Publishing feels like progress. Maintenance feels like admin. That single bias explains most of the flat traffic charts we look at when a new customer connects their Search Console.
The arithmetic is unforgiving. A site publishing four articles a month might add a few hundred clicks. The same site, left alone, will lose a comparable number across the two hundred pages it published in the previous three years. Net zero, four articles a month, and a team that cannot understand why the line does not move.
What decay actually looks like
Decay is rarely a cliff. A page that fell off a cliff was hit by something structural: a redirect gone wrong, a noindex left in a template, a competitor who published something genuinely better. Those are incidents and you find them the same week.
Real decay is a slope. Position drifts from four to six over five months. Click through rate falls because the result above you now has a rich snippet and yours does not. The page still ranks. It simply earns less every month, and no dashboard alarm ever fires, because nothing broke.
Finding it without drowning in dashboards
The comparison that matters is the same page over two equal windows, usually the last twenty eight days against the same twenty eight days a year earlier. Year over year rather than month over month, because month over month is mostly seasonality wearing a disguise.
For every page, pull clicks, impressions and average position for both windows. Then split the losers into three groups, because the fix is different for each.
- Impressions down, position stable. Demand fell or the query itself changed shape. Often nothing to fix. Check whether the query now returns a different page type before you spend an hour.
- Position down, impressions stable. Someone out ranked you. This is the group that responds to a real content update.
- Position and impressions stable, clicks down. You lost the click, not the ranking. An AI overview, a featured snippet or a competitor with a better title took it. Rewrite the title and description first, it is twenty minutes and it works more often than it has any right to.
Deciding what to fix
Not every decaying page deserves rescue. We score each candidate on three things: the clicks it lost, the position it holds today, and whether the topic still matters to the business. A page that slipped from three to five on a commercially relevant query is worth an afternoon. A page that slipped from thirty to forty on a query nobody buys from is worth deleting.
Deleting is a legitimate outcome and teams under use it. Pages nobody reads still consume crawl budget, still dilute internal links, and still make the site look thinner than it is. Merge the salvageable ones into a stronger page, redirect them, and move on.
Make it a habit, not a project
One afternoon a month, one person, the top twenty decaying pages. That is the whole programme. Teams that do this consistently spend roughly a quarter of their content time on maintenance and grow faster than teams that spend all of it publishing.
It is not glamorous work. It is the highest return per hour available to almost every content team we meet.