LotsBlog
You already published the post. It ranked, you moved on, and now it sits in the archive doing nothing while you write this week's new piece. Most operators eventually notice the same quiet problem: a 2023 article is still showing up on page two, slowly decaying, and a fresh post takes months to reach the same position it already had.
That gap — between what the archive could be doing and what it's actually doing — is exactly what a refresh workflow is for. Not a one-off checklist, not a quarterly cleanup. A loop you actually run.
This is the operator version: pick by signal, update substance, restructure for scannability, re-link and re-distribute. Run it monthly and the archive starts compounding again instead of going stale.
Why refreshing beats writing new
A new post starts at zero. It needs links, time, and Google's patience before it earns a real position. A refreshed post already has a foundation: it has indexed URLs, some backlinks, internal-link weight, and a history of clicks. You're not starting over — you're rebuilding on top of authority that already exists.
The decay math is real. Animalz tracked AdEspresso data showing aged content decays at an average of roughly 1.21% per week. That sounds small until it compounds across a year. Meanwhile, Siege Media found that page-one ranking content for popular keywords gets updated within the last two years on average — so the bar for staying on page one is "looks fresh," not "was correct once."
A refreshed post often ranks faster than a brand-new one covering the same topic. That's the case for running the loop, not just writing more.
Step 1 — Pick by signal, not by gut
Most "refresh your old posts" guides start with "update your top 10 posts." That's gut. You want signal.
Open Google Search Console and pull the last 16 months of data. You're looking for three patterns:
- Posts ranking 8–20 that already have impressions. These are the easiest wins. They're one or two solid improvements away from page one.
- Posts with decaying click-through rate over the last 6 months. The post still ranks, but fewer people click. The title or answer-first paragraph has gone stale.
- Posts with stable impressions but dropping average position. A small position drop on page one usually means competitors have added a section, a stat, or a new angle you don't have.
Skip posts with no impressions at all. Without a ranking foundation, a refresh is just rewrite work — and you'll have a hard time telling if it moved anything.
One caution worth naming: Google's Search Console reporting changed materially in 2025 — including a confirmed impression-counting error. Compare like-for-like time windows and don't read a single month in isolation. Use 3- to 6-month deltas, not weekly blips.
Step 2 — Update the substance
Now the real work. Substance, not just dates.
Start with the answer-first paragraph. If the post opens with a definition or a generic intro, rewrite it as a direct answer to the question the reader actually typed. This is also where you tighten the post's promise — make the first two sentences say exactly what the reader gets and why it's still worth their time.
Then walk the body and look for four things:
- Stats older than 12 months. Replace them with current figures and link the source inline. If a stat is a one-off number you can't verify anymore, drop it.
- Anything factually outdated. Tool names, pricing, screenshots, product features, "as of 2023" framings. Anything that screams "old post" in the first scroll.
- Sections that competitors now cover and you don't. Look at the top three results for your target query. If they all have a "common mistakes" section and you don't, that's a gap — add it.
- Broken or weak internal links. Old posts often link to posts you've since updated, retired, or moved. Replace them with the current best fit in your archive.
One pass, in this order, is faster than three passes of polish.
Step 3 — Restructure for scannability
Even when the substance is right, a 2023 post is usually a wall of H2s and 800-word paragraphs. Readers don't trust that anymore. Restructure for the way people actually read now.
- Rewrite H2s as promises, not topics. "Email Deliverability" becomes "Why Your Emails Land in Spam (And How to Fix It)." The heading should answer "what do I get if I keep reading?"
- Break long paragraphs into lists or short tables. Anywhere a paragraph is making a list of things — steps, examples, comparisons — make it a list. Anywhere it's comparing two things — make it a table.
- Add an FAQ section. Three to five questions real readers ask about this topic, with direct answers underneath. Use the exact phrasing people use in search. This is also the easiest way to pick up long-tail traffic the original post missed.
- Tighten the meta title and description. Make the title carry the actual hook — not a generic category name. Keep the meta description under 160 characters and write it as the answer to the query.
The point isn't to make the post shorter. It's to make every scroll worth the reader's time.
Step 4 — Re-link and re-distribute
The post is updated. Now make sure the rest of the system knows.
Re-link inside the cluster. Add 2–4 new contextual internal links from related posts into the refreshed piece, and from the refreshed piece into 2–4 newer posts that didn't exist when this was first written. The compound effect of internal linking is the whole point — a refreshed post that nobody links to is a refreshed post that ranks like an orphan. If you run a 30-minute blog audit on the same cluster, this is where that work pays off.
Update the date signals. Bump the schema dateModified so Google sees the post as current. Add a small "Last updated" line near the top. These aren't ranking tricks, but they set reader expectations honestly.
Re-distribute. Push the refreshed post through the same channels a new post gets:
- One newsletter mention in the next send — even a single line with the new angle.
- One social post that points at the new section, not the post as a whole.
- One internal CTA in a related, currently-trafficked post.
The goal isn't to pretend this is a new post. It's to give the new readers a reason to land on it, and to give the post a fresh signal of life.
The 90-day cadence
A refresh workflow only matters if it runs. Here's the loop that fits inside an hour a month:
- Monthly (30–45 min): Pull Search Console for the last 90 days. Pick three posts with the patterns from Step 1. Add them to a refresh queue.
- Refresh each in 45–60 min. That's the four steps above, top to bottom. Most posts don't need a full rewrite — they need the answer-first paragraph, one section added, and the links cleaned up.
- Quarterly (60 min): Run the topic cluster audit on your top three clusters. Identify posts that should be retired (less than 50 impressions over six months, no clear query fit) and posts that should be merged into a stronger pillar piece. Retirement is part of the loop, not a failure.
- Twice a year: Re-read your top 20 posts by traffic. Not to refresh — to make sure they still match what the blog is actually about. A blog that compounds stays in shape because the operator keeps checking the shape.
If you skip a month, fine. If you skip two, the queue grows and the loop dies. The cadence matters more than the volume.
Where an agent fits in the loop
A refresh workflow is the kind of recurring, rule-based work that eats operator time without needing operator judgment. That's exactly where a system-grade agent earns its keep.
The agent does the recurring work:
- Flags posts in the queue when Search Console signals trigger Step 1.
- Drafts the substance update for each post — fresh stats, tighter answer-first paragraph, the new section that closes a gap.
- Suggests the new internal links from your current archive.
- Queues the re-distribution steps: newsletter mention, social post, internal CTA.
The operator stays in the loop:
- Approves which posts make the queue in the first place.
- Reviews the substance draft before it goes back on the page.
- Approves the republish. Nothing goes live without the human approval step.
This is the difference between a system that runs your blog and a tool that just generates text. The draft-review-approve rhythm is the rhythm. The agent surfaces the work; you decide what ships.
If you want the writing side of the loop to follow the same system, the operator's playbook for writing a blog article that compounds is the natural pair to this piece.
FAQ
How often should I refresh old blog posts?
For fast-moving topics — SEO, AI tools, marketing — every 3 to 6 months. For evergreen content, a yearly pass is enough. The real signal is your Search Console, not a calendar: refresh a post when its impressions or position starts to drop, not because it's been six months.
How do I know which posts to refresh?
Pick posts ranking between positions 8 and 20 that already have impressions, posts with decaying click-through over six months, and posts whose average position is slipping on page one. If a post has no impressions in 16 months, a refresh probably won't move it — consider retiring it or merging it into a stronger pillar post instead.
Does refreshing a post hurt its rankings?
Done right, no. A real refresh — new substance, better structure, current data, cleaned-up links — usually lifts rankings. The risk is cosmetic refresh: bumping the date and changing a word. Google sees through that. Treat the refresh as a real edit or skip the post entirely.
Ready to run this loop on your archive? Start your free blog → lots.blog