SEO positioning is the work of getting a specific page to a specific place in search results for a specific query. The distinction from SEO in general is narrower than most definitions make it sound: SEO is everything you do to a site, positioning is the focused effort on the handful of pages and queries that actually matter commercially.
The phrase is also a direct translation of posicionamiento SEO, which is the standard term in Spanish, so a lot of people arrive at it looking for a plain explanation of what search ranking is. This page covers both readings: what position means and how it is measured, then what actually moves it. If you arrived here in Spanish and the question is really about international content, our guía definitiva de SEO translation covers that instead. If Spanish is your working language for this, our guía definitiva de SEO translation covers optimising translated content for international search.
What a position actually is
Your position for a query is where your page appears in the list of results. Position 1 is the first organic result, position 11 is usually the top of page two.
Three things complicate that simple picture, and all three matter.
Position is personal. Two people searching the same term at the same moment see different results depending on location, search history and device. There is no single true position, only an average across many searches.
Position is not what you think it is in your reports. Average position in Search Console is the average of every position you appeared at, weighted by impression. A page showing at position 3 for one query and position 40 for twenty others reports an average that describes neither.
Organic position 1 is not the top of the page. Ads, AI overviews, featured snippets, map packs, video carousels and People Also Ask boxes all sit above it. On many queries the first organic result is well below the fold.
How to read your real positions
Search Console is the only source that reports your actual positions rather than estimating them. The useful view is not the site-wide average, which is meaningless.
Filter to a single query, then look at the position for that query on that page. Then sort your queries by impressions and find the ones sitting between positions 8 and 20. Those are pages Google already considers relevant and has not quite promoted, and they are almost always the cheapest available wins.
Moving a page from 12 to 6 typically produces more traffic than moving a page from 60 to 30, even though the second looks like a bigger improvement. A site with no history has no near misses yet, which is why SEO strategies for new websites start somewhere else.
What actually moves a position
Seven factors, in rough order of how much leverage each carries for a page that already ranks somewhere.
Search intent match. The single most common reason a good page will not rank. If the results page is full of listicles and you published a how-to guide, no amount of optimisation fixes the format mismatch. Look at what is ranking and ask whether your page is the same kind of thing.
Content depth against what ranks. Not word count. Whether you cover what the ranking pages cover, plus something they do not. Why coverage beats repetition is explained in our guide to semantic search and SEO.
Internal links. The most underused lever available. A page with three internal links from relevant pages on your own site outranks an identical page with none, and this costs nothing but attention.
Backlinks. Still the largest external factor, and the slowest and most expensive to change.
Title and on-page relevance. Cheap to fix, and worth checking that the title actually contains the query someone types.
Page experience. Speed and Core Web Vitals act more as a tiebreaker than a driver. A slow page with the best content usually still wins.
Freshness, on queries where it matters. Anything with a year in it, anything about a changing subject. Irrelevant on evergreen topics.
The intent check, in practice
Before optimising anything, search your target query and read the first ten results properly.
Count the formats. If eight of ten are listicles and yours is a guide, your problem is structure and not wording. Count the entries in the ranking listicles, because if they run fourteen and you have six, you are not competing on completeness.
Note what sections they all have that you do not. This takes fifteen minutes and it is more useful than any tool output, because it tells you what Google has already decided this query means.
Tracking position properly
Rank trackers check positions from a consistent location and device, which is what makes them comparable week to week even though no single reading is objectively true.
Ahrefs, Semrush and SE Ranking all track positions as part of their suites. Dedicated trackers like AccuRanker and Wincher do it faster and cheaper if that is all you need. Two of those tools get confused constantly, which we sort out in Surfer SEO vs Ahrefs. Their volume and difficulty figures are estimates, as our look at how accurate Semrush is explains.
Track fewer keywords than you think. Twenty commercially important queries checked weekly tells you more than two thousand checked daily, and the second costs ten times as much. How to set that up without drowning in data is covered in our guide to tracking SEO rankings.
Why position one is not always the goal
A large share of searches now end without a click, because Google answers the question directly in an AI overview or a featured snippet. Ranking first on a query that Google answers itself produces impressions and very little traffic. The figures behind that shift, with sources, are in our SEO statistics roundup.
This changes what a good target looks like. Position 4 on a commercial query where people have to click through to compare options is worth considerably more than position 1 on a definitional query that gets answered on the results page.
Judge a keyword by whether the click is likely to happen and whether the person clicking could plausibly buy from you. Volume and position are inputs to that, not the answer. How that judgement plays out in one vertical is shown in our breakdown of SEO for real estate agents.
A practical sequence
Week one: find the near misses. Search Console, filtered to positions 8 to 20 with real impressions. Usually between ten and forty pages on an established site. A proper crawl widens that list, and our roundup of the best SEO audit tools covers what to use.
Week two: check intent on the top ten of those. Read the results pages. Some will need a format change rather than an edit, and it is better to know that before you rewrite.
Week three: fix titles and add internal links. The cheapest interventions, and on near-miss pages they frequently move things on their own. Both are covered step by step in how to do SEO optimization yourself.
Week four onward: rewrite the ones with a real intent mismatch. Match the format the results page rewards, then beat it on depth. If the rewriting is more than you can staff, we ranked the best AI SEO services for it.
Then wait. Position changes take weeks, not days, and re-optimising a page every four days makes it impossible to tell what worked.
What to stop doing
Watching daily. Positions fluctuate constantly and daily movement is almost entirely noise. Weekly is enough.
Chasing average position across the site. It goes up when you delete bad pages and down when you publish new ones, which is the opposite of what those two events mean.
Optimising pages nobody searches for. A position 3 ranking on a query with eleven searches a month is not an achievement worth reporting. Run the query through an SEO ROI calculator before spending a week on it.
Ignoring what sits above you. If four ads and an AI overview occupy the screen, being first organically is a different prize than it used to be.
Where positioning work stalls
The near-miss list is easy to produce and hard to work through, because most of the items need a page rewritten or three new pages published, and that is real production work rather than an afternoon of tweaks. Buying that production is one option, and our explainer on AI SEO services covers what it involves.
That is where most positioning efforts stop, usually with a well-maintained spreadsheet. Our guide to SEO reporting tools covers tracking the work properly, and automated SEO software covers the production half.
If the constraint is that the list is long and nothing gets published against it, Distribb runs keyword research, writing, publishing and internal linking on a schedule, with a backlink exchange for the external half. There is a 3-day free trial. If your near-miss list is short and you have time to work it by hand, do that instead, because hand-edited pages beat produced ones every time.