Blog / Web management /

HSTS and HTTPS preload: one step beyond SSL

Your site has HTTPS. Great. But someone can still type the old HTTP address, or follow an ancient link.

A redirect usually sends them to the secure version. Usually. But for a moment, before that redirect, the browser asks for the insecure page. HSTS tells the browser to use HTTPS on future visits. Preloading applies that rule before the browser has ever visited. Handy when it’s set up carefully. Painful when it isn’t.

HSTS is a security policy delivered in an HTTP response header. It doesn’t replace a valid certificate, a working HTTPS site or a redirect plan. The preload list is a separate commitment: browsers that ship it know in advance that a domain should only be reached over HTTPS.

Understand the header

A basic HSTS header has a max-age value telling the browser how long to remember the policy. includeSubDomains extends it to subdomains. The preload token signals that the domain intends to meet the requirements for browser preload lists. Adding that word doesn’t automatically put the domain on any list.

Once a browser stores an HSTS policy, it tries HTTPS for that domain throughout the policy period. That helps against certain downgrade and interception scenarios. It also means a mistake can lock people out of a hostname until the policy expires or you fix the issue. A certificate problem, an HTTP-only subdomain or an overlooked service all get more serious under a strict policy.

Inventory before enabling it

Start with a list of every hostname the organisation controls: the main site, www, subdomains, staging, customer portals, old campaign hosts and any service still using the domain. Check which should work over HTTPS and who maintains them. I don’t turn on includeSubDomains just because a security guide says so. It affects every covered hostname, including ones the team forgot.

Confirm certificates are valid and renewed, HTTP requests redirect cleanly, and HTTPS works for pages, forms, APIs and embedded resources. Hunt for mixed content: a secure page still requesting scripts, images or styles over HTTP. Check links and canonical tags point to the preferred HTTPS URL, and update sitemaps and key external integrations.

Test on a staging hostname where possible, but staging may not represent all your production subdomains. Coordinate with whoever owns DNS, hosting and certificate renewal, and keep a rollback and communications plan. Removing an HSTS policy isn’t instant for browsers that have cached a long max-age.

Roll out in stages

A cautious rollout starts with a short policy on the primary domain and no subdomain directive, then lengthens the duration once you’ve confirmed HTTPS works. The technical owner should choose exact values and sequence for the environment. I wouldn’t paste a long max-age from a tutorial without understanding what it does. Test with browser developer tools and an external header checker, then verify the site from a clean browser profile.

Consider includeSubDomains only after the inventory is complete and every covered host is checked. Preload is a bigger step again. Read the current submission requirements and understand the removal process before applying. The list takes time to update across browser releases, and a removal request doesn’t instantly clear the policy from devices using a bundled list.

If a host isn’t ready, fix it first or leave the stronger directive off. A partial policy that properly protects the main site is safer than a broad promise the organisation can’t keep.

SEO and operations

HTTPS is expected on modern sites, and secure URLs avoid browser warnings and mixed-protocol confusion. Treat HSTS as a security control, not a quick ranking tactic. Search engines need consistent redirects, canonicals and internal links to consolidate URL versions, and a header can’t repair a wrong canonical or a duplicated site architecture.

Document the setting, the hostnames it covers, who renews certificates and how the team will test future subdomains. Put it in migration and onboarding notes, so a new developer doesn’t discover the policy after a service fails. The security benefit is real, but it depends on a maintained HTTPS setup.

Plan for every hostname, not only the website

A domain may host email records, a support portal, a payment flow, a staging environment or an API. Some services are run by vendors with their own certificate management. Before applying a domain-wide policy, map each host to its owner and confirm the vendor supports HTTPS in the way you need. A subdomain gets forgotten precisely because it isn’t part of the public website.

Coordinate with the people who handle email and infrastructure. HSTS affects browser access, but the wider domain setup may involve other services and renewal responsibilities. Watch logs and customer reports after the change, and make sure someone is on call for a hostname that fails. Record the header value and the date you introduced it, so the next maintainer understands the policy instead of changing it casually.

Treat preload as a commitment, not a badge

Preload makes the HTTPS-only instruction work before a browser has seen your header, but it also makes the change harder to reverse quickly. Inclusion means meeting the requirements and submitting through the proper process. The header token alone doesn’t complete it. Read the current instructions before requesting inclusion, because requirements and browser handling evolve.

Before setting a long max-age or enabling subdomains, confirm every hostname in scope is ready for HTTPS, including services other teams manage. A forgotten subdomain may belong to a vendor or legacy system. Ask the owner to verify it. Don’t assume the main website certificate covers everything. If ownership is unclear, sort that out before extending the policy.

Before you enable preload, list every subdomain and verify its HTTPS setup. Which hostname might surprise you?