First 90 days of a new site: the week-by-week SEO plan

TL;DR: A new site will not rank meaningfully in 90 days; Ahrefs’ analysis of 1 million URLs found only 1.74% of newly published pages reach the top 10 within a full year, and John Mueller has said Google can take up to a year to settle a new site’s rankings. What the first 90 days actually buys you is the technical foundation, the first 15-20 published pages, and the crawl and indexing signals that determine whether month four through month twelve go well or badly. Treat day 90 as a setup checkpoint, not a results deadline.

---

Table of contents

• What actually happens in the first 90 days
• Days 1-7: the technical foundation before you publish anything
• Weeks 2-4: publish your first 15-20 pages
• Weeks 5-8: distribution and the signals Google actually watches
• Weeks 9-12: measure, prune, and plan month four
• Why “rank in 90 days” is not a realistic promise
• The realistic 90-day scorecard
• The week-by-week plan as a flow
• Mistakes that reset the clock
• What to do if day 90 shows nothing
• Frequently asked questions
• Key takeaways

What actually happens in the first 90 days

A new site’s first 90 days are for building the technical and content foundation Google needs before it can rank anything, not for winning rankings. Google’s own SEO starter guide tells new site owners to start with a Search Console account and confirm the site is crawlable before worrying about anything else, and its Get on Google guidance states plainly: “It can take a few weeks for Google to notice a new site, or any changes in your existing site.” Weeks just to get noticed. Ranking comes after that.

The data backs up the slow start. Ahrefs studied 1 million random URLs and 1.3 million random US keywords and found that only 1.74% of newly published pages land in the top 10 within a year, down from 5.7% when Ahrefs ran the same study in 2017. Filter for real, non-empty English content and the number rises to 6.11%, still a small minority. Among the pages that do eventually reach the top 10, 40.82% got there within the first month, but that statistic describes pages that already had a reasonable shot, usually on sites with existing authority. A brand-new domain with zero backlinks is not drawing from that same pool.

John Mueller, Google’s Search Advocate, has described what practitioners call the “sandbox effect”: new sites can take up to a year before Google settles on where to rank them. Nobody at Google has published an exact mechanism for this, and Google does not confirm a formal sandbox exists as a named ranking factor. What is confirmed is the outcome: new domains rank inconsistently and often poorly for months regardless of content quality, and that volatility is a normal part of a new site’s life, not a sign something is broken.

None of this means the first 90 days do not matter. They determine whether Google can find, crawl, and understand your site at all, whether your first content batch gives it enough to evaluate, and whether the signals you send in month one set up month four to six months of ranking momentum or a slow, avoidable restart. The plan below is organized in four-week blocks because that is roughly how Google’s own crawl and evaluation cycles operate for a new domain, not because 90 days is a magic number.

Days 1-7: the technical foundation before you publish anything

Set up Google Search Console and verify ownership before you publish a single page. Search Console is how you will see whether Google has found your site, how it is crawling you, and where indexing is failing, and none of that visibility exists until verification is done. Google’s starter guide frames this as the first step for a reason: everything after it depends on being able to see what Google sees.

Confirm your site is actually crawlable. Load your homepage and a handful of interior pages in an incognito browser window to check nothing behind a login wall, a cookie gate, or broken JavaScript is blocking content Googlebot needs to render. Google’s documentation is explicit that CSS and JavaScript resources need to load correctly, because the crawler renders the page much like a browser does, and a page that looks fine to you but fails to render for a crawler is invisible regardless of how good the content is.

Check whether Google already knows you exist by searching site:yourdomain.com. A blank result means you are starting from zero discovery, which is normal for a domain registered in the last few days. It also means your next steps, sitemap and internal links, carry more weight than they would on a site Google already crawls regularly.

Build and submit an XML sitemap through Search Console. Google’s sitemap documentation is careful to call a sitemap “merely a hint,” not a guarantee Google will crawl every URL in it, but for a new domain with no inbound links yet, a sitemap is one of the only discovery paths you control. Submitting it also gives you visibility in Search Console into when Googlebot last accessed it and whether it hit processing errors, which is diagnostic information you will not get any other way in week one.

Fix URL structure now, while changing it costs nothing. Google’s guidance recommends descriptive URLs that include words useful to a reader, not database IDs or randomized strings, because retrofitting URL structure after Google has indexed and ranked pages means managing redirects on top of everything else. A new site has zero indexed URLs to redirect; use that.

Decide on your canonical domain (www or non-www, trailing slash or not) and set the 301 redirects before publishing, not after. This single decision, made wrong, creates duplicate-content signals that Google’s guidance explicitly calls out as something to avoid by consolidating each piece of content to one URL.

Weeks 2-4: publish your first 15-20 pages

Publish enough pages that Google has something real to evaluate. A site with three pages does not give Google’s systems much signal about what the site is about or how good it is at anything; a site with 15-20 focused pages targeting real, researched keywords gives Google enough surface area to start forming a picture. This is the coverage-first logic behind why new sites benefit from higher publishing volume than mature sites do: at low page counts, each additional page still adds meaningful new targeting surface, a dynamic that reverses once a site has hundreds of pages competing with each other.

Do the keyword research before the writing, not alongside it. A full keyword-to-published-post workflow exists precisely because skipping straight to drafting produces pages built around guesses instead of demonstrated search demand. Fifteen pages built on real keyword research outperform thirty pages built on assumptions, and thirty pages built on assumptions is exactly the kind of content Google’s helpful content guidance targets when it describes content “produced primarily to generate ranking signals rather than to help a reader.”

Writing 15-20 genuinely researched pages by hand in under a month is a real labor problem for a solo founder or a two-person team, which is the specific gap Vrid.ai is built for: AI article generation with word-count control, so you can hit a defined length target across a keyword list without the per-article hours a small team does not have in week two of a launch. The tool does not replace the keyword research step above it; it removes the drafting bottleneck once the research is done.

Write unique title tags and meta descriptions for every page from the start. Google’s starter guide calls unique, clear, concise titles a search-appearance priority, and fixing 20 generic titles later is a smaller job than fixing 200 later, so do it correctly the first time while the page count is still small.

Add internal links between your first pages as you publish them, not as a cleanup task later. A new page with zero internal links pointing to it is harder for Google to find through crawling alone and sends a weaker relevance signal than a page linked from two or three related pages on the same site. This is also the point where you start building the topical authority structure that determines how a cluster of pages, rather than any single page, earns Google’s trust on a subject over time.

Skip schema markup entirely in week two if you have to choose, but do not skip it past week four. Structured data is not required for AI search or for Google ranking in the sense of being a hard gate, but it is a low-cost way to help both Google and AI answer engines parse what each page actually is, and it is far cheaper to add while you have 20 pages than to retrofit across 200.

Weeks 5-8: distribution and the signals Google actually watches

Get your first real backlinks. Google’s own guidance is direct about why new sites get missed: “Your site isn’t linked to by other sites on the web.” A brand-new domain with zero external links has almost nothing pointing search engines toward it beyond the sitemap you submitted in week one, and link discovery is one of the two channels Google names explicitly for finding new pages, alongside sitemaps.

Prioritize links that a real person would click, not links built purely to manipulate rankings; Google’s spam policies explicitly penalize paid link schemes, and a young domain has less accumulated trust to absorb a manual action than an established one does. Directories relevant to your niche, guest contributions to sites your audience already reads, and genuine partnerships are slower than buying links but do not carry the risk of erasing three months of setup work in one penalty.

Check your Search Console coverage report weekly during this stretch, not monthly. This is where you catch indexing failures early: pages stuck in “Discovered, currently not indexed,” pages blocked by an accidental robots.txt rule, or a canonical tag pointing somewhere it should not. A systematic Search Console review that most site owners never run at this stage catches problems in week six that would otherwise go unnoticed until month four, when they have already cost weeks of crawl time.

Watch average days-to-index on each new page you publish during this period. If pages you published in week two are indexed within days, you have healthy crawl allowance. If pages from week five are still not indexed by week eight, that is a signal worth investigating now, not after publishing another 20 pages on top of the backlog.

Run the technical basics you deprioritized in week one if you have not already: confirm mobile rendering works cleanly, check Core Web Vitals in Search Console’s report even though a low-traffic new site will have limited field data, and verify HTTPS is configured correctly sitewide with no mixed-content warnings.

If your business has any physical location or service area component, this is also the window to claim and complete a Google Business Profile, since local pack visibility runs on a mostly separate, and often faster, evaluation cycle than organic web rankings do.

Weeks 9-12: measure, prune, and plan month four

Pull every published URL from Search Console with impressions and clicks, even if the click numbers are close to zero. Impressions without clicks at this stage are not a failure signal; they mean Google has indexed and is occasionally surfacing the page, which is exactly the evaluation phase a new domain is supposed to be in during its first quarter. Zero impressions on a page published six weeks ago is the signal actually worth investigating.

Check average query position, not just impressions. A page sitting at position 60-90 for its target keyword at week 10 is behaving normally for a new domain; Google is still gathering signal on it. A page with zero impressions at all for its target keyword after six weeks, despite being indexed, usually points to a content-relevance or on-page targeting problem worth fixing before you write ten more pages with the same issue.

Decide what to do with your weakest early pages now rather than letting them sit. If two or three of your first 15 pages are clearly thin, rushed, or built around a keyword with real demand you misjudged, this is the cheapest point in the site’s life to rewrite or consolidate them; a content refresh versus new content decision made at page count 15 costs a fraction of the same decision made at page count 150.

Set your month-four publishing cadence based on what weeks 2-8 actually taught you about your own research and production capacity, not on a number you picked before launch. If your keyword research pipeline supports four well-researched pages a week, publish four; a cadence that outpaces your research depth produces exactly the kind of thin, ranking-signal-chasing content Google’s helpful content guidance targets, regardless of how fast a tool can generate the words.

Review your KPIs against what is realistic for a 90-day-old domain, not against a mature-site benchmark. The SEO KPIs that actually matter once raw clicks stop telling the whole story, impressions, indexed page count, average position trend, and referring domain count, are more honest indicators of quarter-one progress than organic sessions, which will legitimately be near zero for most new sites at this stage regardless of how well the first quarter went.

Why “rank in 90 days” is not a realistic promise

Every agency pitch that promises page-one rankings in 90 days is selling against data that does not support it. Ahrefs’ age analysis of Google’s top 10 results found 72.9% of pages currently ranking in the top 10 are more than three years old, up from 59% when Ahrefs ran the same analysis in 2017, and the average page holding the #1 spot is now five years old, up from two years in the earlier study. Only 13.7% of top-10 pages are under a year old, down from 22%. The competitive set a new site is trying to break into has gotten older and more entrenched, not younger, over the past decade.

That does not mean nothing happens in 90 days. It means what happens in 90 days is setup, not victory. The gap between the two matters because a founder or a client evaluating “did SEO work” against a 90-day ranking promise will conclude it failed, when the actual failure was setting the wrong expectation at day zero. Understanding what SEO content actually costs and running an honest ROI calculation both depend on using the real 6-to-12-month timeline as the input, not a 90-day fiction.

The mechanism behind the slow start is not a Google conspiracy against new sites; it is a trust problem Google has no shortcut for solving. A domain with no publishing history, no backlink profile, and no user-behavior data has given Google nothing to evaluate except the content itself, and content alone has never been the only input to ranking. Every one of the levers a new site can pull in its first quarter, technical setup, initial content, early links, is aimed at giving Google that missing history as fast as legitimately possible, not at forcing a faster verdict.

The realistic 90-day scorecard

What you’re checking at day 90	Realistic target	Red flag
✓ Pages indexed in Search Console	80%+ of published pages	✗ Under 50% indexed after 8+ weeks live
✓ Search Console impressions	Rising week over week, even from near zero	✗ Flat at zero impressions past week 6
✓ Average query position	Anywhere in positions 20-90, trending down over time	✗ No ranking data at all for target keywords
✓ Referring domains	3-10 legitimate, relevant links	✗ Zero external links, or links from clearly spammy sources
✓ Published page count	15-25 researched, unique pages	✗ Fewer than 10, or 30+ thin pages published on a fixed calendar
✗ Organic sessions in Google Analytics	Low, often single or double digits	Expecting meaningful session volume this early
✗ First page-one ranking	Not expected yet for most competitive terms	Judging the whole quarter’s work against this one metric

Read the table by row: the checked rows are what a healthy new site actually looks like at day 90, and the unchecked rows are the metrics most owners mistakenly use to judge the quarter, then conclude the work failed when the real indicators were fine.

The week-by-week plan as a flow

flowchart TD
    A["Day 1: Domain live"] --> B["Days 1-7:\nSearch Console + sitemap\n+ crawlability check"]
    B --> C["Weeks 2-4:\nKeyword research\n+ publish 15-20 pages\n+ internal links"]
    C --> D{"Indexing on track?\n(check Search Console)"}
    D -->|"Yes"| E["Weeks 5-8:\nEarly backlinks\n+ weekly coverage checks"]
    D -->|"No"| F["Fix crawl/index issue\nbefore publishing more"]
    F --> E
    E --> G["Weeks 9-12:\nReview impressions + position\nPrune weak early pages"]
    G --> H{"Impressions rising,\npages indexed?"}
    H -->|"Yes"| I["Month 4:\nScale cadence to\nresearch capacity"]
    H -->|"No"| J["Diagnose: crawlability,\nrelevance, or link gap\nbefore scaling volume"]
    J --> I

Mistakes that reset the clock

Publishing before the technical foundation is fixed is the most common mistake, because it feels productive while actively working against you: pages published before a canonical domain decision or a broken robots.txt rule is caught get indexed under the wrong URL structure or not at all, and fixing that after the fact means Google has to re-crawl and re-evaluate pages it already partially processed.

Buying links to force early authority is the second, and it is the one most likely to end a new site’s chances entirely rather than just delaying them. Google’s spam policies treat paid link schemes as a manual-action risk, and a domain with almost no established trust has far less room to absorb a penalty than a site with years of legitimate signal behind it.

Publishing on a fixed calendar instead of a research pipeline produces the thin-content problem in miniature during the exact window a new site can least afford it. Twenty pages published in three weeks because a plan called for twenty pages, with keyword research done after the fact or not at all, is worse for a new domain’s early trust signals than twelve pages that each answer a real, researched query.

Ignoring Search Console coverage reports until month three means any indexing problem from week two has been compounding, unaddressed, for ten weeks by the time it is caught. Weekly checks in the first quarter are not excessive; they are the only way to catch a robots.txt mistake or a canonical error while it has cost days instead of months.

Comparing month-one traffic to a competitor’s current traffic is a mistake of framing, not execution, but it does real damage: it convinces founders and stakeholders the strategy is not working during the exact window when abandoning it is most costly, right as the technical and content foundation is finally in place for month four to compound on it.

What to do if day 90 shows nothing

Check Search Console coverage first, before assuming the content is the problem. If a meaningful share of your published pages are not indexed at day 90, the issue is discovery or crawlability, not content quality, and no amount of additional writing fixes an indexing problem.

Check whether your pages are indexed but simply not ranking for anything yet. Indexed-but-invisible at day 90 is close to the expected state for a new domain per the data above; it is not evidence of failure, it is evidence you are on the normal timeline and need more time and more signal, not a different strategy.

Audit your first content batch honestly against real search intent, not against how much effort went into writing it. A well-written page targeting the wrong query, or a query with intent that does not match your page format, will sit with impressions and no clicks regardless of prose quality; this is a targeting fix, not a rewrite-everything fix.

Check your referring domain count. Zero external links after 90 days on a competitive niche is a genuine gap worth addressing directly in month four, through the same channels described in weeks 5-8, rather than assuming more content alone will close it.

If the technical foundation, indexing, and initial link profile all check out and the site still shows nothing at day 90, that is not a failure signal yet either; it is confirmation you are inside the normal range Ahrefs and Mueller both describe, and the correct move is to keep executing the plan into month four through month six, when the real ranking signal starts to show up for sites that did the first quarter correctly.

Frequently asked questions

How long does it actually take a new website to rank on Google?

Most new sites see meaningful rankings between 6 and 12 months, not 90 days. Ahrefs’ 1-million-URL study found only 1.74% of new pages reach the top 10 within a year, and John Mueller has said Google can take up to a year to settle a new site’s rankings. The first 90 days build the foundation that timeline runs on top of.

What is the Google sandbox and is it real?

The “sandbox” describes new domains ranking inconsistently or poorly for months regardless of content quality, a pattern John Mueller has acknowledged can last up to a year. Google has not confirmed a formal, named sandbox mechanism, but the outcome practitioners describe is well documented and should be treated as a normal, expected phase, not a bug.

Should I submit a sitemap for a brand-new site?

Yes. Google’s sitemap documentation calls a sitemap “merely a hint,” not a guarantee, but for a domain with no inbound links yet, it is one of only two discovery channels Google names, alongside link discovery, so submitting one in week one costs nothing and helps.

How many pages should a new site publish in the first month?

15-20 focused, keyword-researched pages gives Google enough surface area to start forming a picture of the site, based on the coverage-first logic that applies specifically to low-authority new domains. Fewer than 10 gives Google too little to evaluate; 30+ pages published without real keyword research behind them creates thin-content risk instead of coverage.

Does Google penalize new sites for using AI-generated content?

No, not for the production method itself. Google’s helpful content guidance states policy targets content produced to manipulate rankings, regardless of who or what wrote it. AI-assisted writing with word-count control can help a small team hit a first-quarter coverage target faster, provided the underlying keyword research and editing are still real.

How many backlinks does a new site need in the first 90 days?

There is no fixed number; Google’s own guidance simply states that sites often get missed because “your site isn’t linked to by other sites on the web.” A handful of legitimate, relevant links, in the single digits, is a reasonable first-quarter target; the source and relevance of each link matters more than hitting a count.

Is it bad if my new site has zero organic traffic after 90 days?

Not necessarily. Given that 72.9% of current top-10 pages are over three years old per Ahrefs, near-zero organic sessions at 90 days is close to the expected state for a new domain. Check Search Console impressions and indexing status before concluding anything is wrong; those metrics move earlier than session counts do.

Should I set up Google Search Console before or after publishing content?

Before. Search Console verification is the only way to see whether Google can find, crawl, and index your pages, and catching a crawl or indexing problem in week one costs far less than catching the same problem after publishing 20 pages on top of it.

What should I prioritize in week one: content or technical setup?

Technical setup. Search Console verification, crawlability checks, sitemap submission, and canonical URL decisions all cost less to get right before any pages exist than to retrofit afterward, and content published on top of an unresolved technical issue gets indexed incorrectly or not at all.

How do I know if my new site’s slow start is normal or a real problem?

Check Search Console coverage first. Indexed pages with low impressions and no clicks at 90 days is normal for a new domain. Pages that are not indexed at all after 8+ weeks, or zero Search Console impressions for target keywords despite indexing, point to a real crawlability or targeting problem worth investigating immediately.

Do I need schema markup on a brand-new site?

It is not required to rank, but it is inexpensive to add correctly while you only have 15-20 pages, and far more expensive to retrofit across hundreds later. Add it by week four if you skipped it in the rush of initial publishing.

How often should I check Search Console during the first 90 days?

Weekly, not monthly. Indexing problems, robots.txt mistakes, and canonical errors caught in week two cost days to fix; the same problems caught in month three have been silently limiting every page published since, at real cost to the whole quarter.

What’s the biggest mistake new site owners make in the first 90 days?

Judging the quarter against a page-one ranking that the data says is not realistic yet, and concluding the strategy failed when the actual indicators, indexing rate, rising impressions, growing referring domains, were all healthy. The second most common mistake is buying links to force faster trust, which risks a manual action a young domain has little room to absorb.

Should I focus on one page type or a mix in the first 90 days?

A mix built around real keyword research, not a single format applied uniformly. Google’s helpful content guidance rewards pages that genuinely answer the query behind them; a diverse first batch targeting different real search intents gives Google more to evaluate than 20 pages built off one template.

How does site age affect ranking versus content quality?

Both matter, but age is a proxy for accumulated trust signals, not a direct ranking factor Google names. Ahrefs’ data shows the average #1-ranking page is five years old, which reflects years of accumulated links, updates, and user signal, not that Google explicitly favors older domains for their age alone.

Can a new site rank faster in a low-competition niche?

Yes, in relative terms. The mechanisms described above (indexing, trust-building, link accumulation) still apply, but a niche with fewer established, multi-year-old competitors gives a new site a realistic shot at page one sooner than a saturated niche where the top 10 is dominated by five-year-old domains, per the age data above.

Should I pause publishing if my pages aren’t ranking by week 8?

No, unless Search Console shows an indexing or crawlability problem. Position in the 20-90 range at week 8 is on the normal timeline described by the 6-to-12-month data; pausing publishing removes the coverage-building advantage a new site specifically benefits from during this phase.

How do I set expectations with a client or stakeholder for the first 90 days?

Frame the deliverable as foundation and indexing health, not rankings: percentage of pages indexed, impression trend, referring domain count, and the published page count against a real keyword list. Set the ranking conversation for month six, backed by the Ahrefs and Mueller data above, not for the 90-day mark.

What role does content refresh play this early?

A limited one, but not zero. If two or three of your first pages are clearly weak by week 10, refreshing them now costs a fraction of what the same fix costs once the site has 150 pages and a backlog of similar issues.

What happens in month four if the first 90 days went well?

Publishing cadence scales to match the research capacity the first quarter revealed, distribution continues past the initial link batch, and the site starts accumulating the age, link, and user-signal history that the Ahrefs data shows is what actually correlates with top-10 rankings, none of which the first 90 days could produce on their own.

Key takeaways

• A new site’s first 90 days build the technical and content foundation Google needs, not rankings; only 1.74% of new pages reach the top 10 within a year, and John Mueller has described up to a year before Google settles a new site’s rankings.
• Technical setup (Search Console, crawlability, sitemap, canonical URLs) comes before publishing, because fixing it after pages are indexed costs far more than getting it right in week one.
• 15-20 real, keyword-researched pages in the first month gives Google enough coverage to evaluate; 30+ pages published to a calendar without research behind them creates the exact thin-content risk Google’s helpful content guidance targets.
• Weekly Search Console checks catch indexing and crawl problems while they cost days, not months, and the realistic day-90 scorecard is indexing rate, impression trend, and referring domains, not organic sessions or a page-one ranking.
• Bought links and calendar-driven publishing are the two mistakes most likely to reset the clock on a new domain that has the least trust reserve to absorb either one.

Run your own first quarter against the week-by-week plan above, then check your Search Console data against the realistic scorecard before deciding whether month four needs a course correction or just more time.