The best way to rank on a keyword is to pick a different keyword

There is no best way to rank on any keyword, and the honest version of that sentence is the useful one. Ahrefs sampled 1 million URLs its crawler saw in September 2023 and followed them for a year. 1.74 percent reached Google's top ten. Filtered down to English pages that were not empty, the figure rises to 6.11 percent, which the author calls the fairer number (Ahrefs, How Long Does It Take to Rank in Google, 15 May 2025). Either way you round it, the base rate for a new page is somewhere between one in sixteen and one in fifty-seven.

So the question worth asking is not how to win a keyword. It is which keyword is winnable at all, and what work moves the odds. I went looking for numbers I could check. The clearest one I have is my own site's Search Console export, which shows the ranking problem in its early and least flattering form.

Everything below is either a Google statement about its own systems or a study with a published sample size and method. Google's documentation carries no publication date and changes without notice, so every quotation from it here was checked on 27 August 2026 against Search Central and the Search Console help pages. Where a claim is mine rather than sourced, I say so and show what it rests on.

Most pages get nothing

96.55 percent of pages in Ahrefs' 14 billion page Content Explorer index get zero Google organic traffic. Another 1.94 percent get between one and ten visits a month, leaving 3.45 percent with anything you could call traffic (Ahrefs, 96.55% of Content Gets No Traffic From Google, 1 December 2023).

The author flags the limits himself, and they matter. 14 billion pages is small against the 340.8 billion in Ahrefs' larger index and skewed towards the better end of the web. Traffic is estimated from a database of roughly 651 million keywords, so a page living entirely on long-tail queries outside that set reads as zero when it is not.

Both caveats push the same way. The sample is biased towards good pages, which means the figure across all pages is worse, not better. Publishing a page does not enter you into a contest. It buys a ticket in a draw where reading the rules first changes the odds.

The base rate for a page is zero. Everything you do is an argument against that default.

Age is the variable nobody wants

The average page ranking first in Google is five years old. In Ahrefs' 2017 run of the same study it was two. 72.9 percent of top ten pages are more than three years old, against 59 percent in 2017, and only 13.7 percent are under a year old, down from 22 percent (Ahrefs, same study, across 1.3 million random US keywords).

There is a second number in that study that cuts the other way and is easy to miss. Of the pages that did reach the top ten, 40.82 percent got there within a month. So the distribution is not a slow uniform climb. Either a page moves early or it mostly sits.

That combination is the single most actionable thing in the research. If a page has not shown movement in Search Console within roughly a month, the problem is usually not patience. It is the target. Six months of waiting on a query you were never going to win is six months you could have spent on four queries you could.

What Google says does not work

Google's SEO starter guide names things people still sell as tactics and says flatly that they do nothing. Worth reading in the original, because the phrasing is blunter than the summaries of it.

The keywords meta tag: Google "doesn't use" it. Keyword stuffing: "excessively repeating the same words over and over (even in variations) is tiring for users" and breaches the spam policies. Keywords in the domain or URL: "hardly any effect beyond appearing in breadcrumbs." Word count: "the length of the content alone doesn't matter for ranking purposes" and there is "no magical word count target, minimum or maximum." Heading count and order: useful for screen readers, out of order headings "don't matter to Search." And on whether E-E-A-T is a ranking factor: "No, it's not."

That last one gets misread constantly. Google's helpful content guidance says trust is the most important of the four and that related signals weigh more on health, finance and safety topics. What it denies is that there is an E-E-A-T dial in the ranking system. The concept describes what the systems are trying to approximate. It is not itself a lever.

The starter guide, on duplicate content: "it's inefficient, but it's not something that will cause a manual action." There is no duplicate content penalty. What exists is a deduplication system that suppresses near identical results. That is a different mechanism and it has a different fix. Canonical tags, not paraphrasing.

The list of systems is public

Google publishes the names of its ranking systems, and reading that page is a faster education than most SEO courses. It tells you which of your beliefs are about live systems and which are about systems that were folded into core ranking years ago.

Live and named: link analysis systems and PageRank, still part of core ranking. Freshness systems, described as "query deserves freshness," which is why a news query and an evergreen query behave differently on the same words. The passage ranking system, which scores individual sections of a page, so one strong section can carry a page whose overall topic is broader. Neural matching and RankBrain for concepts rather than strings. The site diversity system, which usually caps one site at two results in the top set. The original content systems, which try to keep the source above the sites citing it. The exact match domain system, which specifically limits credit for domains built to match a query word for word.

Retired, and listed as retired: Panda, folded into core ranking in 2015. Penguin, folded in 2016. Hummingbird, from 2013. The helpful content system, launched in 2022 and merged into core ranking in March 2024. If your plan involves recovering from a Penguin penalty, your plan is ten years stale.

The site diversity cap is the one that changes strategy most and gets discussed least. Two results per site in the top ten means the marginal fourth article on the same query is not competing with other sites. It is competing with your own other three for two slots that were always going to be two.

The click you are competing for is smaller now

Pew Research Center installed a browsing meter on 900 US adults from KnowledgePanel Digital and logged 68,879 Google searches during March 2025, of which 12,593 returned an AI summary. On result pages with a summary, 8 percent of visits included a click on a traditional result link. Without a summary, 15 percent did, close to double. Clicks on links inside the summary accounted for 1 percent of visits to those pages (Pew Research Center, 22 July 2025).

Read the rest of that study before you panic about it, because the parts that get quoted least are the parts that tell you what to do. Around two thirds of all searches ended without a click on any result, with or without a summary. The zero-click majority predates AI. What the summary changed is the size of the minority, not its existence.

Summaries appeared on 8 percent of one and two word searches and 53 percent of searches of ten words or more, and on 60 percent of searches starting with a question word. Length and question form predict whether a summary shows up better than topic does.

That points somewhere specific. Short commercial queries still hand you a normal result page. Long question-shaped queries mostly get answered above you. If you write nothing but question-shaped headings, you are volunteering for the surface with the 8 percent click rate.

There is no separate optimisation for AI results

Google's own documentation on AI features is unusually direct about this. "There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary." And specifically: "You don't need to create new machine readable files, AI text files, or markup to appear in these features," with "no special schema.org structured data that you need to add."

Eligibility is stated as one condition. A page has to be indexed and eligible to appear with a snippet. That is the whole gate. If you have applied nosnippet or max-snippet limits, you have opted out of these surfaces as a side effect, which is worth checking before you conclude you were passed over.

Google also claims clicks from AI Overview pages "are higher quality (meaning, users are more likely to spend more time on the site)." I have no way to verify that from outside, and neither does anyone else quoting it, so treat it as the vendor's characterisation of its own product rather than a finding. The Pew number is measured. This one is asserted.

My own index report, which is mostly not indexed

Here is the worked example. My Search Console coverage export dated 24 August 2026 shows 13 pages indexed and 29 not indexed. The breakdown of the 29: 21 discovered and currently not indexed, 4 excluded by a noindex tag, 3 with a redirect, 1 crawled and currently not indexed. Impressions across the tracked window: zero. That zero is a starting line rather than a collapse, because I had only just connected the property, and it is the more useful case to look at for exactly that reason.

Two of those categories are not problems, and telling the difference is most of the skill. The 4 noindex pages are noindex because I set them that way. The 3 redirects are redirects because I moved things. Google's own documentation says of the whole not indexed group that "these pages have not been indexed, but not necessarily because of an error."

The 21 discovered pages are the signal worth reading, and the meaning is narrower than it looks. Google defines the state as "the page was found by Google, but not crawled yet," and gives a cause: "typically, Google wanted to crawl the URL but this was expected to overload the site," so "Google rescheduled the crawl." That is a statement about crawl budget and host capacity on a site with almost no crawl history. It is not a verdict on the writing.

The single crawled and currently not indexed page is the one that would worry me if there were twenty of them. Google's advice there is to do nothing: "it may or may not be indexed in the future; no need to resubmit this URL for crawling." One page in that state on a young site is noise. A pattern of it is Google having looked and declined.

Zero impressions on a site that is 31 percent indexed is not a ranking problem. Ranking is a thing that happens after indexing.

The order matters more than any tactic in this post. Nothing about keywords, headings, internal links or schema does anything for a URL Google has not fetched. Work the index report until the pages you care about are in it, then start reading position data. Doing it in the other order produces months of edits to pages that were never in the running.

What moves it

Google's guidance and the two Ahrefs studies agree on more than they disagree, and the overlap is short enough to list.

Get indexed first. Internal links are the mechanism, not sitemaps. Google says "the vast majority of the new pages Google finds every day are through links," and calls sitemaps optional in the same guide. A page reachable only from a sitemap is a page you have told Google about and not given it a reason to fetch.

Pick queries where you can plausibly be the best result rather than the tenth adequate one. Google's self-assessment questions are the honest test: does the page offer original information, reporting, research or analysis, does it provide insight "beyond the obvious," would you expect to see it cited in "a printed magazine, encyclopedia, or book." Most pages targeting a competitive head term fail all three, which is why they are in the 96.55 percent.

Write the thing you can write and nobody else can. Original data, a documented process, a result with numbers attached. Google's warning signs for search-engine-first content name the alternatives directly: publishing across unrelated topics hoping something ranks, "using extensive automation to produce content on many topics," mostly summarising others, entering a niche purely for traffic.

Then wait a specific amount of time rather than an anxious one. Google says some changes take hours and others months, and advises waiting "a few weeks" before assessing. For a core update it recommends waiting at least a week after the rollout completes, then comparing that week with a week before it started. The 40.82 percent figure suggests a month is a reasonable point to judge a new page rather than nurse it.

The trap in the phrase

"Rank on any keyword" contains an assumption worth naming. It treats keywords as interchangeable and rankable with enough effort, which the data does not support. A five year old page with links, a site that already holds two of the ten slots, and a query that returns an AI summary on 53 percent of its long-form variants are three fixed conditions you do not out-effort.

What you choose is which queries to enter. That choice is worth more than everything downstream of it combined, and it is the only part of this that is fully under your control.

Google puts a caveat at the top of its own starter guide that most SEO articles would never survive printing: "There are no secrets here that'll automatically rank your site first in Google (sorry!)". Fifteen years of tactics posts exist because that sentence is unsatisfying. It is still the accurate one.