Skip to content
Client login

Free Audit

Digital marketing

Content marketing: the one asset you keep

Everything else in this pillar is rented and stops with the invoice. A written page stays, provided it answers a question somebody genuinely asks.

Published on 4 August 2026 — Algeria Agency

The article on digital marketing describes four floors of spending and shows that the loss nearly always happens downstream of the first. This one takes a particular line in that pillar and explains why it does not compare to the others.

The difference comes down to one word: ownership. An ad account, an audience on a network, a place in a ranking — all of that is rented, and all of it stops on the day the invoice stops. A page written and published at an address you own stays, and goes on working without being topped up.

That does not make content a superior expense; it makes it an expense of a different nature, judged over a different span and with different numbers. Confusing the two is the source of the commonest disappointment in this trade.

This article fixes the vocabulary the pages beneath it reuse — search, copywriting, social content, email — and it begins by refusing a promise the profession makes constantly and that no dated figure supports here.

Rented against owned: the only distinction that matters

Take every line of your marketing budget and ask a single question: if I stop paying tomorrow morning, what is left? The answer sorts the whole thing into two piles, and the sort is more useful than any classification by channel.

What stops: advertising, in all its variants. Distribution ceases within hours, the audience built belongs to the platform, and the account itself can be taken away from you — the article on search advertising describes how it regularly ends up in somebody else’s name.

What stays: the pages of your site, the product records, the written answers, the list of addresses you collected, the photographs you took. Those are objects you hold, can move, can rewrite, and can carry to a different supplier.

Content belongs to the second pile, and that is its whole economic justification. It costs more to produce per unit, it acts more slowly, and it does not switch off. A sound page written this year is still answering people three years later, which no campaign does.

The corollary is a spending rule, and it is simple: never fund content that will be published at an address you do not control. A text published only on a social network is an expense from the left-hand pile, whatever its quality.

The unit is not the article, it is the question

The service is sold almost everywhere by volume — so many articles a month — and it is the same unit error the article on social media management identifies for its own subject. A production unit says nothing about the work accomplished.

The right unit is the question. A real question, asked by a real person, whose answer is currently absent or scattered. One page that answers such a question completely is worth more than twelve that circle a theme.

That immediately changes how a plan is built. You do not start from a list of topics; you start from the inbox, the salesperson’s notebook, the questions asked at the counter, and whatever after-sales repeats every week.

It also changes the stopping criterion. A page is finished when the question no longer needs asking, not when it reaches a word count. Some answers fit in six hundred words and deserve no more.

And it settles the question of how many: if you have eleven recurring questions, you have eleven pages to write and then nothing to do for a while. An honest service has to be able to say "we are done", and almost none of them do.

Three things all called content

The word covers three objects that share neither purpose, nor author, nor measure. Confusing them produces services that miss all three at once, and it is the commonest fault in the offers we see go past.

The first is content that has to be found. It targets a search, it is written to be indexed and cited, and the pages on search and on citation by answer engines describe how it works. Its measure is appearing in a result.

The second is content that has to persuade. Nobody finds it: it is sent, shown, opened at the moment of a decision. Its measure is the conversation it shortens — how many questions it removes before the quotation.

The third is content that simply has to exist. The complete product record, the pricing page, the delivery terms, the area covered, the warranty. Nobody reads it with pleasure and its absence costs sales every day, which makes it the most profitable of the three.

The sequence we apply is the reverse of the one being sold: third first, then second, first last. Writing to be found before having anything to answer with amounts to opening a door onto an empty room.

The figure we will not quote

There is a statistic the entire profession recites in this context: a proportion of the buying journey supposedly completed before the first contact with a seller. You will see it in nearly every proposal, with a decimal place and no date.

We will not quote it, and it is better to say why here than to let it slip past elsewhere. The studies that produce it cover North American or European business markets, they survey professional buyers, they are often commissioned by vendors of marketing software, and the most repeated one is a decade old.

Any single one of those objections would be enough to make it useless for deciding anything in Algiers or Oran. All four together make it an ornament: a number that gives a recommendation the appearance of a measurement.

The house rule is the one applied throughout this series: when a figure cannot be dated and placed, it does not get published and the reason gets said. That is why this page carries no chart at all — there is nothing dated to put in one on this particular subject.

What remains in the absence of a figure is not nothing: it is your own data. The number of identical questions received this month, the number of quotations lost for want of a written answer, the time your salesperson spends re-explaining the same thing. Those three are dated, placed, and yours.

Where your content actually lives: inside a message thread

Here is this article’s local inversion, and it is the observation the pages below should take up rather than rediscover. A great deal of the useful content in Algeria is never found through a search at all. It is sent.

The mechanism is ordinary and visible in any company: a customer writes on a messaging app, asks a question, and whoever replies pastes a link. That page does the commercial work, and the number of times it gets sent appears on no search dashboard.

The article on search advertising explains why this is structural here: for many trades the search volume does not exist in sufficient quantity, and demand travels through conversation. A content strategy that counts only what is searched for therefore misses most of what is read.

That imposes precise constraints on form. A page sent in a thread opens on a phone, in a fraction of a second, and has to answer within the first screen. It has to make sense without the rest of the site, without navigation, and without an accompanying message written to explain it.

And it changes quality control: the right question is not "will this page rank", it is "would I be comfortable sending it as it stands to a customer who is hesitating". If the answer is no, the page is not finished, whatever its ranking.

Write first for the person who has already decided

The usual order starts with broad subjects — guides, trends, lists — and works down towards the sale. It has to be taken the other way round, and the reason is arithmetic rather than theoretical.

Pages close to the purchase are few, quick to write, and read by people about to pay. Broad pages are countless, slow to write, and read by people gathering information. At equal budget, starting with the first produces a result within the month and with the second within the year.

The list of pages close to the purchase is nearly always the same: what it costs, how long it takes, what is included, what happens if it does not work, how to order, where you operate. Six pages, and most Algerian sites we open have none of them.

Then come the comparison pages, which are the hardest to write honestly: this one or that one, repair or replace, rent or buy, do it yourself or hand it over. They work because they take the risk of recommending, and they do not work any other way.

Broad subjects come last and only if budget remains. Not that they are useless: they are slow, and a business that needs customers this quarter should not fund them first.

Price, lead time, area: the three pages nobody writes

Three pieces of information are missing from nearly every company site we open, and their absence costs more than everything written in their place. They deserve a section of their own because the reluctance is always the same.

The price. The argument against is familiar: "it depends", and "our competitors would see". The first is settled by a range, a starting rate, or three real projects with their amounts. The second is false in practice: your competitors already know your prices, it is your customers who do not.

The measurable effect of that page is not more enquiries. It is fewer enquiries and more sales, because people outside the budget stop writing in and your salesperson stops spending days qualifying them. It is the same mechanism as the priced caption described in the content-creation article.

The lead time. How long between order and delivery, between call and intervention, between deposit and start. A business that writes down its lead times looks organised before it has proved anything else, and it removes the question asked most often after price.

The area. Which wilayas, what radius, on what terms you travel and above what amount. It is a three-line page that avoids impossible conversations every week, and its absence is the silent cause of a share of the messages that go nowhere.

Three readers, three pages, never one bilingual page

The language question arrives early and it is badly put: "should we write in French or in Arabic". The right question is who reads, and the answer is nearly always that there is more than one reader.

A single subject can have three readerships: the one searching in French, the one searching in Arabic, and the one who writes in transliterated dialect and reads whatever they find. Those are three different entrances to the same information, and one page does not serve them.

What must not be done is writing a bilingual page. A paragraph in French followed by the same paragraph in Arabic gives a page twice as long as it should be, unreadable in both languages, and badly ranked because engines cannot tell what it is. Two separate pages beat one doubled page.

Translating is not sufficient either. Questions are not put in the same terms from one language to the next, and a literal translation of a French title produces a phrase nobody types in Arabic. The Arabic version has to start from the question as it is actually worded.

Our default position: begin with the language of your current customers, not that of your sector. You already know it — it is in your received messages, and looking at the last thirty settles it without a debate.

Cadence: four pages that hold rather than forty that die

The cadence promised in proposals — four articles a month, indefinitely — is not an ambition, it is a forecast of failure. It assumes useful subjects are inexhaustible, which they are not, and it produces padding as soon as the real list runs out.

Our observation, repeatedly: a business owns between eight and fifteen questions worth a page. Past that stock, invention is required, and invented content is recognisable immediately — it talks about trends, it opens with a definition, and it contains no number from the business itself.

The right cadence is therefore a decreasing one. A great deal of work at the start, then very little: the real stock is written within a few months, then maintained. That is bad for a monthly contract and good for the client, which explains why almost nobody offers it.

Maintenance, on the other hand, is real and permanent. Prices change, lead times change, a range disappears, a regulation moves. A wrong page is worse than a missing one because it creates a precise expectation that reality disappoints on delivery.

So fix a review: twice a year, the pages close to the purchase get reread and their figures corrected. It is half a day, it interests nobody, and it is what separates a living site from an archive with a logo on it.

What makes a page reusable

A useful page gets sent several times by several people. It is the most reliable measure of value available to us, and it is built from four properties that can be checked before publication.

It answers within the first screen. No introduction setting the context, no history of the trade: the answer, then the detail. A page that scrolls before answering will not be sent, because whoever sends it has to be able to guarantee that it answers.

It stands on its own. It does not refer to three other pages to be understood, it does not assume the previous one has been read, and it restates in one sentence what it is. A link always arrives alone in a conversation.

It contains at least one verifiable element: a price, a duration, a dimension, a list of wilayas, a photograph of the real object. That is what separates it from atmosphere writing, and what makes it quotable — by a customer to their partner, and by an answer engine, as the page on that subject explains.

Finally, it has a stable address. A page moved at every redesign loses every link that pointed at it, including the ones pasted into private conversations two years earlier. Address stability is a technical decision that protects an editorial asset.

Measuring when nobody is searching

The standard measure of this service is search traffic, and it is insufficient here for the reason set out above: a significant share of the reading arrives through sent links, which appear nowhere as an intent.

Three measures hold up better, and none of them requires a complicated tool. The first: how many times your own team sent a link to a page this week. That gets counted by hand for a month, and the result is generally a surprise.

The second: the length of conversations. If the pricing page exists, exchanges before the quotation get shorter. Count the messages between first contact and decision, before and after publication. It is the measure closest to money.

The third: the questions that disappeared. Keep the list of recurring questions and cross off the ones that stop arriving. A question no longer being asked is content that worked, and it is the only measure on this list that depends on no platform.

Search traffic remains useful and is handled by the pages devoted to it. But it arrives later, it depends on outside factors, and taking it as sole judge in the first six months leads to stopping a service that was already working elsewhere.

What we do, and what we will refuse to do

What we will refuse: selling a subscription to a number of articles per month. It is the market’s standard format and it is structurally dishonest beyond the fourth month, because it obliges production when there is no real question left to treat.

We will also refuse to write without access to your field data: the messages received, the salesperson’s objections, the reasons for refusal, the questions at the counter. Content written from the outside is recognisable, and it is recognised by your customers first.

We will refuse to publish a figure we cannot date and place, including when it would make your page more persuasive. Section 3 explains why, and the rule applies to our own proposals before it applies to your pages.

What we do: the list of real questions established with your teams; the six pages close to the purchase written first, price included; one version per language where the readership justifies it, never a bilingual page; stable addresses; and handover of the source texts, to you, with no usage limit.

And what you can do this week without us: open your last thirty incoming messages and note every question asked. Count the repetitions. The three commonest are your first three pages, and you have just done the work we bill at the highest rate.

Frequently asked questions

How many articles a month should we publish?

No number is a good answer. A business owns between eight and fifteen questions that deserve a page; write those, then maintain them twice a year. An indefinite monthly cadence produces padding as soon as the real stock runs out.

Should prices be published?

Yes, at least a range or three worked examples. The effect is not more enquiries but fewer enquiries and more sales: people outside the budget stop writing in, and your salesperson stops spending days qualifying them.

Should we write in French or in Arabic?

Look at your last thirty incoming messages: the answer is in them. If both readerships exist, make two separate pages, each starting from the question as it is worded in that language — never a bilingual page, which is unreadable in both.

Does content replace advertising?

No, it does not act at the same speed. Advertising produces within the week and stops with the invoice; a page produces slowly and never switches off. The useful question is what remains if you stop paying tomorrow morning.

How do we know it works without search traffic?

Count three things: how often your team sends a link to a page, the length of exchanges before a quotation, and the recurring questions that stop arriving. All three are countable by hand and depend on no platform.

Who owns text written by an agency?

You do, if you have it written down — and require the source files plus unlimited-duration use. Content published only on a social network is not yours whatever its quality: it lives at an address you do not control.

Where we come in

The count hands you three subjects and an order. It does not say which of them you could write yourself in an afternoon.

  • We separate what answers itself from what needs figures only you hold.
  • We write the pricing page first where it is missing, because that is what gets searched for.
  • We hand each page over dated, with the name of whoever will keep it true.

Thirty messages asking thirty different questions yield no page at all: your subject is qualification, not content.

Read next

Let us talk about your project

A free audit, no commitment: we look at your online presence and tell you what is holding it back.

We measure how this site is used with Google Analytics, to learn which pages actually help. You can stop that measurement at any time from the footer. Cookie policy