What actually moves your search positions

The chart of your average position is not a picture of your website. That is the assumption worth putting down before you open one of those charts again. A position is a ranking, a seat in a row of ten, and where you sit depends as much on who else showed up as on anything you did. Your page can get better in every way that matters and still slide two places because a competitor published something sharper on the same Tuesday. The number is not wrong. It is simply not the measurement most people think they are taking.

Start with what the number is made of. An average position averages across every search term you appeared for, on every device, in every country, and only across the searches where you appeared at all. Four of those five can shift in a week without you touching a single page. A phone-heavy week pulls the average one way. A new country finding you pulls it another. And appearing for a term you never used to appear for adds a fresh, low number to the pot, which drags the average down at the exact moment your visibility went up.

Do that arithmetic once and the effect stops being abstract. Suppose you rank fourth for one term that shows you a hundred times a month. Your average position is four. Now a second term starts surfacing you at position thirty, four hundred times, because a page you wrote became newly relevant to something broader. Your average position is now just under twenty-five. On the chart that looks like a collapse. In reality you gained four hundred appearances and lost nothing whatsoever. This is the most common way a ranking chart lies to somebody who is doing everything right.

Which is why the weekly wiggle deserves almost none of your attention. Results get reordered constantly, layouts change, the mix of people searching changes, and the difference between position nine and position eleven on a Wednesday is frequently nothing at all. Watching a single term day by day is the search equivalent of judging a product from three signups: a sample too small to carry the weight you are putting on it. If you would not conclude anything from sixty-seven ad clicks, do not conclude anything from a two-place move over four days.

a position is not a score you earn. it is a seat, and somebody else is already sitting in it.

Month to month, the list of things that genuinely move positions is short and unglamorous. A page that answers the search better than the pages currently above it. Internal links, so your own site says clearly which page is the authoritative one on a topic. Mentions and links from elsewhere, which arrive slowly and rarely on the schedule you want. The boring technical floor: the page loads, it can be indexed, and it is not accidentally hidden. And time, which is the ingredient people account for least. A new page that lands at position forty and climbs to fifteen over two months is behaving normally, not failing.

Two specific things are worth checking that most people never look at. The first is whether two of your own pages are competing for the same search, splitting the signal so neither one wins it. When that happens, the fix is editorial rather than technical: decide which page owns the topic, and make the other one point at it. The second is the page-one edge. Terms where you sit between roughly eight and fifteen are where a modest improvement pays best, because you are already close enough that a better title, a clearer opening, or a stronger section can move you into view. Improving something at position sixty is usually a much worse hour of work than improving something at position eleven.

Split brand from everything else before you draw any conclusion at all. Searches for your own product name will rank first, convert well, and make every average look healthy, and they mostly measure how many people already know you. Non-brand terms are the ones that tell you whether strangers can find you. A month where brand searches grow and non-brand stays flat is a month your existing audience talked about you, which is good news about a different thing entirely. Averaging the two together hides both stories.

What this means for a weekly routine is less than you might expect, and that is the point. Positions belong to the monthly review, next to the pages you shipped, where a trend has had enough time to become one. The daily glance is for the numbers that move fast and mean something quickly, the small set that tells you whether the product is fine this morning. Rankings are simply not that kind of number, and treating them as one converts noise into anxiety and anxiety into pointless rewrites.

This is roughly how we set it up in SiteOps. Search position data is pulled in from the tools you connect and lands on the same screen as your traffic and revenue rather than in a separate login you visit when you are already worried. Before there is enough data to say anything, the panel says so plainly instead of drawing a confident line through three points. And when the system drafts a piece of content aimed at a term you are close to winning, it waits for you to approve it, because which topics are worth owning is a judgment about your business rather than a calculation about a chart.

Weekly, look at what you shipped. Monthly, look at where you rank. Getting those two on their proper clocks removes most of the false alarms and nearly all of the panicked rewrites, and it leaves you with the only ranking question that has ever been worth answering: on the searches you actually want, is the page above yours better than yours, and what would it take to fix that?