SEO for new websites: what to do in your first 90 days

A new domain starts with nothing Google trusts: no history, no links, no evidence anyone finds it useful. That's not a penalty — it's a blank file. The first 90 days won't make you rank; that takes longer, on any site, with any budget. What the first 90 days decide is when the ranking eventually starts, and the difference between a good start and a wasted one is roughly a year.

Week 1: plumbing

Two evenings of setup that everything else depends on:

• Search Console verified, sitemap submitted, analytics installed, one conversion tracked — the full setup list stages one and two.
• HTTPS working, site usable on a phone, no leftover noindex from development — the classic new-site killer, worth double-checking.
• Titles and URLs that say what each page is. Boring and load-bearing.

Weeks 2–3: the pages that earn their keep

Before any blogging: one solid page per service or product, each answering what it costs, how it works, and who it's for, each with a visible next step. These pages are why the site exists — content marketing's whole job is eventually delivering readers to them. A new site with five strong service pages and no blog beats one with fifty posts and a thin 'Services' page.

Week 3: pick ten keywords you can actually win

New domains lose head-to-head fights, so don't schedule any. Build your first list from what customers say they searched and buyer-intent patterns, then keep only queries where the current first page looks beatable — other small businesses, forum threads, thin generic pages. Specific beats broad: 'emergency water heater replacement cost' over 'plumbing'. Low search volume is fine; low-volume specific queries convert, and ten small wins compound into the authority that makes bigger fights winnable later.

Weeks 4–12: publish weekly, no exceptions

One good post a week, each answering one of your ten queries better than what currently ranks: direct answer up top, concrete numbers and first-hand specifics throughout, internal links to your service pages and between related posts. Eight to ten posts by day 90 is enough — the schedule matters more than the count, because Google is watching whether this site is alive and because the habit is what most new sites fail at, not the writing.

The first 90 days aren't for ranking. They're for becoming the kind of site that ranks: set up, indexed, publishing on schedule, and pointed at winnable queries.

Throughout: collect the few links that matter early

You don't need a link-building campaign in the first 90 days. You need the obvious ones: your Google Business Profile and the two or three directories that matter in your industry, your suppliers or partners who list who they work with, your chamber of commerce or trade association. A handful of real links tells Google the site exists in the actual world — which measurably shortens the everything-takes-forever phase.

What to expect on the scoreboard

• Days 1–30: pages get indexed (chase any that don't). Traffic: none. Normal.
• Days 30–60: first impressions appear at positions 30–60 for long, specific queries. Traffic: still nothing worth mentioning.
• Days 60–90: impressions climbing, distinct queries widening, maybe the first few clicks. That's exactly on schedule — the signals to watch are impressions and query count, not visits.

The four ways new sites waste their first 90 days

• Chasing head terms a new domain can't win, then quitting when nothing moves.
• Publishing twenty posts in week one and nothing after — bursts don't build trust, schedules do.
• Buying cheap links or $99 'complete SEO' packages — a new domain has no cushion for spam signals.
• Redesigning the site in month two instead of publishing. New sites procrastinate with polish.

Day 91 isn't a finish line; it's when the flywheel is finally assembled. Keep the weekly habit, expand from your first ten keywords to the next ten, and start judging posts by what they convert as the traffic arrives. The sites that win year one are almost never the ones that knew secrets — they're the ones that were set up correctly in week one and still publishing in week fifty.