The article that brought 3,000 clicks last year only has 600 this month. It wasn’t deleted, and it took no penalty — it just walked onto the downhill section of the content decay curve. Judging when an old article should be refreshed and how deep to go comes from reading this curve, not from gut-feel rewrites. Evergreen content isn’t naturally decay-proof either — the evergreen-vs-trending comparison covers that.
Here’s the bottom line: content decay has four causes, and the treatment differs completely — seasonal fluctuation means just wait, competitive replacement means go deeper, stale information means swap the data, technical breakdown means fix the page. First use three signals — click month-over-month, ranking range, and click-through rate — to locate the cause, then decide between small fixes, rewrites, or merges. Rewriting before you’ve nailed the cause is the most common form of wasted work in a content team.
What a typical curve looks like
Most informational articles’ lifecycles fit into four stages. Knowing which stage you’re in tells you whether to wait or act.
| Stage | Approximate timing | Typical behavior | What to do now |
|---|---|---|---|
| Climbing | 0–90 days after launch | Ranking crawls forward from outside 30, clicks unstable | Don’t touch the body; add internal links and distribution so crawling keeps up |
| Plateau | Months 3–12 | Ranking stable in top 10, clicks form a baseline | Record the baseline as the reference for detecting decay later |
| Declining | Months 12–24 | Ranking slides to 10–20, clicks drop 30%+ | Locate the cause, do one refresh with real substance |
| Long tail | After 24 months | Low but stable traffic, occasionally triggered by long-tail words | Assess merge value — is it worth maintaining on its own |
The curve’s shape correlates strongly with topic type. Tool-review articles may have a plateau of only six months; methodology articles can hold for three-plus years. For your own site’s real curve, export two years of monthly clicks and draw it — that’s more accurate than any industry average.
Four causes, four treatments
- Seasonal fluctuation: year-over-year it’s not down at all, just lower than last month. Do nothing; just distribute a month ahead.
- Competitive replacement: your ranking barely moved, but several newer, more complete pages squeezed in ahead. What’s missing is depth and exclusive data, not a new title.
- Stale information: prices, version numbers, screenshots, and policies in the article no longer match reality. Readers see through it instantly, bounce rate rises first, ranking falls after — swapping the data is the only fix.
- Technical breakdown: a redesign dropped internal links, images broke, the page got slower, or it was accidentally noindexed. This loss is the most unjust, and the easiest to fix in a day.
Pinpointing the cause takes two steps: read the article side-by-side with the current top three search results — if the gap is obvious, it’s competitive replacement or stale info; if the content looks fine, check crawl logs and page status — likely technical. Both steps together take under 15 minutes and can save a whole day of ineffective rewriting.
Three signals for deciding whether to act
- Click month-over-month: clicks in the last 90 days down 30%+ vs the prior 90 days — add to the watchlist; below 15% — just observe.
- Ranking range: dropped from top 10 to 11–20 — a salvageable range, prioritize it; straight to beyond 50 — intent has likely changed, consider rebuilding rather than patching.
- Click-through rate: impressions barely moved but CTR halved — the problem is title and snippet, not the body. Rewrite the meta description per the CTR approach for titles and meta descriptions; an hour’s work.
Read the three signals together. Watching clicks alone wrongly penalizes topics where overall search demand is shrinking — if impressions and clicks drop together while ranking doesn’t move, the topic itself went cold and no refresh can save it; moving resources to another topic pays off better.
Grade the refresh, don’t jump straight to rewriting
| Level | Trigger | Workload | Reasonable expectation |
|---|---|---|---|
| Small fix | Stale info, CTR drop | Within 1 hour | Recover 60–80% of original clicks |
| Medium fix | Competitive replacement, gap is in breadth | Half a day, add 2–3 sections | Ranking back to top 10 |
| Rewrite | Stale structure, shifted intent | 1–2 days, keep URL, rewrite 60%+ | Re-enter the climbing stage |
| Merge | Two articles heavily overlapping, both past 20th | Half a day, plus a 301 | Consolidate authority — one article doing the work of two |
At any level, don’t change the URL. Change the URL and all accumulated backlinks and history signals have to go through redirects again — the cost dwarfs the structure you wanted to optimize. Use the old-article refresh checklist for the execution list, ticking each item to avoid missing internal links and structured data.
Build a monitoring rhythm that doesn’t need daily babysitting
- Export the site-wide last-90-days data on the 1st of each month, compute month-over-month with one template, and have a list in five minutes.
- Set a threshold alert: clicks down over 30% and ranking out of top 10 automatically enters the to-do list.
- Handle quarterly in batches: classify the to-do by the four levels above and schedule 8–12 at once — far more efficient than scattered patching.
- Archive every refresh: change date, what changed, 30-day click comparison before and after — after three months you’ll have your own refresh-return table.
- Feed the archive into the content ROI data model, and work out how many clicks each hour of labor buys back — that’s your evidence the next time you request budget.
The update cadence itself needs restraint too. Revising the same article repeatedly within one quarter makes it impossible to tell which change worked and tends to loosen the structure — the controlled experiments in the content-refresh-frequency study show you need at least a 30-day observation window after a substantive change to judge anything.
You can start today: export the last 180 days of data, compute click month-over-month, and pick the top 10 articles down over 30%; judge each of the 10 for its cause, spending only 15 minutes per piece to classify without acting; finish all small-fix articles this week first — lowest cost, fastest effect; add a fixed task for the 1st of each month to the calendar, turning the data pull into a process rather than an inspiration.
Drawing the decay as a chart
An article’s organic traffic usually isn’t a plateau — it spikes after publishing, then declines month by month. The illustrative curve below helps you judge: once it drops to a certain threshold, it’s time to refresh, not wait until it hits zero. The slope differs by industry, but the “it will decay sooner or later” shape is universal.
