When an SEO says "sitemap", they almost always mean the XML feed at /sitemap.xml that lists every page on your site for search engines. They almost never mean the planning tool used by web designers to lay out a site's navigation. Two different things, same word. A lot of older SEO writing confuses them.
This guide is about the first one. The XML sitemap. What it does, when you actually need one in 2026, and how to build one that Google and the AI crawlers will respect.
The fundamentals haven't changed. A sitemap is still a list of URLs you want indexed, the search engine reads it, and pages get discovered faster than they would by crawling alone. What has changed is everything around it. Google now treats the <lastmod> timestamp as a binary trust signal. Two of the four optional XML tags have been quietly killed off. IndexNow has split the discovery game between Google and the rest. And a new generation of AI crawlers (GPTBot, PerplexityBot, ClaudeBot) is using your sitemap as a primary discovery channel without ever showing up in your Search Console reports.
If your last mental model of XML sitemaps was set in 2019, this is the catch-up.
Do you actually need an XML sitemap?
Honest answer: probably yes, but not always.
Google's own Search Central documentation lists four conditions that make a sitemap genuinely load-bearing. Your site is large (1,000+ URLs). Your site has pages that aren't well linked internally. Your site is new and hasn't picked up external links yet. Your site has lots of media (image, video, news) you want indexed separately.
If none of those apply, you can get away without one. A 25-page brochure site for a local plumber with sensible internal links will be crawled and indexed perfectly well by Google without an XML sitemap. The official guidance says as much.
For everyone else, the calculus is simple. The cost of producing a sitemap in 2026 is close to zero (every modern CMS and framework ships one) and the cost of not having one on a site that needs one is invisible: pages just don't get crawled, or get crawled weeks late. If you're running e-commerce, news, multi-region, or anything content-heavy, treat the sitemap as mandatory.
XML sitemap vs HTML sitemap vs information-architecture sitemap
Three different artefacts, all called "sitemap". The disambiguation up front saves a lot of confusion later.
The XML sitemap is a machine-readable file at a URL like /sitemap.xml. It exists for search engines. Nobody reads it directly. This is the one this article is about.
The HTML sitemap is a regular web page listing your site's links, usually in the footer of older sites. Built for humans, scraped incidentally by search engines. In 2026 it's mostly obsolete, except as a belt-and-braces internal-linking aid on very large sites.
The information-architecture sitemap isn't a live file at all. It's a planning diagram (in tools like Slickplan, Octopus, or just a whiteboard) used by designers and developers to lay out a site's structure before building it. Useful in the design phase, irrelevant to SEO.
The first one is what search engines want. The other two are about humans and process.
What's actually changed since 2023
Most of the XML sitemap content still on page one of Google was written before 2023 and hasn't caught up with what Google has done to the spec since.
The lastmod trap
Google's current guidance is explicit: it will use the <lastmod> value "if it's consistently and verifiably accurate." If you bump <lastmod> on every page after a minor template change or a republish-button click, Google eventually stops trusting the timestamps entirely. Site-wide. There's no per-page granularity to it.
<lastmod> is now binary. Either Google trusts your timestamps and uses them to prioritise crawl, or it doesn't and falls back to its own heuristics. A 2024 HTTP Archive study found that 58% of sitemaps had stale or missing <lastmod> values, which means the majority of sitemaps in the wild aren't even doing the one thing the tag is supposed to do.
<priority> and <changefreq> are dead tags. Google has confirmed publicly it ignores both. Bing barely uses them either. Stop generating them. Stop populating them. The CMS plugin that fills them in is wasting CPU.
IndexNow has split the discovery stack. Bing, Yandex, Naver, and Seznam all support IndexNow, a push protocol where you ping the engine when a page changes instead of waiting for it to crawl your sitemap. Google doesn't support IndexNow. So the modern setup is XML sitemap for Google, IndexNow for everyone else, both running in parallel.
AI crawlers are using sitemaps too. GPTBot (OpenAI), OAI-SearchBot, ClaudeBot (Anthropic), PerplexityBot, and Bingbot for Copilot all read sitemaps as a discovery channel. None of them give you Search Console-style reports back. You build for them anyway, because the alternative is your content being invisible to AI search.
What goes in (and what doesn't)
A sitemap should be a curated list of the pages you want indexed. Not a dump of every URL your CMS can generate.
Include canonical URLs that return a 200 status code and are marked as indexable. Each URL should appear in exactly one sitemap. If you have a URL in your sitemap that's noindexed, blocked in robots.txt, redirects, or returns a 404, you're sending Google a contradictory signal and the sitemap loses credibility.
The minimum viable URL entry is two tags: <loc> and <lastmod>. That's it.
<?xml version="1.0" encoding="UTF-8"?> <urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9"> <url> <loc>https://example.com/services/seo</loc> <lastmod>2026-05-12</lastmod> </url> <url> <loc>https://example.com/blog/keyword-research-2026</loc> <lastmod>2026-05-16</lastmod> </url> </urlset>
Four common mistakes worth checking for on any existing sitemap:
- Noindexed pages in the sitemap. They contradict each other.
- Non-canonical URLs (tracking parameters, session IDs, faceted-search variants).
- Dead URLs (redirects or 404s). Sitemaps shouldn't contain URLs that don't resolve.
- URLs blocked in
robots.txt. Sitemap says "index this", robots.txt says "don't crawl this", Google trusts robots.txt.
The single most useful tag is <lastmod>. The others (<priority>, <changefreq>) are decorative at this point.
Specialised sitemap types most sites overlook
The base XML sitemap covers regular pages. For everything else, there are specialised formats that surface content into specific Google indices.
Image sitemaps declare your images explicitly. Useful for e-commerce, portfolios, news, anywhere image search drives meaningful traffic. Without one, Google still finds your images, but slower.
Video sitemaps are effectively required if you want videos to appear in video SERP features. They include duration, thumbnail URL, and content URL per video.
News sitemaps apply only to publications approved for Google News. They have a strict spec: only articles published in the last 48 hours, with publication metadata in each entry.
Hreflang in sitemaps is the cleanest way to declare alternate-language and alternate-region pairs on multilingual or multi-region sites. Cleaner than rel="alternate" link tags on every page, easier to audit, and the canonical recommendation in Google's own docs.
Sitemap index files matter the moment you cross the size limit. A single XML sitemap is capped at 50,000 URLs or 50MB uncompressed. Above that, you split into multiple sitemaps and reference them from a sitemap_index.xml. A single index file can reference up to 50,000 individual sitemaps, so the protocol scales to roughly 2.5 billion URLs without breaking.
Below the limit, sitemap index files are still useful for segmenting by content type (sitemap-products.xml, sitemap-blog.xml, sitemap-pages.xml). Easier to audit, easier to diagnose when something goes wrong.
Building and submitting a sitemap in 2026
The workflow is mature. Five steps from nothing to a live, monitored sitemap.
Use your CMS plugin (Yoast, Rank Math, All in One SEO for WordPress) or your framework's built-in feature (Next.js, Astro, Hugo all support this natively). Avoid free online sitemap generators for sites that update frequently. They produce a static file that goes stale within hours.
Run Screaming Frog or Sitebulb against the sitemap to confirm every URL returns 200, is canonical, and isn't noindexed. The first sitemap audit on a site usually surfaces a dozen contradictions you didn't know were there.
Open Search Console, navigate to Indexing → Sitemaps, paste in the sitemap URL, submit. Then add a `Sitemap:` line to your robots.txt file so other crawlers find it without being told.
Most modern CMS plugins ship an IndexNow integration. Turn it on. Bing and Yandex will index changes within hours instead of days, and IndexNow's push model is more reliable than waiting for sitemap recrawls.
Search Console's Sitemaps report shows submitted vs. indexed counts per sitemap. If submitted is 500 and indexed is 80, the sitemap isn't the problem. The URLs in it are. Audit content quality, duplication, and crawl budget before re-submitting.
# Tell every crawler where to find the sitemap User-agent: * Allow: / Sitemap: https://example.com/sitemap.xml Sitemap: https://example.com/sitemap-news.xml
The whole workflow takes about an hour for a small site, half a day for a large one. The maintenance burden after that is close to zero if the CMS or framework is regenerating the file automatically.
When Google ignores your sitemap
You submit a sitemap and nothing happens. Search Console shows "Couldn't fetch", or it shows the sitemap was processed but the indexed count is far below the submitted count. Four diagnostics, in this order:
First, fetch the sitemap URL yourself. If it returns a 404, the wrong content-type, or a redirect chain, fix that before anything else. Google won't index what it can't fetch.
Second, check the URLs inside it. Spot-check 10 random URLs in incognito. Are they all 200? Are they all canonical? Are any of them noindexed? Any single one of those problems drags the whole sitemap's reputation down.
Third, look at the <lastmod> values. If most of them are today's date despite no content changes, Google has probably stopped trusting them. The fix is unglamorous: update your CMS or template to only bump <lastmod> when real content changes (main body copy, structured data, or links). Then wait. Trust takes weeks to rebuild.
Fourth, check robots.txt for accidental crawl blocks on the URLs in the sitemap. Sitemap says "please index"; robots.txt says "don't crawl". Robots wins. The pages stay invisible.
Most "Google is ignoring my sitemap" problems are one of these four. Genuinely opaque sitemap failures are rare.
The minimum sitemap hygiene checklist
A monthly five-minute check. Run it. Most sitemap problems compound silently otherwise.
- ✓Sitemap lives at a discoverable URL (typically /sitemap.xml or /sitemap_index.xml)
- ✓Listed in robots.txt with a Sitemap: line
- ✓Submitted in Google Search Console
- Every URL returns 200, is canonical, and is indexable
- No duplicate URLs across multiple sitemaps
- <lastmod> reflects real content updates, not republish-button presses
- Under 50,000 URLs / 50MB per file (sitemap index if over)
- IndexNow enabled (CMS plugin or server-side) for Bing/Yandex/Naver
- Specialised feeds (image / video / news / hreflang) where they apply
A sitemap won't lift your rankings. What it will do is quietly cap how much of your site Google bothers to crawl, and how quickly AI search models discover your new content, if you let it drift. The work is unglamorous and the upside is silent compliance with how search engines actually work in 2026.
FAQ
Do small websites need an XML sitemap?
Probably not, if you have fewer than about 500 pages and your internal linking is solid. Google's own guidance is that small, well-linked sites can be discovered without one. That said, the cost of generating a sitemap with a modern CMS is essentially zero, so most small sites still produce one as a low-cost backstop.
Does a sitemap improve my Google rankings?
No. A sitemap is a discovery aid, not a ranking signal. What it does is help Google find and re-crawl your pages faster, which means new content gets into the index sooner and updated content gets refreshed faster. Indirectly that affects visibility, but the sitemap itself isn't a ranking factor.
What's the difference between an XML sitemap and an HTML sitemap?
The XML sitemap is for search engines. It lives at a fixed URL like /sitemap.xml, is machine-readable, and nobody browses to it directly. The HTML sitemap is a regular web page listing your site's links for human visitors, usually buried in the footer. In 2026 the HTML version is mostly obsolete for SEO and only useful as an internal-linking aid on very large sites.
How often should I update my sitemap?
Automatically, every time a page is published, updated, or removed. Modern CMS plugins and frameworks do this for you. The only thing you should never do is bump
Is IndexNow replacing XML sitemaps?
No. IndexNow is a push protocol that Bing, Yandex, Naver, and Seznam support. It's complementary to sitemaps, not a replacement. Google does not support IndexNow at all, so the XML sitemap remains the only way to tell Google about your URLs at scale. The modern stack runs both in parallel.
Should I include noindexed pages in my sitemap?
Never. The sitemap is for pages you want indexed. Including a noindexed page sends Google a contradictory signal: 'index this / don't index this'. Repeated contradictions erode the sitemap's credibility, which can lead Google to deprioritise the whole file. Audit your sitemap once a quarter for these.
Technical SEO audit (free 7 days)
We'll surface your top 5 highest-impact technical SEO fixes (sitemap, internal linking, crawl issues, schema, the lot) in a week.
Book an audit
