Most SEO advice is written for sites that already have some authority. If your domain is three weeks old, that advice is not just unhelpful — following it will waste months.
A new site has no backlinks, no history, no topical authority and no trust. That changes what you should target and in what order.
Accept the constraint first
You cannot rank for competitive commercial queries. Not because your content is bad, but because Google has no reason to trust a domain with no track record over one with a decade of signals.
This is not permanent. It is a starting condition, and the whole strategy is about working within it rather than against it.
The practical consequence: you win with specificity, not volume.
Choosing queries you can actually win
Ignore search volume for the first six months. Volume is a measure of competition as much as opportunity.
What makes a query winnable for a new site:
- Long-tail — four or more words. "How to fix layout shift in WordPress" rather than "WordPress SEO".
- Specific intent — the searcher wants one particular answer, not an overview.
- Weak existing results — go and look. If page one is forum threads, thin articles and outdated posts, there is room.
- Low commercial competition — informational queries have fewer agencies fighting over them than "buy" queries.
- Niche-technical — combining two specific things ("Shopware 6 product feed for Google Shopping") often has almost no competition.
How to check competition without paid tools:
- Search the query. Read page one properly.
- Are the results from major brands and established sites, or from forums, Reddit and small blogs?
- Are the top results actually answering the question well, or approximately?
- Check the dates. Page one full of 2019 content is an opening.
If page one is Wikipedia, Forbes and three agencies you have heard of, pick a different query.
The 90-day plan
Days 1–14: foundation
- Google Search Console and Bing Webmaster Tools verified, sitemap submitted.
- Analytics installed.
- Confirm HTTPS and one canonical domain (www or non-www, not both).
- Every page has a unique, specific title tag and meta description.
- Organization and WebSite schema in place.
- Site loads fast on mobile — see Core Web Vitals.
- Every page reachable within three clicks of the homepage.
This is a few days of work. It will not make you rank, but nothing else works without it. Our SEO foundations guide covers this in more detail.
Days 15–45: build the first cluster
Do not publish scattered articles on unrelated topics. Pick one narrow topic and cover it thoroughly.
A cluster is one broad "pillar" page plus five to eight specific pages that link to it and to each other. This does two things: it gives Google a coherent signal about what you are an authority on, and it means each new page benefits from the ones already published.
Publish six to eight pieces in this window. Every one should genuinely be the best answer to its specific query — not the longest, the most useful.
Days 46–90: expand and start earning links
- Add a second cluster, adjacent to the first.
- Update the earliest pieces based on what Search Console shows people actually searching.
- Begin link work in earnest.
Earning the first links
This is the part with no shortcut, and the part that determines whether any of the rest works.
Things that work and are legitimate:
- Do work for people and get credited. If you build sites, a footer credit with permission is the most natural link there is.
- Claim your profiles. Google Business Profile, LinkedIn, industry directories, Clutch, Behance, GitHub. Individually low-value, collectively they establish that you exist.
- Be genuinely useful in public. Answer questions properly on Reddit, Stack Overflow, industry forums and Slack communities. Do not drop links; put the domain in your profile and let good answers do the work. This is slow and it compounds.
- Publish something original. Original data is the only reliable way a small site earns links from larger ones. Benchmark something. Survey something. Test fifty of something and publish the results.
- Guest post on genuinely relevant smaller sites. Not link farms — real blogs with real readers.
- Digital PR on a small scale. Local press, industry newsletters, podcast appearances.
Things that do not work: buying links, private blog networks, mass directory submissions, comment spam, "1,000 backlinks for $20". These range from useless to actively harmful, and a new domain has no trust buffer to absorb a penalty.
Ten relevant links from real sites will do more than a thousand from directories nobody reads. This has been true for fifteen years and is unlikely to change.
What to measure in the first 90 days
You will not have meaningful traffic. Measure leading indicators instead:
- Indexed pages — is Google finding and keeping your content?
- Impressions in Search Console — rising means you are appearing for more queries.
- Number of ranking queries — twenty to two hundred is real progress even at position 50.
- Average position — trending from 70 to 40 is movement.
- Referring domains — even three or four real ones matter early.
Clicks come later. See how long SEO takes for the full timeline.
Common mistakes on new sites
- Chasing head terms. Targeting "web design" on a two-month-old domain is a wasted quarter.
- Publishing thin content quickly. Twenty 400-word posts perform worse than five thorough ones.
- Spreading across unrelated topics. No cluster, no authority signal.
- Rewriting what already ranks. If you cannot add something, pick a different query.
- Giving up at month three. Month three is when it looks like failure and is not.
- Buying links out of impatience. The fastest way to make months 4–12 worse.
The honest summary
A new site with no links takes six to twelve months to become a real traffic channel, and that is with consistent work. Plan for other lead sources in the meantime.
The compensation is that it compounds. Content published in month two will still be earning traffic in year three. Very few marketing channels do that.