Your Best Page Is Three Years Old and Quietly Dying

The problem nobody has a dashboard for

Content decay is the slow loss of traffic by a page that has not changed, because everything around it did. Competitors published better material, the query's intent shifted, the facts aged, and the ranking systems reweighted. No single day looks bad, which is exactly why it runs for two years before anyone investigates.

Most businesses monitor total traffic, and total traffic hides this completely. New pages add traffic while old ones lose it, the line looks flat, and underneath it your best asset is being replaced by weaker ones.

Why 2026 accelerated it

Four Google updates landed through 2026: a Discover-only update in February, a spam update and a core update in March, and a further core update in May. The March core update rolled out from 27 March to 8 April and produced widespread movement across essentially every industry; the May update was reported as more volatile still.

The consistent finding across analyses of both was that topical authority, content freshness, and demonstrable expertise carried more weight than before. Sites with regularly updated, expert-led content weathered them better; thin and generic content continued to lose ground.

Read carefully, that is not a new ranking factor to chase. It is a statement that the gap between maintained and unmaintained content widened. Publishing and walking away used to cost you slowly. It now costs you faster.

Two different problems that look identical

Before fixing anything, separate these, because they need opposite responses and a traffic report cannot tell them apart.

Falling impressions mean you are being shown less often. The page is losing relevance or coverage — competitors are answering the query better, or the query itself has changed. This needs real content work.

Stable impressions with falling clicks mean you are still being shown and no longer being chosen. That is a title, description, or SERP-feature problem: an AI Overview may now be answering above you, or your title may be truncated. This often needs nothing more than a rewritten title, which is twenty minutes rather than a day.

Businesses routinely rewrite entire articles that had a title problem, and equally often tweak titles on pages whose content had genuinely been overtaken. Check Search Console page by page, comparing the last three months against the same period a year earlier, before deciding anything.

Which pages are actually worth saving

Not all of them, and the instinct to refresh everything is how refresh programmes stall.

Refresh pages that still target a query your buyers actually search, that once performed, and whose decline is explainable by staleness or competition. These have accumulated history and links, so improvement compounds from an existing base.

Consolidate where several thin pages cover overlapping ground and compete with each other. One strong page with the others redirected into it usually outperforms the sum, and this is the highest-return move on most older blogs.

Retire pages targeting queries that never brought buyers. A page ranking well for a search your customers do not run is not an asset. Redirect it somewhere relevant and stop maintaining it.

The filter to apply first is commercial, not technical: does this page attract the person who hires you? A large share of decaying content was never going to convert, and refreshing it is expensive housekeeping.

What a real refresh involves

Changing the date does nothing. Search systems evaluate whether the content meaningfully changed, and a new timestamp on identical text signals low quality rather than freshness — while destroying your own ability to tell which pages actually need work.

A refresh that works usually means most of the following. Update every fact that has an expiry — prices, statistics, product names, regulatory positions, tool comparisons. Re-examine the query's intent, because what people want from a search shifts, and a page written for a how-to intent that has become a comparison intent needs restructuring rather than editing. Add what competitors now cover and you do not. Restructure for extraction, since a self-contained answer directly under each heading is what AI systems quote, and older posts that open narratively rarely have one. Fix the internal links in both directions, particularly pointing newer related pages back at this one.

Then set the revision date honestly, because at that point it is true.

How not to lose what already works

The real risk in refreshing is destroying a page that was ranking for reasons you had not identified.

Before editing, record what the page currently ranks for, including the queries you never targeted — pages frequently earn traffic from something other than their intended keyword, and rewriting blindly removes the passage responsible. Keep the URL unchanged unless there is a compelling reason, since changing it forfeits accumulated authority for a marginal gain. Preserve the sections that are working rather than rewriting wholesale. And if a page is significant, change it in stages so you can attribute the effect.

Then wait properly before judging. Meaningful re-evaluation takes weeks, and reverting after five days because rankings wobbled is how businesses conclude that refreshing does not work.

A maintenance rhythm that survives contact with a real workload

Refresh by decay signal, not by calendar. Quarterly, pull the pages that lost the most year-on-year impressions and take the top handful. That is a manageable amount of work that targets the pages where recovery is worth most.

Two exceptions worth diarising regardless. Anything carrying prices, statistics, or regulatory dates needs a check every six to twelve months because its facts expire on a schedule — that includes the trend-led material on this blog, whose numbers decay faster than its arguments. And anything that has become commercially important since publication deserves attention whether or not it is declining.

The underlying shift is one most content strategies have not made: the value is in a maintained library, not an accumulating archive. Twenty pages kept genuinely current outperform two hundred abandoned ones, and they are cheaper to run. If you are unsure which of yours are decaying, our diagnostic order of operations for search visibility separates this from the other three reasons traffic disappears, and we are happy to look at your Search Console and tell you which pages are worth the effort.