Most content does not fail loudly. It fades. A page that brought in steady visits for a year slowly slips down the results, and by the time your traffic report shows it, the damage has been building for weeks. The good news is that there is usually an earlier signal, and it is easy to watch for.

What content decay is

Content decay is the slow loss of search performance on a page that used to do well. It happens for ordinary reasons. The information gets old. Competitors publish something fresher or more complete. Search results change shape. The questions people ask shift.

No single event causes it, which is exactly why it goes unnoticed. Nothing breaks. There is no error message. The page just earns a little less every week.

Why traffic is a late signal

Clicks only happen after a page shows up in the results and someone chooses it. So there are two steps before a click, and the first step moves first.

Google Search Console reports impressions, which is how often your page appeared in results, along with clicks and average position. When a page starts losing ground, it often shows up as fewer impressions or a worse position before it shows up as fewer clicks. That is what we mean by impressions move first. It is a pattern we rely on, not a guarantee, so treat it as an early warning and not as proof.

Traffic tells you what already happened. Impressions tell you what is starting to happen.

What the watcher does

The watcher is a simple agent that runs on a schedule, usually weekly. Search Console has an API, so the data can be pulled automatically instead of someone opening a dashboard. The steps are straightforward.

  1. Pull the data. Impressions, clicks and average position by page, and by search query for each page.
  2. Compare two windows. For example, the last four weeks against the four weeks before. Choose windows long enough to smooth out daily noise.
  3. Flag the pages that are slipping. A page is flagged when impressions or position drop by more than a threshold you set.
  4. Filter out the noise. Ignore very small pages, brand new pages, and anything with a known seasonal pattern.
  5. Rank by what matters. A decline on a page that drives leads is more urgent than a decline on a page nobody reads.

We do not have a universal threshold to give you. A good starting point is whatever level you would care about for your own site, and then you adjust it after a few weeks of seeing what gets flagged. Treat any number you pick at first as a guess to tune.

From flag to brief

A list of slipping pages is useful, but a brief is better. For each flagged page, the agent can prepare a short brief that a writer can act on right away:

  • The page, its main topic and what it is meant to do for the business.
  • The queries where it lost impressions or position, and by how much.
  • What currently ranks for those queries, and how those pages differ.
  • Parts of your page that may be out of date, such as old facts, old tools, missing sections or broken links.
  • A suggested set of changes, in priority order.

The agent does the digging. A person still decides what is worth changing. Agents are good at gathering and comparing, and less reliable at judging what your audience will find useful, so keep the final call with someone who knows the business.

Things that can fool the watcher

  • Seasonality. Some topics rise and fall every year. Compare against the same period last year where you can, not only the previous weeks.
  • Tracking changes. A site migration, a new page structure or a change in how pages are indexed can look like decay when it is not.
  • Search result changes. New features on the results page can reduce clicks or shift impressions without any fault of your page.
  • Small numbers. A page with a few dozen impressions will swing wildly. Set a minimum before you trust a trend.
  • Data delay. Search Console data arrives with a short lag, so do not read the most recent days as final.

Close the loop

The watcher is only half the system. When a page gets refreshed, log what was changed and the date. Then have the watcher look at that page again a few weeks later. Did impressions and position recover? If yes, you have learned what works for your site. If no, you have learned the cause was something else.

Over time that log becomes a record of which kinds of updates actually move results for you, which is far more useful than general advice.

Where to start

You can begin without any automation. Open Search Console, compare the last four weeks with the four before for your top twenty pages, and write down which ones slipped. Pick two and work out why. Once you have done that by hand, you will know exactly what you want the agent to do, and it will be much easier to write the spec.