Update an old post only if it meets two conditions at once: it has rankings to lose (it shows up in Google, even on the second page) and something real has changed in the topic, in your catalog, or in search intent. If the first fails, the problem isn't the content. If the second fails, you're repainting a wall that's perfectly fine.
Summary
- Age is a bad criterion. The publication date says nothing about performance: there are posts from four years ago still bringing in visits and posts from last year that no one has read.
- The most profitable material sits between positions 5 and 20. It has already cleared the hard part —it exists, it's indexed, and Google considers it relevant— so moving it up costs less than starting from scratch.
- There are four possible decisions, not one. Update, merge, reconnect to the catalog, or leave alone. Choosing wrong costs more than doing nothing.
- If a post leads to no product page, updating it is makeup. Fix the commercial route first; the text comes later.
- Changing only the date doesn't work. Google compares the content, not the label, and the reader who expected this year's data leaves.
Why age is a bad criterion
The usual way to approach this is to sort your articles by date and start with the oldest. It's convenient and it's wrong. The publication date doesn't tell you whether an article works: there are posts from four years ago still bringing in visits every month and posts from last year that no one has read.
What does matter is the trajectory. An article that was at position 6 and is now at 14 is warning you about something: someone wrote something better, or the question changed, or your content aged badly. That's the post that deserves your time, whatever its age.
The cheapest material an online store has is the content that's already indexed and nobody has looked at again.
The data you need before touching anything
This part can't be skipped, and it's the reason most update plans fail: they're run by guesswork. You need three things from Google Search Console, all free:
- Average position by URL, comparing the last three months with the three before that. That's where you see who's dropping.
- Impressions and clicks. Lots of impressions with few clicks is a title-and-description problem, not a content one — and it's fixed in ten minutes.
- The actual queries each post shows up for. There are almost always surprises: people arrive searching for something you didn't write.
Without this data you can write better, but you can't prioritize. And prioritizing is the whole game when you have forty articles and time for three.
The decision tree
For each post you're thinking of touching, in this order. The first answer that applies decides.
What "updating" really means
Updating isn't adding two paragraphs at the end. If the post passed the decision tree, this is what you need to review, more or less in order of impact:
1. Intent, before anything else
Look at the actual queries the article shows up for and compare them with what you wrote. If people arrive searching for a comparison and you wrote a definition, no amount of style tweaking will save you: the piece has to be refocused.
2. The title and the meta description
It's the cheapest thing and usually the most profitable. If you have impressions but few clicks, your problem is here. And make sure the Google title and the page's H1 say the same thing: when they contradict each other, Google rewrites the title on its own and almost never gets it right.
3. What's no longer true
Prices, years, deadlines, screenshots of interfaces that changed and —the most common one in stores— products you mention and no longer sell. Those broken links to discontinued product pages are a sign of neglect for the reader and for Google.
4. Structure, so they can cite it
This point didn't exist three years ago and today it decides quite a lot. ChatGPT, Perplexity and Google's AI summaries need to be able to extract a concrete answer from your page. An article with the direct answer up top, headings in the form of questions and a frequently-asked-questions section gets cited. A wall of text with no hierarchy doesn't.
5. The links to the catalog
We leave it for last on the list but it's first in importance if you run a store. Every article should lead the reader to the product page or category it belongs to, from inside the text and at the moment it makes sense. Not with a banner at the end that no one looks at.
If you want your store's actual number before you start, our commercial linker test goes through your latest posts and tells you how many lead to a product page or category. Free, no sign-up.
The most repeated mistake: changing the date
Changing the publication date without touching the content doesn't work. Google compares the content, not the label, and what you get is worse than nothing: the reader arrives expecting this year's information, finds data from three years ago, and leaves. That bounce confirms the page doesn't answer.
Update for real and then, yes, change the date. In that order.
What this looks like in a real store
Imagine a blog with forty-seven articles. Applying the decision tree, a typical split ends up looking more or less like this:
| Situation | Decision | Effort |
|---|---|---|
| Ranks between 5 and 20, and lost positions | Update | Medium — this is where the return is |
| Two or three articles on the same topic | Merge into one and redirect | Medium |
| Brings in visits but links to no product page | Reconnect to the catalog | Low — the best return per hour |
| Stable and bringing in visits | Leave alone | Zero |
| Zero impressions and no demand for the topic | Delete or merge | Low |
Look at the third row. It's usually the most common one in an online store and the least worked on, because it doesn't look like an SEO problem. And it's exactly the one that separates a blog that entertains from a blog that sells.
Frequently asked questions
How often should you update a blog's content?
There's no frequency that works for everyone. What works is reviewing your entire inventory twice a year and updating only what meets the two conditions from the start. A post that holds steady at position 4 doesn't need you to touch it.
Is updating an old post better than writing a new one?
Almost always, if the post already ranks. An article on the second page has already cleared the hard part: it exists, it's indexed, and Google considers it relevant for something. Moving it up costs less than starting from scratch and the result comes sooner.
Do you need to change the URL when updating?
No. If the post already has rankings and links, changing the URL throws away that history. You update while keeping the URL. The exception is merging several articles into one: there you do need 301 redirects from the old ones to the one that survives.
What do I do with a post that gets no visits at all?
It depends on why it gets none. If nobody searches for that topic, delete it or merge it with another. If the topic does get searched but your article doesn't show up, the problem isn't the content: either it has no internal links reaching it, or it competes with another page of yours.
Want to know which of your posts fall into each box?
We send you your blog diagnosis by email: what's losing traffic, what's cannibalizing, how many posts lead to your catalog, and where we'd start. Free, for the first thirty.
Get my diagnosis