SEO & Content StrategyContent refresh, consolidation and pruning · Lesson 13 of 17
Content decay and the refresh playbook
Video lecture
Content decay and the refresh playbook
The narrated lecture is in production
Every chapter is scripted and ready. Browse the chapters and read the full transcript now — the video will appear here when it’s published.
Chapters
Transcript of the narration, chapter by chapter.
0:00 Content decay and refresh
Your best article from two years ago used to bring in hundreds of visits a week. Now it brings in half that. Nobody noticed, because everyone was busy publishing new content. That's content decay, and on established sites it's often where the biggest, cheapest wins are hiding. In this lecture you'll learn how to spot decay, how to diagnose its cause, a nine step refresh playbook, and a Python script to find decaying pages.
0:32 Why it matters
Why does this matter? Because an existing page already has history, links and relevance. Refreshing it well is often faster and cheaper than creating something new, and the return can be bigger. Yet most content plans are all new content and no maintenance. It's like a shop that keeps adding new shelves while the best selling shelf gathers dust.
0:58 Diagnose first
Here's the key idea. Decay has different causes, and each needs a different fix. Think of a doctor. A cough could be a cold, an allergy or something more serious. You don't prescribe until you diagnose. The same with content. Is it a SERP change, competitors improving, demand falling, intent shifting, or a technical problem? Search Console patterns give you strong clues.
1:25 Read the pattern
Here are the patterns. Impressions stable, position stable, but clicks and CTR down: probably a SERP change, like more features or AI answers satisfying the query. Impressions and position both down: competitors improved or your content is outdated. Impressions down but position stable: demand fell, so check Google Trends. Queries shifting to a different intent: the page needs reworking. And a sudden drop across many pages: check technical issues and the timing of algorithm updates.
1:58 Compare like for like
One rule before any diagnosis: compare like for like periods. The last three months against the same three months last year, not against the previous three months. Otherwise seasonality looks like decay. A Ramadan recipe page will always fall after Eid. That's not decay. That's the calendar.
2:18 The 9-step refresh
Now the refresh playbook, nine steps. Re read the SERP. Update facts against sources. Add what others lack: first hand examples, data, photos, a calculator or an expert quote. Improve structure with answer first summaries and clearer headings. Tighten by removing outdated or padded sections. Refresh internal links, to newer pages and from newer pages back. Update the title and description if needed. Update the modified date honestly, only for material changes. And request indexing for important pages.
2:52 Honest dates
A word on dates. Google advises against changing dates to make pages look fresh without substantial changes. Readers notice too. A page that says updated this month but still quotes last year's prices destroys trust instantly. So only show a new last updated date when the content has materially changed, and make sure the structured data's date modified matches.
3:18 Example 1 (simple)
Worked example one, simple. A recipe blogger's Ramadan iftar ideas page falls every year after Eid. She compares year on year and sees that the peak this year was actually higher than last year. There's no decay. Instead, she schedules a refresh six weeks before next Ramadan: new photos, two new recipes she's tested, and updated links to her newer posts.
3:45 Example 2 (illustrative)
Worked example two, realistic and illustrative. A UK tax advice blog's self assessment deadlines guide loses a third of its clicks year on year, while impressions are flat and position unchanged. The SERP now shows an AI Overview answering the basic deadline question. So the team refreshes around what still earns a click: a downloadable deadline calendar, worked examples for first time filers, a penalties calculation explained step by step, and a chartered accountant's tips, with every date checked against official guidance. Clicks partially recover, and enquiries rise.
4:23 Find decay with Python
How do you find decaying pages at scale? The Python script in the lesson loads two Search Console page exports for comparable periods, merges them, calculates the change in clicks, and labels likely patterns: a CTR drop suggesting a SERP change, or a ranking loss. It filters pages that had meaningful traffic and lost at least a quarter of their clicks, and saves refresh candidates. The labels are prompts to investigate, not diagnoses.
4:55 Diagnose this
Let's pause and diagnose a real looking case. A UAE car rental company's page on driving licence rules for tourists loses forty percent of clicks year on year. Impressions are also down forty percent, but position is stable. What does that suggest? Demand fell, not ranking. Checking Google Trends confirms fewer searches this year for that topic in the UAE. So the right action isn't a big rewrite. It's a light accuracy check, and moving effort to topics where demand is growing. Diagnosis saved them a wasted rewrite.
5:33 Watch me do it: find + diagnose decay
Watch me do it. I'll find and diagnose decaying pages for a UK tax advice blog. First, in Search Console, I export the Pages table for June to August this year, and again for June to August last year. Same months, so seasonality is controlled. Next, I run the Python script from the lesson. It merges the two exports, calculates the click change, and labels patterns. Fifteen pages lost at least a quarter of their clicks. Then I read the labels. The self assessment deadlines guide shows a CTR drop, impressions flat, position stable. I search the main query, and an AI Overview answers the basic deadline question. So the fix isn't rewriting the basics. It's adding what still earns a click. A second page, on expenses for landlords, shows a ranking loss. I compare it with the pages that now outrank it: they're more current and have worked examples. That's a substance refresh. A third page lost impressions but kept its position. Google Trends shows lower interest this year, so I leave it. Finally, I schedule the two refreshes, apply the nine step playbook to the first one, and write the refresh date in the annotation log, so I can measure the result in four to eight weeks.
7:04 Refresh calendar
Make it routine with a refresh calendar. Tier one pages, revenue or high traffic, get reviewed quarterly. Tier two, twice a year. Tier three, annually, or when a trigger fires, like a law change or product update. Seasonal pages get refreshed six to eight weeks before their peak. Common mistakes: refreshing by changing the date and a few words, treating seasonality as decay, rewriting a page that still performs just because it's old, and forgetting links from newer pages.
7:38 Recap
Recap. Content decay is normal. Diagnose the cause with Search Console patterns and like for like comparisons before acting. A real refresh updates facts, adds original elements, improves structure, tightens and refreshes internal links, with honest dates. And a tiered calendar keeps it routine. Try this now. Compare this quarter with the same quarter last year in Search Console, find five pages that lost the most clicks, and label each with a likely cause.
What content decay is
Content decay is the gradual decline in a page's search performance after it peaks. It is normal: facts date, competitors publish better pages, intent shifts, SERP features (including AI Overviews) change how many people click, and the page itself accumulates broken links or outdated screenshots. On many established sites, refreshing existing pages is one of the highest-return activities available — the page already has history, links and relevance.
Diagnose before you refresh
Decay has different causes, and each needs a different fix:
| Pattern in Search Console | Likely cause | Fix |
|---|---|---|
| Impressions stable, clicks and CTR down, position stable | SERP change — more features or AI answers satisfying the query | Improve title/snippet appeal; add depth, tools or first-hand elements that earn the click; consider targeting related sub-questions |
| Impressions and position down | Competitors improved, or content outdated | Refresh substance: current facts, better structure, original elements |
| Impressions down, position stable | Demand fell (seasonality, trend) | Check Google Trends; maybe nothing to fix |
| Queries shifted to a different intent | Intent change | Rework the page or create a better-matched page |
| Sudden drop across many pages | Technical issue or site-wide change | Check indexing, robots, canonicals, redirects; algorithm update timing |
Always compare like-for-like periods (for example the last 3 months vs the same 3 months last year) to avoid mistaking seasonality for decay.
The refresh playbook
- Re-read the SERP. What ranks now? Has the dominant type or format changed? What would still earn a click?
- Update facts. Prices, laws, dates, product details, screenshots — verified against sources.
- Add what others lack. New first-hand examples, data, photos, a calculator or checklist, an expert quote.
- Improve structure. Answer-first summaries, clearer headings, tables where they help.
- Tighten. Remove outdated or padded sections. Longer is not better.
- Refresh internal links — to newer pages, and from newer pages back to this one.
- Update on-page elements — title and meta description if they no longer reflect the content.
- Update
dateModifiedhonestly. Google advises against changing dates to appear fresh without substantial changes; show a "last updated" date only when the content materially changed. - Request indexing for important pages via URL Inspection, and annotate the change date.
Hands-on: find decaying pages with Python
Using two Search Console exports (Pages tab) for comparable periods, or the API:
import pandas as pd
prev = pd.read_csv("pages_2025-06_to_2025-08.csv").rename(columns=str.lower) # columns: page, clicks, impressions, position
curr = pd.read_csv("pages_2026-06_to_2026-08.csv").rename(columns=str.lower)
df = prev.merge(curr, on="page", how="outer", suffixes=("_prev", "_curr")).fillna(0)
df["click_change"] = df["clicks_curr"] - df["clicks_prev"]
df["click_change_pct"] = df["click_change"] / df["clicks_prev"].where(df["clicks_prev"] > 0)
df["pattern"] = "other"
df.loc[(df["impressions_curr"] >= 0.9 * df["impressions_prev"]) & (df["click_change_pct"] <= -0.25), "pattern"] = "CTR drop (SERP change?)"
df.loc[(df["impressions_curr"] < 0.75 * df["impressions_prev"]) & (df["position_curr"] > df["position_prev"] + 2), "pattern"] = "ranking loss"
decaying = df[(df["clicks_prev"] >= 50) & (df["click_change_pct"] <= -0.25)]
print(decaying.sort_values("click_change").head(25)[["page", "clicks_prev", "clicks_curr", "pattern"]])
decaying.to_csv("refresh_candidates.csv", index=False)(Thresholds are illustrative. Note that position values from exports are averages; treat the pattern labels as prompts to investigate, not diagnoses.)
Refreshing for AI answers without chasing gimmicks
When a query now shows an AI Overview, the refresh question is "what does our page offer that a summary can't?" Useful answers include a downloadable checklist or calendar, a calculator, current local prices, original photos, a worked example with real numbers, or a named expert's judgement on trade-offs. Google's May 2026 guidance says you don't need to rewrite pages into "AI-friendly" fragments; a clear answer-first structure plus genuinely original elements serves both readers and AI features.
Measure the refresh
Record the refresh date in an annotation log, then compare the 28 days after re-indexing with the 28 days before and with the same period last year. Watch impressions, clicks, CTR and the mix of queries (new sub-questions appearing is a good sign), plus conversions in GA4. Give it 4–8 weeks before judging, and resist refreshing the same page again before then.
A refresh calendar
Tier pages: Tier 1 (revenue or high-traffic pages) reviewed quarterly; Tier 2 twice a year; Tier 3 annually or when a trigger fires (law change, price change, product update). Seasonal pages get refreshed 6–8 weeks before their peak.
Worked example (illustrative)
A UK tax-advice blog's "self-assessment deadlines" guide loses a third of its clicks year-on-year while impressions stay flat and position is unchanged. The SERP now shows an AI Overview answering the basic deadline question. The team refreshes the guide around what still earns a click: a downloadable deadline calendar, worked examples for first-time filers, a penalties calculator explained step by step, and a named chartered accountant's tips — with every date checked against GOV.UK. They rewrite the title to promise the calendar and examples. Clicks recover partially, and time on page and enquiries rise.
Common mistakes
- Refreshing by changing the date and a few words.
- Treating seasonality as decay.
- Rewriting a page that still performs because it's "old".
- Forgetting internal links from newer pages.
Key takeaways
- Content decay is normal; diagnose the cause (SERP change, competitors, demand, intent, technical) before acting.
- Compare like-for-like periods to separate decay from seasonality.
- A real refresh updates facts, adds original elements, improves structure, tightens and refreshes internal links.
- Only update dates when content materially changes; tier pages into a refresh calendar.
Check your understanding
Quick questions to lock in the lesson. They don’t count towards your certificate.
Put it into practice
Export two comparable periods of page data from Search Console, find five decaying pages, label each with a likely cause, and write a refresh plan for one using the nine-step playbook.
Enrol for free to save your progress
Reading is always free. Enrol to keep your place, take the final assessment and earn a verifiable certificate.