Updating one old page beats publishing three new ones

The cheapest ranking work you are not doing

Every small business content plan looks the same. Publish something new. Publish something else new. Keep publishing.

Meanwhile, the pages that are already bringing in traffic sit exactly as they were written, slowly losing ground to newer pages from competitors who covered the same question more completely.

This is backwards, and the reason is straightforward. A page that already ranks has something a new page does not: proven relevance. Google has already decided that page is a reasonable answer to a real query, and has the crawl history, the internal links, and the engagement data to back it up. Improving it starts from that position. A new page starts from zero.

For a small business with a handful of hours a month for this work, the highest-return use of those hours is usually not a new post. It is making an existing one deserve the position it is close to.

Finding the candidates

Two queries in Search Console find almost every page worth touching.

Query one: what is slipping? Open the Performance report, set the date comparison to "last 6 months vs previous 6 months," switch to the Pages tab, and sort by impressions difference, ascending. The pages at the top are shedding visibility.

A slow, steady decline usually means the page is being out-covered — someone wrote something more complete and Google noticed. That is the ideal refresh candidate, because the fix is editorial and within your control.

A sudden cliff is a different problem. Check whether the URL changed, whether the page is still indexed, and whether a redirect broke. Do not refresh a page that is actually 404ing.

Query two: what is nearly there? Filter to positions 8 through 20, sort by impressions descending. These are queries where Google already associates you with the topic but ranks you on page two. The relevance exists; the depth does not.

Both lists will be short on a small site. That is the point — you want four or five names, not forty.

What counts as a refresh

Here is the line, stated plainly: if a reader who saw the old version would not notice the change, it is not a refresh.

Changing the date is not a refresh. Google's own guidance on publication dates asks that dates reflect when content was genuinely published or meaningfully updated, and that the visible date match what is in your markup. A site that re-dates everything monthly without changing anything is describing work it did not do.

Real refreshes look like some combination of:

Correcting what is now wrong. Prices, hours, service areas, staff, platform names, anything with a version number. This is the highest-value edit and the one most often skipped, because nobody re-reads their own pages.

Adding the sections the page is missing. Look at what people actually search around the topic — the queries in your own Search Console data are the best source, better than any keyword tool — and answer the ones the page currently does not.

Rewriting the introduction. Introductions age worst. They were written for a moment ("with the recent changes to…") and read as dated years later, at the exact point where a visitor is deciding whether to keep reading. This is where the twelve-word clarity test applies again.

Adding a real example. A worked example, a photo of an actual job, a number from your own records. This is usually the difference between a page that reads as generic and one that reads as written by someone who does the work.

Fixing the internal links. Link out to the pages you have published since, and — more importantly — link in from newer pages. A page that nothing links to is a page Google has little reason to re-crawl.

Adding an FAQ, if customers genuinely ask. Not for the rich result, which no longer exists, but because six real questions answered plainly is often the most useful part of a page.

What to do with pages that should not survive

Not everything gets refreshed. Two other outcomes are legitimate and both are better than leaving a stale page live.

Consolidate. If three thin pages all circle the same topic, they are competing with each other for the same queries and splitting whatever signals they have. Merge the best material into one substantial page and 301-redirect the other two to it. This is the same duplication problem that scaled location pages fall into, and the fix is the same: fewer, better pages.

Retire. A page about a service you no longer offer, an event that happened, a product that was discontinued. If something current covers the same ground, redirect to it. If nothing does, and the page has no traffic and no inbound links, remove it. Leaving it published means a customer will eventually find it and call about it.

The one thing not to do is nothing. A site accumulating obsolete pages gets steadily harder to navigate and steadily less trustworthy, and it is one of the warning signs that a redesign conversation is coming.

A rotation that survives contact with a real business

Ambitious refresh schedules fail for the same reason ambitious publishing schedules fail. Here is a cadence we have found sticks, offered as a working routine rather than a rule:

Quarterly — four pages. Pull the two Search Console lists. Pick four pages: two that are slipping, two that are in positions 8–20. Refresh those four properly. Four pages a quarter is sixteen a year, which on a thirty-page site means every page gets meaningful attention roughly every two years.

Annually — the facts sweep. One session, every page, checking only for things that are factually wrong: prices, hours, service areas, team members, phone numbers, third-party names. This does not require rewriting anything and it catches the errors that quietly cost you calls. Pair it with your annual domain and hosting check so it happens at a fixed time rather than when someone remembers.

Whenever something changes in the business. New service, dropped service, new suburb, price change, moved premises. Same day, not next quarter. These are the updates where being out of date is not a ranking issue but a customer-service one.

Expectations

Two honest caveats.

First, refreshed pages do not move instantly. Google has to re-crawl, re-evaluate, and re-rank, and on a small site with modest crawl frequency that is weeks rather than days. Refresh in a batch, then leave it alone for a couple of months before judging.

Second, not every refresh works. Sometimes a page is losing ground because the query itself is dying, or because the results page now has an answer box where the click used to be, or because a competitor genuinely covers the topic better than you reasonably can. Some pages should be consolidated rather than saved, and recognising that early is worth more than a fourth attempt at rewriting them.

What is reliable is the direction: hours spent on pages that already have traction produce more than the same hours spent on pages that have none. The publishing treadmill is more visible work. This is more useful work.