmile.exampletowww.brightsmile.exampleis the same site.brightsmile.exampletosmilegroup.examplemeans the business moved, rebranded or was bought. - Cap the chain. Ten hops is plenty; loops and very long chains are a fault, not a site.
- Look for redirects in the page as well: a meta refresh tag or a script that sends the visitor elsewhere.
- Flag redirects to unrelated content, such as gambling or adult sites. That usually means the domain lapsed and was bought by someone who wants its old links. Do not contact or link to it.
Content checks
Most parked and placeholder pages give themselves away:
- Very little visible text, often under a hundred words.
- Phrases such as "domain may be for sale", "buy this domain", "parked free", "renew now", "account suspended", "website coming soon", "default web page".
- The page title is just the domain name.
- Most links point to one advertising or marketplace domain.
- The same HTML template appears on several unrelated domains in your list.
- No business name, address or phone that matches the listing.
Combine signals. Thin content alone also describes a one-page site for a sole trader; thin content plus parking name servers plus a "for sale" phrase is conclusive.
Running the checks politely in bulk
Checking hundreds of domains is light work compared with a full crawl, but still be considerate:
- Fetch only the homepage on the first pass.
- Run a small number of checks in parallel and spread requests across different hosts.
- Respect
robots.txtif you go beyond the homepage. RFC 9309 describes how crawlers should read it. - Do not retry a host that refused you straight away; queue it for later.
- Keep the results: status, final URL, number of redirects, title, word count and the state you assigned.
False positives to avoid
- Script-built sites. Some sites send an almost empty HTML shell and build the page with JavaScript. A plain HTTP check sees no text. Render the page in a browser before calling it parked; see why a scraper gets blocked or returns nothing.
- Geo-blocking. Some sites refuse visitors from other countries. Check from the country the business serves.
- Maintenance pages. A "back soon" page with the business's own branding is a live business having a bad day.
- Single-page sites. Short is not the same as parked if the page names the business and its contact details.
Using them as leads
For most campaigns, parked and dead sites are noise: there is nothing to crawl, and emails at a lapsed domain bounce. For some sellers they are the list:
- Web designers and developers. A business that once had a website and lost it has already shown it values one. "Your website at kirkgate-plumbing.example has been showing a hosting suspension page" is a specific, useful opener.
- IT support and hosting providers. Expired certificates, suspended hosting and broken email all point to an owner without technical help.
- Marketing agencies. A Maps listing that sends customers to a dead page is losing them every day.
Before contacting:
- Confirm the business is still trading from the map listing: recent reviews, an open status, a working phone.
- Do not email addresses at a domain that has expired or changed hands. They will bounce or reach the new owner. Use the phone or post instead.
- Check whether the business is a company or a sole trader before choosing email; UK rules differ. The channel and law section of how to find businesses without a website covers this, and the same applies here.
How Map Extract helps
- DNS, HTTP and content checks on every site. The website data extractor checks DNS and reachability, then records the HTTP status, whether the site redirected and the effective final URL, whether it is parked, blocked or showing a CAPTCHA, and whether its certificate is valid and when it expires.
- Recorded, not mistaken for data. Parked, dead and blocked sites are stored as outcomes with a status, and are quarantined so they are not crawled again and again. Only you can choose to retry a quarantined domain.
- Filter them into a list. Because the outcome is a field on the record, you can export businesses with a parked or dead website as their own list, alongside their map listing details.
- Rendered fallback. Script-only sites are given a rendered browser pass before being judged, which reduces false "empty site" results.
- Run control. The run centre shows progress and items that need attention, and lets you pause and resume runs.
What it does not do: its parked and dead labels are signals to review, not certainties, and it does not tell you who now owns a lapsed domain.
Public business and social results the app collects are uploaded automatically to Map Extract's servers while it is licensed and running; the privacy notice explains this.
Key takeaways
- Check DNS, HTTP and content together; a 200 status proves nothing on its own.
- Follow and record redirect chains, and compare the final registrable domain.
- Treat bot challenges, timeouts and script-only shells as unknown, not dead.
- Keep dead and parked outcomes on the record instead of deleting the business.
- For web designers and IT providers, a lost website is a lead; never email a domain that has changed hands.
Frequently asked questions
How can you tell if a domain is parked?
A parked domain usually resolves and returns HTTP 200, but the page has little real text, phrases such as 'this domain may be for sale' or 'domain parking', links to advertising or a domain marketplace, and the same template as many other parked domains. Its name servers often belong to a registrar or parking service. Check DNS, the HTTP response and the page content together, because no single signal is conclusive.
How do I check if a list of websites is still live?
For each domain, check that DNS resolves, request the homepage with a short timeout and record the status code, every redirect and the final URL, then read the page title and text for parking, hosting default or suspension phrases. Run the checks slowly, a few at a time, and mark uncertain results such as timeouts or bot challenges as unknown rather than dead.
Should I remove parked domains from a lead list?
Remove them from website-based steps such as email extraction, because there is nothing to extract, and do not email addresses at a domain that has expired or changed hands. Keep the business itself if the map listing shows it is still trading. For web designers and IT providers, a business that has lost its website is one of the best leads on the list.
Can a business with a dead website still receive email?
Sometimes. Websites and email are configured separately, so a domain whose website is down may still have working MX records and mailboxes. If the domain has expired, been parked by the registrar or sold, email will usually fail or, worse, reach a new owner. Check the MX records and verify the address before sending, and never email a domain that now belongs to someone else.
Sources
Written by the Map Extract editorial team. Found an error? Email [email protected] and we will check and correct it. How we research guides: editorial policy.