The decisions that are expensive to reverse
Most new-site SEO problems are not things that went wrong later. They are choices made in the first week that became costly to change once content existed.
Four worth getting right before you publish anything:
Domain. An exact-match keyword domain confers no advantage worth having and constrains you if the business changes. A brandable domain is almost always the better choice. Whatever you pick, decide www or non-www and be consistent from the first day.
URL structure. Readable, lowercase, hyphenated, shallow. Avoid dates in URLs — they age content visibly and complicate updating it later. Changing this after publishing means redirecting every URL.
Platform. Whatever your team can actually maintain. The SEO differences between mainstream platforms are small next to the difference between a site someone updates and one nobody touches.
Site structure. Which sections exist, what lives under what. This is the one people most regret, because restructuring after two years of content is a migration.
Before launch
Block the staging site properly. HTTP authentication, not robots.txt and not noindex. A blocked robots.txt does not prevent indexing — a URL can be indexed without ever being fetched. A password can.
Set up analytics and Search Console before you need them. Verify the property, and do it early — Search Console only reports from verification onwards, and there is no backfill. Verifying on launch day means your first weeks have no data at all.
Decide what is indexable. Tag archives, author pages, internal search results, thank-you pages, paginated series. Decide deliberately rather than accepting platform defaults.
Write the titles and meta descriptions. Not "optimise them later". Templated titles are what most new sites launch with and what most never revisit.
Check the basics render. One <h1> per page, images with alt text, a mobile layout that works, HTTPS everywhere with no mixed content.
Launch day
Short list, and it is the same short list that goes wrong every time.
- Remove the staging block. The single most common launch failure is a site launched still carrying
Disallow: /or a sitewidenoindex. Check the live site, not the config file. - Confirm the live site is crawlable. Load
/robots.txtin a browser. Check the homepage source fornoindex. - Submit the sitemap in Search Console.
- Crawl your own site and check for broken links, redirect chains, missing titles.
- Request indexing for the handful of pages that matter most. Not every page — for a new site, the homepage and your main service or product pages.
If you are replacing an existing site rather than launching a brand new one, stop here and read the migration guide instead. Redirects change everything about this sequence.
What actually matters in month one
New sites have no history, no links and no track record. That is a real constraint and no amount of on-page work removes it.
What helps:
- Publishing pages that answer specific questions, rather than a thin page for every keyword
- Internal links, so what you do publish is connected rather than orphaned
- Getting a few genuine external links — listings, partners, suppliers, anyone with a real reason to mention you
- Being findable for your own brand name, which is the easiest thing to win and the first thing people search
What does not help:
- Checking rankings daily. They will not be there yet, and watching costs attention that publishing would use better
- Buying links, which is the fastest way to turn a slow start into a penalty
- Publishing volume for its own sake. Twenty thin pages perform worse than five good ones and are harder to fix later
How long before anything happens
There is no honest fixed answer, and anyone giving you one in weeks is guessing.
What is reliable is the *sequence*: indexing comes first, then impressions for long-tail and brand terms, then positions for terms with real competition. Each step depends on the one before, and the last is slowest because it depends on accumulating signals a new domain has not had time to earn.
What you can control is whether the site is crawlable, whether pages answer something specific, and whether anything links to them. Those are the inputs. The timeline is not one of them.
Frequently asked questions
How long does SEO take for a new website?
There is no reliable fixed figure — it depends on competition, content quality and how quickly the site earns external links. The order is dependable even when the timing is not: indexing, then long-tail impressions, then competitive positions.
Do I need to submit my site to Google?
Not usually — Google finds sites through links. Verifying Search Console and submitting a sitemap speeds up discovery and, more importantly, gives you the reporting you will need. Paid submission services are unnecessary.
Should a new site block search engines before launch?
Block staging with HTTP authentication, not robots.txt. A URL blocked in robots.txt can still be indexed if something links to it, because Google can list a URL it has never fetched. A password cannot be crawled around.
What is the most common new-site SEO mistake?
Launching with the staging block still in place — a sitewide noindex or Disallow: / carried into production. It is easy to miss because everything looks normal to visitors, and the fix is instant once someone checks.
Launch it set up correctly
The expensive new-site mistakes are structural and early: a URL structure that has to be redirected later, Search Console verified too late to have baseline data, or a staging block that survives launch.
The New Site SEO Launch Setup covers the configuration, the indexing decisions and the launch checks in one pass — $197, one-off, no retainer.