Algeria AgencyClient acquisition
A budget split four ways learns nothing
Being everywhere on a small budget buys impressions and produces no knowledge. Why one properly funded channel beats four funded symbolically.
The most common advice, and the most expensive, fits in one sentence: be everywhere. Facebook, Instagram, TikTok, Google, a newsletter, and why not LinkedIn. The reasoning looks solid — more surfaces, more chances — and it is wrong the moment the budget is small.
It is wrong for a mechanical reason rather than an ideological one. An advertising campaign is not a billboard that works as soon as it is up: it is a system that learns, and it needs a certain number of results per week to learn anything at all. Below that threshold a campaign does not improve, it broadcasts at random.
Four campaigns each running below their threshold do not add up to one above it. They add up to four blind systems, four invoices, and at the end of the quarter four sets of numbers none of which supports a decision. This is the most common way to spend a year of advertising budget without ever finding out whether advertising works for you.
This article describes how the single channel gets chosen, what the first three months actually produce, how to read the one number that matters, and when opening a second channel becomes legitimate. It is written from the work we do in the entry tier, and it contains everything needed to do it without us.
The platforms are not the same size in Algeria
Before discussing trade-offs it is worth looking at what there is to trade off. Algerian audiences for the major platforms are published, and the gap between them is wider than most media plans assume.
That gap does not say where to go. It says only that "being on social media" covers decisions of very different sizes, and that a budget split evenly across those surfaces treats audiences as equivalent when they are not.
The useful way to read these numbers is by subtraction. A platform whose national audience is half the size of another is not eliminated for that reason alone: it is eliminated if, on top of that, your customers are not there. A national ranking is a starting point, never a conclusion.
One thing to hold onto before the rest: these are total audiences, not addressable ones. Nobody sells to twenty-five million people. What concerns you is a fraction defined by your area, your customers’ age and your price — and that fraction appears in no public report.
DataReportal, Digital 2026 Algeria — October 2025 figures
A channel is chosen on three questions, not on a trend
The first question is age. Customers over forty are overwhelmingly on Facebook; customers under twenty-five spend most of their time elsewhere, and a beautiful campaign on the wrong platform is still a campaign addressed to nobody.
The second is intent. On a search engine somebody is already looking for you: they have a problem and they type the words of that problem. On a social network nobody is looking for you: you interrupt. Both work, but not at the same price, not at the same speed, and not for the same products.
The third is visual proof. Some products sell because they are seen — a dish, a garment, a finished site. Others sell because they are explained — insurance, an audit, a piece of software. A product that needs explaining, on a platform that rewards imagery, starts with a handicap no budget compensates for.
Three answers, one decision. If all three point at the same platform the question is settled. If they diverge, the rule we apply is to follow intent: a channel where people are looking for you produces rarer enquiries that are far easier to convert, which is exactly what a business starting out with no team to handle volume needs.
The first month returns nothing, and that is its job
This belongs before the contract rather than in the first report: the first month of a campaign is not a performance month. It is a collection month. Audiences are not yet qualified, messages have not been compared, and the platform’s algorithm has too few examples to know who to show what to.
That month nevertheless produces the most useful thing in the quarter: a baseline. Without one, any judgement made in month two is an impression. With one, a change in cost per enquiry becomes a fact, and a decision can rest on it.
The classic mistake is correcting every week. Each significant change puts the system back into learning and destroys the comparison under way. A campaign reworked every eight days stays permanently in its first month — it looks busy and it never learns.
Our rule is to touch nothing in the first thirty days except what is plainly broken: geographic targeting that overspills, a landing page returning an error, a budget exhausted by midday. Everything else waits until there are numbers to set against it.
Conversion tracking goes in before the first dinar
A campaign without conversion tracking spends money and returns platform metrics: impressions, reach, clicks, engagements. None of those words describes a customer. They describe opportunities to get one, which is not the same thing and does not compare across platforms.
Tracking means connecting a spend to an event you actually care about: a form sent, a call started, an order placed, a message received. Once installed, it turns an advertising dashboard into a business dashboard, and that is the only moment advertising stops being a matter of faith.
It also has to be accepted as imperfect. Some customers will see the ad on their phone and call from the office landline; those appear nowhere. Honest tracking always undercounts slightly, and that is preferable to tracking that overcounts — the second error pushes you to spend more, the first to spend better.
This is why we launch no campaign before tracking works, including when the client is in a hurry. A week’s delay at the start costs a week; three months of unmeasured spend costs the whole quarter, because at the end there is nothing to conclude from it.
What a boosted post hid from you
Most businesses that come to us have advertised already, in one specific form: the "boost post" button. It is the best-selling advertising product in Algeria, and it is also the one that teaches you least.
A boost optimises what is visible in the interface: reactions, comments, shares. Those are real objectives, and they are almost never yours. A post heavily commented on by people who will never buy is a success for the platform and an expense for you.
The second problem is the absence of structure. A boost leaves no reusable audience, no comparison between two messages, no usable history. Six months of weekly boosts produce an invoice and nothing resembling an asset — whereas six months of structured campaign leave audiences, compared creatives and a known cost per enquiry.
That does not make a boost always a mistake. To announce a dated event to a local audience you already have, it is a sound and cheap tool. What it should not be asked to do is constitute an acquisition channel: that is not what it does.
The cost of an enquiry is the only number that decides
Once tracking is in place, a single number governs every decision that follows: what a qualified enquiry costs you. Not a click, not an impression, not a follower — a person who contacted you with an identifiable intention.
That number has no universal reference value. The 2025 figures published by WordStream, American and in dollars, give the measure of the gap between sectors: the most expensive click they record is for dental practices, the cheapest for restaurants. The ratio between them is more than ten to one.
Those figures are not yours and must not serve as a target. They come from another market, in another currency, with another level of competition. All they prove is this: asking "what does a click cost" without naming a sector is a question with no answer, and any agency that replies with a single figure has invented something.
The right way to use your own cost per enquiry is to compare it with what a customer is worth to you, not with what one costs somebody else. If a customer brings you thirty thousand dinars over their lifetime and an enquiry costs you eight hundred, the question is no longer the price: it is whether your team can handle more.
WordStream, 2025 benchmarks — US market, in dollars
The threshold below which a campaign cannot learn
Advertising platforms optimise by learning: they show the ad, observe what happens, and narrow. That mechanism needs a minimum number of conversions per period to produce anything other than noise, and below that number optimisation never engages at all.
We do not publish that threshold as a chart, and that is deliberate. Each platform sets its own, documents it itself, and changes it without notice — which makes it a number produced by the seller to describe the minimum you must buy from them. Citing it as data would treat a commercial argument as an independent measurement.
What remains true without the exact figure is the shape: the threshold exists, it is counted in tens of conversions per week per campaign, and it does not divide. Two campaigns at half the threshold are worth zero campaigns above it, not one.
The practical consequence fits in a sentence, and it is the one that gives this article its title. On an entry budget you choose between one channel that clears the threshold and several that do not. This is not a trade-off between caution and ambition: it is a trade-off between learning and not learning.
Publishing regularly beats publishing a lot
The social side of this tier is not chasing virality. It is chasing regularity, because regularity is what decides what a customer finds when they check that you exist before buying — and almost all of them check.
A page whose last post is seven months old communicates something precise, and it is not "we are busy". A page posting twice a week, however modestly, communicates the opposite. The content matters less than the proof of activity, at this stage at least.
Regularity is also what makes a calendar survivable. Posting four times a day for three weeks and then stopping is the most common pattern among businesses running their own social media, and it produces a page that looks abandoned eleven months out of twelve.
The second job, less visible, is answering messages. An enquiry received in an inbox and read three days later is a lost enquiry, whatever advertising budget produced it. Response time is the first place well-run acquisition gets squandered.
The ground is a phone, not a desk
All of the above happens on a six-inch screen, held in one hand, on a mobile network. The split of internet subscriptions in Algeria leaves no room for doubt on this point.
That changes how an ad is judged. A visual designed on a twenty-seven-inch screen, with thin type and three levels of information, disappears in a feed scrolled with a thumb. The question is not "is it beautiful" but "is it legible at thumbnail size, in one second".
It changes the destination too. An ad that lands on a slow page is an ad paid for twice: once to the platform, once in visitors lost during loading. The advertising budget then funds a demonstration of a technical problem.
Finally it changes the means of contact. A tappable number and a monitored inbox convert better than an eight-field form, not because the form is badly made, but because filling in eight fields with a thumb, standing up, is an effort few people make for a first contact.
- Mobile subscriptions88.71%
- Fixed subscriptions11.29%
ARPCE, internet market observatory, second quarter of 2025
Three questions at the end of the quarter
A quarter of acquisition is judged on three answers, and they are the same in every sector. The first: what does an enquiry cost? If that number does not exist, the quarter has failed, regardless of how many posts went out and how good the visuals were.
The second: where do the best ones come from? Not the most numerous — the best. An audience producing ten enquiries of which one buys is worth less than one producing three of which two buy, and only the second deserves more budget.
The third: what would happen if you doubled it? Some channels take the increase, others degrade immediately because the useful audience is already saturated. Knowing which one you have is the difference between investing and wasting the next quarter.
None of these three answers needs a particular tool or supplier. They need tracking installed and three months of patience. If you get them on your own, you do not need us for this stage, and that is a conclusion we would rather see written down somewhere than discovered too late.
What this tier cannot repair
An acquisition tier brings visitors. It does not correct what they find on arrival, and that is the most important limit to understand before signing anything.
If your site takes eight seconds to appear, if your catalogue has no prices, if nobody answers the phone between noon and two, then paid traffic only increases the number of witnesses to the problem. The advertising budget funds the leak instead of plugging it.
The symptom is recognisable: a normal cost per click, a visitor volume in line with expectations, and almost no enquiries. When we see that combination in month one we say to suspend the spend and fix the landing page first — advice that reduces our own invoice the following month, and there is no other honest advice to give.
The general rule: advertising multiplies what already exists. It multiplies a clear offer as readily as a confused one, and in the second case the result is simply a more expensive confusion.
What we do, and what we refuse to do here
Our work in this tier amounts to few things: choose the channel and write down why that one, install tracking before the first spend, hold a publishing rhythm, correct once a month on numbers, and deliver a report that ends in a decision rather than a table.
What we refuse is shorter and more precise. We do not open a second channel during the first quarter, even where the budget allows it and even when you ask. The reason is the one in section seven: two channels below the threshold are not worth one above it, and we would be paid to degrade the only result the tier is meant to produce.
Nor do we promise a cost per enquiry before we have a month of your data. An agency that names a figure at signature took it from somewhere else — another sector, another market, or nowhere. We give a reasoned range after the audit, and we say when it is out of reach.
Finally, we say stop when the numbers say stop. A channel that has not cleared its threshold after a properly run quarter will not clear it in month four on the same budget, and continuing out of commercial politeness costs the client an entire season. That is the part of the work that costs us contracts, and it is also the only part that makes the rest credible.
Frequently asked questions
Why one channel if I have the budget for two?
Because the learning threshold does not divide. Two campaigns at half the threshold produce two unusable datasets; one above it produces a decision. The second channel opens when the first is profitable and measured, not before.
How long before I see results?
Enquiries can arrive in the first week, and that proves nothing. What takes three months is a stable cost per enquiry — that is, a number you can decide to scale or stop on.
Is the advertising budget included in the price?
No. What you pay us pays for the work; what you pay the platform buys the audience. Both are in your budget, only one is on our invoice, and the ad account is created in your name.
Can I keep my current page and posts?
Yes, and it is preferable: a page’s history has value and a new page starts from zero. We take over what exists and change the rhythm, not the identity.
What happens if the chosen channel does not work?
We say so in the report for the month it becomes legible, with the number that shows it, and we propose either a change of channel or a stop. The bad scenario is not that a channel fails; it is that it fails for six months without anybody writing it down.
Do I need a website before advertising?
You need somewhere to land that gives a price, a proof and a way to make contact. A well-made page is sometimes enough. What is never enough is a homepage that talks about the company and says neither what it costs nor how to order.
Where we come in
One channel held for a whole quarter leaves you a figure that is yours; four channels held for three weeks leave you an impression.
- We choose the channel with you and write in one line why that one.
- We install the measurement before the first advert goes out.
- We set in advance the threshold below which that channel stops.
If you have already decided to be everywhere at once, do not call us: we would not know how to run that.
Read next
Your side of it: what has to be true at your end for this level to produce anything
At this level the outcome is decided at your end more than at ours. Advertising is half an act, and you hold the other half.Traffic is not the problem; the page is
Acquisition cost and conversion rate multiply. A point gained on the page is worth the same as an equivalent discount on the click, costs less, and does not stop when the invoice does.What advertising budget to set — and how to know
There is no published rate card for Algeria. Here is what the available data actually lets you conclude.
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.