SEO for a new website: a 90-day plan, week by week

Short answer: SEO for a new website is three jobs, in this order. In weeks 1–2, make the site easy to crawl and set up Search Console. In weeks 3–6, choose searches you can win, build the pages that answer them, and earn your first relevant links. From week 7, publish at a steady pace and run a weekly Search Console loop: fix what isn't indexed, then rewrite titles on pages that get impressions but few clicks. Expect pages to be crawled within days to weeks, and real search traffic to take months.

A new website has no history, no links and no data. Most SEO advice assumes you have all three. So founders either copy a plan meant for an established site, or do nothing and hope.

This is the plan we'd follow if we started a domain tomorrow. It's based on what we did on Versely, our own product, including the part we got wrong.

Your new website SEO checklist, week by week

Set expectations first

Google's SEO starter guide says "some changes might take effect in a few hours, others could take several months," and that "you likely want to wait a few weeks to assess whether your work had beneficial effects."

Crawling is the quicker part. Google says that crawling a requested URL "can take anywhere from a few days to a few weeks," and that asking again for the same URL won't make it faster. Ranking is the slower part.

That's the honest frame. In 90 days, a new website can realistically get:

• Every important page crawled and indexed.
• A first set of relevant links and brand listings.
• Impressions for long, specific queries, and a few clicks.
• Enough Search Console data to know what to do next.

What it usually won't get is meaningful traffic on competitive head terms. Plan for that, and you won't quit in week 8.

For context: Versely was flat at about 2 clicks a day until around August 10, 2026, then climbed to roughly 60–90 clicks a day. That's six flat weeks at the start of the three months we measure.

Weeks 1–2: the technical base

Nothing else works if Google can't crawl and understand your pages. This is a one-time job, and it's small on a new site. The full list is in our technical SEO checklist.

Search Console and Bing. Verify a domain property in Google Search Console: it covers every subdomain and protocol, and you verify it with a DNS record. Then import it into Bing Webmaster Tools, which verifies the site and brings your sitemaps across. Our $0 tool stack covers both, and Search Console for founders shows which reports to open first.

A sitemap with honest dates. Generate it from the same source as your pages, so it's never stale. Google's sitemap overview names new sites: "Googlebot might not discover your pages if no other sites link to them." A sitemap helps discovery but doesn't guarantee crawling or indexing. Submit it, then check it with our sitemap checker.

Metadata on every page. A unique title and description, a canonical URL, and no stray noindex. Run your key pages through the meta tag checker, or your homepage through the website SEO checker.

robots.txt that lets in the right crawlers. Including the AI search crawlers you want to be cited by. Use the AI crawler checker to confirm.

Basic structured data. Organization on the homepage, SoftwareApplication or Product on product pages, Article on posts, BreadcrumbList everywhere. It won't rank you by itself. It helps search engines read what each page is.

A first indexing check. Google says it primarily finds pages through links from pages it has already crawled, and suggests searching site:yourdomain.com to see what's in. That search is a quick look, not proof: Google says the site: operator "doesn't necessarily return all the URLs that are indexed." For one page, use the URL Inspection tool. Our free Google index checker combines a live site: search with noindex, robots.txt and canonical checks. If pages are still missing after a few weeks, work through how to get Google to index your site.

That's the same base we built on Versely: a metadata and canonical helper, JSON-LD, a sitemap index with honest lastmod, a generated llms.txt, and AI-crawler tracking. None of it is exotic.

Weeks 3–6: searches, core pages and first links

Choose searches a new website can win

A new site rarely beats established brands on broad terms, so start with specific searches where the results are weak. For each search you're considering, check three things:

1. Intent. What kind of page ranks: guides, tools, product pages, forum threads? If every result is a tool, a blog post won't compete. See search intent.
2. Who ranks. If page one is all large brands, pick a narrower search. Small sites, forum threads and thin pages in the results are a better sign. This is a rule of thumb, not a Google rule.
3. Difficulty. Keyword difficulty is an estimate that SEO tool vendors calculate in their own ways. It is not a Google metric. Use it to compare searches with each other. Our free keyword difficulty checker scores up to 20 keywords at once and shows monthly searches and intent.

Start with three to five searches your buyers really type. Sales calls and support questions are the best source. Once you have impressions, Search Console's query report shows more, and long-tail searches are often where a new site gets its first clicks.

Core pages

Before any blog post, make sure these exist and are good:

1. Homepage that says what you do in one sentence.
2. One page per main use case or audience. "For agencies," "for podcasters." Only if you'd genuinely say something different on each.
3. Pricing. Reviewers, directories and AI assistants all look for it.
4. Comparison and alternatives pages for the two or three products people compare you with. These match real search intent, and they're where buyers are close to a decision.

Each page should answer the query it targets in its first paragraph. Google's guide to optimizing for generative AI features calls this kind of work "non-commodity content": pages with "unique expert or experienced takes that go beyond common knowledge." As a founder, you have that. Use it. Our on-page SEO checklist covers titles, headings and the rest, page by page, and internal linking for small sites shows how to tie the pages together.

First links

A new domain needs a few relevant links so search engines discover it and have some reason to trust it. The honest ways to get them early:

• A launch. Launch free on LaunchRanked for a permanent product page and a followed link while our badge is on your site. Add one or two other platforms. Our Product Hunt alternatives roundup compares them.
• Relevant directories. 30 to 100 reviewed, niche-matched directories. Our directory list shows which links are followed and how each one reviews. The reasoning is in directory submissions in 2026.
• One free tool or resource. A calculator, checker or template people in your niche will link to. It keeps earning links long after launch.

For the full order, from launch week to unlinked mentions, see how to get backlinks for a new startup.

Avoid bulk link packages and link exchanges. Google's spam policies list "buying or selling links for ranking purposes" and "excessive link exchanges" as link spam. See is backlink exchange safe? for the detail.

Weeks 7–12: publish at a pace, then run the loop

Pick a pace you can keep

Start with a small number of good articles per week. Each one should answer a question your buyers actually ask, link to at least three relevant pages on your site, and cite its facts.

Then earn speed. This is the lesson we paid for: we published about 2,500 posts to Versely in one month, and Google crawled many of them but didn't index them. We now publish about 20 posts a day on Versely, a domain with history. A brand-new domain should start far lower. The full story is here.

The rule we use: raise the pace only while new pages are getting indexed. The moment "Crawled – currently not indexed" starts climbing in the Page indexing report, slow down and improve what you have. Our guides to Crawled – currently not indexed and Discovered – currently not indexed cover what to do.

Run a weekly Search Console loop

Once a week, 30 minutes:

1. Page indexing. Are new pages getting indexed? If not, are they thin, too similar to another page, or poorly linked? Merge, improve or noindex.
2. Performance, by page. Sort by impressions. Pages with impressions but almost no clicks are your best opportunity.
3. Performance, by query. Which queries do you appear for that you didn't target? Some deserve their own page. Most deserve a better paragraph on an existing one.
4. Links. Which new sites link to you? Which pages attract links?
5. AI answers. Once a few pages are indexed, check whether ChatGPT, Claude, Perplexity and Gemini name you for your buyers' questions. Our free AI visibility checker puts four of them to each.

Rewrite titles on pages that already rank

On Versely, the title and snippet rewrites on pages already ranking on page 1 or 2 had the clearest payoff. We can't cleanly split the credit between everything we changed, but those pages were already indexed and ranking. They just weren't getting chosen.

By week 10 or so you should have a few such pages. Look at the queries they get impressions for, and rewrite the title to match the words people actually use. We walk through the method in title tags that get clicked.

What to skip in the first 90 days

• Chasing a third-party authority score. Domain Rating and similar metrics are estimates from SEO tool companies, not Google signals.
• Hundreds of programmatic pages on day one. They can work later, from real data, with quality bars. Read programmatic SEO without getting deindexed first.
• "AI SEO" hacks. Google's guide to generative AI features says there's no requirement to break your content into tiny pieces, that you don't need new machine-readable files or AI text files to appear in Google Search, and that seeking inauthentic mentions across the web isn't as helpful as it might seem.
• Daily rank checking. Positions on a new site jump around. Weekly is plenty.

If you want help running it

This plan is simple, but it doesn't run itself. The weekly loop is the part most founders drop by week 6.

That loop, the paced publishing and the title rewrites, is what we built into Autopilot: $49/mo at the founding price, with a 7-day free trial. If you'd rather do it yourself, everything above works with free tools and a spreadsheet. The key is to still be doing it on day 90.