Evergreen or Expiring? Decide Before You Commission

Most businesses commission content the same way they commission tea for a meeting: someone asks for four pieces this month, four get delivered, nobody checks what happened afterwards. Six months later the blog has twenty-four posts, two of them bring visitors, and nobody can say in advance which two it was going to be.

You can usually say in advance. Not perfectly, but well enough to decide how much a piece is worth and who should write it. The question is not is this good content. It is when does this stop being true.

The only test that matters

Before you approve a topic, ask one thing: if someone reads this page three years from now, what in it will be wrong?

Then be honest about the answer. Three kinds come back.

Nothing will be wrong. "How to check whether a transformer is undersized for your load" will read the same in 2029. Physics does not get revised. This is durable content.
The details will be wrong but the structure will hold. "GST on restaurant services, explained" will need the rates checked and perhaps a paragraph rewritten, but the page is still the page. This is maintainable content — durable with a service interval.
The whole thing will be pointless. "Our stall at the Delhi trade fair this February" is dead on 1 March. So is "what the Budget means for exporters", nine days after the Budget. This is perishable.

That is the entire framework. Everything below is about what to do once you have sorted a topic into one of those three buckets.

What durable content actually looks like

Durable content is nearly always one of four things, and it is worth knowing them because they are the things small businesses are best placed to write and most reluctant to.

The decision page

Someone is choosing between two options and does not know how. Powder coating or anodising. Modular kitchen in plywood or in HDHMR. Lease the machine or buy it. These searches never stop, because there is always a new person making the decision for the first time. Nobody outside your trade can write this page properly, which is exactly why it ranks.

The specification page

Sizes, grades, load ratings, compatibility, what fits what. Boring to write, endlessly looked up. If a customer has ever rung you to ask which size they need, that call is a page.

The diagnosis page

"Why is my RO unit making a knocking sound." "Why does the invoice show a different taxable value than my PO." People search symptoms. If you fix things for a living, you have thirty of these in your head and none of them on your website.

The local reference page

"Fire NOC process for a commercial kitchen in Delhi." Partly perishable — the forms change — but a page that is genuinely about your city, kept current, is hard for a national competitor to match and valuable to the handful of people searching it. This is also the kind of page that tends to get quoted back by AI assistants, though as discussed below, nobody can promise that.

What these share: the question is stable, the answer is specific, and you know it. Those three together are the whole recipe. If a topic fails any one of them, you are commissioning filler.

Perishable content is not a mistake

None of this means you should never publish anything with an expiry date. Perishable content does a job that durable content cannot: it proves you are currently alive, trading, and paying attention. A prospect who lands on a site whose latest news item is from 2021 draws a conclusion, and it is not a kind one.

The mistake is paying evergreen prices for perishable work, and expecting perishable work to produce search traffic. A festival offer post, a new-branch announcement, a photo set from a site visit — these belong on your social channels and perhaps a simple updates feed. They should be cheap, fast and written by someone inside the business in twenty minutes. They should not be handed to a writer at a research rate, and they should not be counted as SEO deliverables, because they will not behave like SEO deliverables.

A reasonable split for a business publishing four things a month: one durable piece, properly researched, built to be linked to. Three short perishable ones, done in-house. That ratio costs less and compounds more than four middling posts from an agency content calendar.

How to price the two differently

Think of it as asset versus consumable, and account for it that way in your head.

	Durable	Perishable
Who writes it	Someone with access to a person who knows the trade	Anyone in the office
Time to write	Hours, with an interview	Twenty minutes
Where it lives	A permanent URL, linked from your navigation	A feed, a social post, maybe nowhere permanent
When you judge it	Six to twelve months on	Within a week, or not at all
Worth paying for	Yes, properly	Rarely

The practical implication: stop buying volume. A package of ten posts a month at a low per-post rate almost guarantees perishable output, because nobody can research ten durable topics in a month on that budget. You are buying word count. Word count does not rank.

What I would not commission

These are all things agencies sell, including ones that look like ours. They do not work, and knowing why saves you the fee.

Festival listicles. "Top 10 Diwali gifting ideas for corporates." Thousands of identical pages, written by people with no stake in the answer, published every October. You will not outrank the ones that have been doing it for eight years, and the page is dead by November anyway.
News commentary with no angle. Repeating what the Budget said two days after every newspaper said it. If you have a genuinely specific read — what this means for a plywood trader's working capital, from someone who finances plywood traders — that is worth writing. Summary is not.
"Ultimate guide to [your whole industry]". Enormous, generic, and competing with every well-funded site in the category. Three narrow decision pages beat one enormous guide almost every time, because narrow pages match the way people actually search.
Posts generated to hit a monthly count. If the brief originated in a calendar rather than a customer question, the page is already weak.
Durable content with no maintenance budget. This is the quiet one. A page citing a tax rate, a scheme name, a statutory limit or a screenshot of a portal is durable only if someone reviews it. An out-of-date page is worse than no page — it costs you the trust of the one reader who notices.

The AI search question, honestly

Owners keep asking whether durable content matters more or less now that assistants answer questions directly. The honest answer is that nobody outside those companies knows how their systems weight sources, and the systems change without notice.

What can be observed, rather than claimed: assistants tend to surface pages that answer a specific question plainly, in the page's own words, with concrete detail — specifications, processes, numbers that can be checked. That is the same thing that makes a decision page or a diagnosis page useful to a human. It is also, notably, nothing like a festival listicle.

So the reasonable position is that durable, specific content is a bet that pays off under several possible futures, and perishable content is a bet that pays off under none of them. That is a weaker claim than "do this and AI will cite you", and it is the one the evidence supports. If you want to think about this more deliberately, that is what AI visibility work involves — and it starts with having something worth citing.

A maintenance habit worth building

Put a review date in the page itself — a line at the bottom saying when it was last checked. Then once a quarter, open the five pages that bring you the most traffic and read them as a customer would. You are looking for three things: facts that have changed, prices or rates you no longer charge, and links that now return a 404 error (page not found). Fifteen minutes a page.

When you genuinely retire something — a product line you no longer sell — redirect the URL to the nearest living page with a 301 (a permanent redirect), rather than deleting it and leaving a dead end. If there is no sensible destination at all, a 410 (gone) is cleaner than letting a broken page sit there. Your developer will know; if the site is already set up sensibly, this is a two-minute job. If it is not, that is a structural problem worth fixing before you commission any more content.

What to do next

Open your blog or news section. Go through every post and write one letter beside it: D for durable, M for maintainable, P for perishable. Count them. If the P column is longest, you now know why the content budget has not produced anything.

Then write down the five questions customers ask you on the phone most often, in their words, not yours. Those are your next five durable pages, and the brief for each one is already in your head.

If you would rather work through that list with someone — deciding which ones are worth building out, and what the site needs to support them — that is a conversation about search and content strategy, and you can get in touch to start it. There is more on related decisions over on the blog.

Questions we get asked

How often should I publish on my business blog?

Less often than most agencies will suggest, and with more care. One well-researched page a month that answers a question your customers actually ask will usually outperform four quick posts written to fill a calendar. Publishing frequency is not itself a ranking factor you can rely on; usefulness is what gives a page a chance.

Should I delete old blog posts that are out of date?

Check first whether they bring any traffic or links. If a post is genuinely obsolete and brings nothing, retiring it and redirecting the URL to the nearest relevant live page with a 301 permanent redirect is tidier than leaving it. If it still brings visitors, update it instead — a refreshed page keeps whatever authority it has built.

Can I use AI to write evergreen content for my website?

You can use it to draft and structure, but the thing that makes a durable page work is specific knowledge only your business has — tolerances, local processes, what actually goes wrong. A model does not have that, so an unedited AI draft tends to produce exactly the generic page that fails to rank. Treat it as a typing aid, not a source.

How long before a new page starts bringing traffic?

Nobody can promise a timeline, and anyone who does is guessing. New pages on established sites generally move faster than pages on brand-new domains, and competitive national topics take far longer than narrow local or technical ones. Judge a durable page at six to twelve months, not six weeks.

Is it worth putting a date on blog posts?

For perishable content, yes — readers need to know whether news is current. For durable reference pages, a visible "last reviewed" date is often better than a publish date, because it tells the reader the information has been checked recently without making a useful 2022 page look stale.