Algeria AgencyDigital marketing
Why they do not buy, and how to find out
Almost every optimisation programme starts by changing things. The diagnosis costs a week and makes the rest useful.
The usual conversation about conversion is about changes: move a button, shorten a form, try another colour. Those changes are sometimes good and they all share one flaw — they are proposed before anybody has established why people are not buying.
The number of reasons is small, though. In this market five objections cover very nearly every abandonment, and they are not mysterious: price, trust, delivery, payment, and doubt about whether the product matches. Each is corrected differently, and correcting the wrong one produces nothing.
This article is about the diagnosis: the five objections, how to ask people who did not buy, what the price question really means, what happens at the moment of payment, and how to test only what the diagnosis found. The companion article covers the mechanics — the rate as a fraction, the unit of counting, the price of a form field, counting by hand — and none of that is repeated here.
One figure appears in it, in the payment section, and it changes a decision rather than illustrating one.
Optimising without a diagnosis
An optimisation programme nearly always starts with a list of ideas, and that list is made of things somebody read or saw elsewhere. It is produced before any contact with the question of what is blocking, and its order is that of ease of execution rather than importance.
The result is predictable: you get variations with no effect, conclude that optimisation does not work in this market, and go back to spending on advertising — that is, to bringing more people to the same obstacle.
What the diagnosis changes is not the quality of the ideas but their order. A business whose obstacle is doubt about delivery can shorten its form for a year and gain nothing, while one sentence about the real lead time, placed in the right spot, shows within the week.
The diagnosis costs a week and it is dull: it consists of talking to people and watching what they do, rather than changing something. That is precisely why it gets skipped, and the reason it is nearly always profitable.
Five objections, and they are not mysterious
The first is price, and it is the most badly understood — section four returns to it, because “it’s expensive” rarely means what the word says.
The second is trust: the person is not sure you really exist, or that somebody will answer if something goes wrong. It is the dominant objection for a business the buyer is discovering, and it is handled through details people think insignificant.
The third and fourth belong to distance selling and are particularly present here: doubt about delivery — when, by whom, how far, at what price — and doubt about payment, which is a technical objection as much as a trust one.
The fifth is fit: the person is not certain the product is what they think, in the right size, the right quantity, the right compatibility. It is very frequent and is nearly always corrected by information rather than persuasion, which makes it the cheapest to handle.
Ask the people who did not buy
The most reliable source is also the least used: the people who wrote and did not go through with it. They are identifiable, they are few, and they answer more readily than expected because the question is unusual.
The wording matters and must contain permission not to buy. “You wrote to us about the blue model — did you end up getting it elsewhere, or did something stop you? It helps us to know.” It does not restart the sale, which is what makes it effective: as soon as a message looks like a follow-up, the real answer disappears.
Twenty messages is enough, and that is an important point because it makes the exercise feasible. This is not a study: the five objections are few and repeat very quickly. After fifteen replies the distribution is established and stops moving.
There are two other free sources nobody uses. The questions asked before purchase — the same question asked ten times is missing information on the page — and the reasons for returns, which say which objection should have been lifted before rather than after.
The price question is not always price
“It’s expensive” is the most frequent sentence and the most badly translated. It is heard as a request for a discount, and it is one in a minority of cases. In the others it means something different: I cannot see what justifies that difference.
The distinction is checkable and tests with one question: expensive compared with what? The answer says everything. Compared with a named competitor, it is a price objection. Compared with another product that is not comparable, it is an objection of understanding. Compared with nothing, it is a doubt looking for a justification.
The practical consequence is that a price cut is rarely the answer, and it is irreversible in practice. What settles the majority of these cases is an explanation of what the price contains — the guarantee, the service, the origin, the installation — placed next to the price rather than three pages away.
You also have to accept the case where the objection is real: your price genuinely is above the market for an identical product. No amount of page work corrects that, and acknowledging it saves months for a business that would otherwise try every technique on this list in turn.
Trust is proved by dull details
The trust objection is the most frequent for a business being discovered, and it is handled through elements whose effect looks disproportionate to their banality. A full physical address. A phone number that answers. Somebody’s name.
Those elements do not reassure because they are checked — almost nobody checks them. They reassure because their presence signals that somebody accepts being findable, which is exactly what the hesitant buyer doubts.
Two things work better than the rest in this market, and they cost nothing. Real photographs of the business — the premises, the workshop, the people — rather than bought images, which produce the opposite effect as soon as they are recognised. And reviews, particularly the ones that are not perfect: an average rating that is too high reads as manufactured.
What degrades trust is equally well known and easy to fix: an absent price, an empty “about” page, a last post eight months old, a dead link. None of those is an argument, and their accumulation produces an impression nothing else makes up for.
Doubt about delivery
Delivery is an objection in its own right, not a logistics detail. Somebody hesitating asks three precise questions — when, how much, and what happens if I am not there — and the absence of an answer to any one of them is enough to stop a purchase.
In this market two further uncertainties add to the usual three. The real coverage: do you deliver to my wilaya, and how far exactly, which is almost never written precisely. And the mode: home delivery or collection from a pick-up point, which are not the same thing from the buyer’s point of view.
The correction is textual and costs an hour: one sentence per question, written plainly, next to the buy button and not in a terms and conditions page. The content matters less than the precision — “two to four days to Oran” beats a “fast delivery” that says nothing and reads as evasion.
It is better to announce an honest lead time than a flattering one, for a reason that is economic rather than moral. A short lead time announced and missed produces a message, often public, and sometimes a return; a long lead time announced and met produces nothing at all, which is the intended result.
What happens at the moment of payment
Payment is the most concrete of the five objections, because it does not rest on a feeling: either the person can pay the way they usually do, or they cannot. And the choice of accepted methods is made, in most cases, without looking at what people actually hold.
Cards in circulation in Algeria at the end of 2025 are distributed very unevenly: 21,899,581 cards in total, of which 17,655,039 are Edahabia and roughly 4.2 million CIB. In other words, about one card in five belongs to the banking network and the rest to the postal one.
The consequence is direct and rarely drawn: a merchant accepting only one of the two networks is not making a technical choice, they are excluding most cardholders in the country. It is not a question of interface preference or commission, it is a question of who you are selling to.
Cash on delivery has to be added, which remains dominant and is often treated as a fallback. It has a real cost — returns and uncollected parcels — and it lifts both the payment objection and part of the trust objection at once, which explains its persistence. Removing it to save those returns moves the loss rather than removing it, and that arithmetic is done beforehand, not afterwards.
2025 card-stock review, GIE Monétique and Algérie Poste
What they want to know and you do not say
The fifth objection — fit — is the cheapest to correct and the most systematically ignored, because it does not look like a commercial problem. The person wants to buy and does not know whether this is the right item.
The missing information is nearly always the same and it is dull: the real dimensions, the material, what is in the box, compatibility with a specific model, the capacity. It is not absent through negligence but because it seems obvious to whoever is selling.
Spotting it is free and immediate: it is the questions asked in comments and private messages. A question asked ten times is missing information on the page, and treating it as a question to answer rather than a page to fix means paying for the same defect indefinitely.
There is a frequent special case here and it is easy to handle: scale. A product photograph with no size reference leaves an uncertainty the text does not always lift, and it is one of the most banal causes of returns — therefore a legitimate objection in somebody who has already been disappointed once.
The moment they leave
Knowing which objection dominates is not enough: you have to know where it shows, because the same objection is corrected in different places depending on when it stops the person.
Three moments concentrate abandonment and are easy to tell apart. On the product page, where people leave because information is missing — that is the fit objection. At the point of making contact or filling the basket, where they leave from lack of trust or doubt about delivery. And at the payment screen, where they leave for a mechanical reason.
The distinction is useful even without sophisticated measurement, and it comes from section three’s conversation: people tell you where they stopped when you ask them. The word “when” in the question does more work than any report.
There is an abandonment that is not one and has to be set aside: the person comparing who will come back. They are numerous, they produce an alarming figure, and they fall under none of the five objections. That is one more reason to talk to a few people rather than concluding from a percentage.
Test only what the diagnosis found
Once the dominant objection is identified, the correction is often obvious and needs no testing: add the delivery time, write the dimensions, accept the second card network. Those are additions of information, and missing information has no superior version to test.
Testing becomes useful afterwards, when two plausible wordings are in opposition and nobody can settle it by reasoning. It is then an arbiter, which is its role, and not a method of discovery — the companion article explains separately why the necessary volume is rarely available.
The order that follows is simple and is the reverse of the usual one: fix what is missing first, measure whether the effect shows, test last and only the ambiguous cases. Most businesses do the opposite and conclude that nothing works.
You also have to measure the right thing. A correction that lifts the delivery objection shows in the number of accepted orders, not in time on page or click rate — and confusing the two causes working corrections to be abandoned.
Taking over a site that converts badly
The ordinary case is a site receiving traffic and producing little, with a request phrased as “it needs redoing”. A complete rebuild is nearly always the wrong answer, and it is the most expensive: it changes everything at once, which makes it impossible to know what acted.
The sequence that works starts with section three’s conversation, before any change. Twenty messages to people who did not buy give, within a week, a distribution of the five objections — and that distribution surprises the team in a majority of cases, which is precisely the point of the exercise.
Then comes one correction, the one addressing the dominant objection, with a before-and-after measurement on the unit that matters. One, because it is the only way to learn anything, and because a business that sees a real effect accepts the second correction without discussion.
The rebuild, if it remains desirable, comes afterwards and costs less, because it is done knowing what was blocking. Redoing a site without that diagnosis amounts to changing all the answers on a questionnaire whose questions were never read.
What we do, and what we refuse
What we take on is bounded: the twenty conversations, reading the questions already asked and the reasons for returns, the distribution of the five objections, and the first correction with its measurement — then the second, if the first showed something.
We do not run tests before a diagnosis. It is a refusal that costs, because testing sells well and looks rigorous, and because it usefully occupies a team for weeks. Testing variants without knowing which objection dominates amounts to arbitrating between two answers to a question nobody asked.
We do not remove a price from a page to increase contact requests. It is a technique that works in the sense that the number of messages rises, and it moves work towards conversations that will not conclude while degrading the trust objection among precisely the people who were looking for a price. We measure accepted orders, not conversations started.
Finally, we do not promise a gain in rate. A rate is a fraction whose two halves move — the companion article starts there — and an advertising campaign launched the same month is enough to lower it while sales rise. What we track with you is the number of accepted orders, before and after, over a comparable period.
Frequently asked questions
Where do we start when a site converts badly?
With a week of diagnosis, before any change. Write to twenty people who contacted you without buying, read the questions asked before purchase and the reasons for returns. The five objections — price, trust, delivery, payment, fit — distribute very quickly, and the distribution surprises the team in most cases. Correcting the wrong objection produces nothing, however good the correction.
How do you ask somebody why they did not buy?
By giving them permission not to have bought: “You wrote to us about the blue model — did you end up getting it elsewhere, or did something stop you? It helps us to know.” As soon as a message looks like a sales follow-up, the real answer disappears. Twenty messages is enough: the objections are few and repeat very quickly.
“It’s expensive” — should we cut the price?
Rarely, and ask a question first: expensive compared with what? Compared with a named competitor, it is a price objection. Compared with a non-comparable product, an objection of understanding. Compared with nothing, a doubt looking for a justification. Most cases are settled by an explanation of what the price contains, placed next to the price. A cut is irreversible in practice.
Which payment methods should we accept?
Both networks. At the end of 2025, of 21,899,581 cards in circulation in Algeria, 17,655,039 are Edahabia and roughly 4.2 million CIB: accepting only one network is not a technical choice, it excludes most cardholders. Keep cash on delivery too, which lifts the payment objection and part of the trust objection at once — removing it moves the loss rather than removing it.
What should we write about delivery?
One sentence per question, next to the buy button and not in the terms: when, how much, what happens if I am not there — plus two uncertainties specific to this market, the real coverage and the mode. Precision matters more than content: “two to four days to Oran” beats a “fast delivery” that reads as evasion. Announce honest rather than flattering: a short lead time missed produces a public message, a long one met produces nothing.
Should we run A/B tests?
After the diagnosis, and only for ambiguous cases. Missing information has no superior version to test: adding a delivery time or dimensions is done, not tested. A test is an arbiter between two plausible wordings that reasoning cannot separate — the companion article explains separately why the necessary volume is rarely available.
Where we come in
Fifteen replies are enough to learn which of the five objections holds you. Writing them is easy; reading them without reassuring yourself is much less so.
- We read your replies cold and hand back the one that keeps recurring.
- We check which card networks your payment actually accepts.
- We fix the first objection before the second one is opened.
- We send the same message again three months later, on the same terms.
If you do not have twenty people to write to, this diagnostic cannot be run: wait until you do, which will come sooner than you expect.
Read next
Conversion rate: the unit is the accepted order
In this market the event on the page is a message, not a purchase. The two biggest losses come afterwards, and neither one shows on the site.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.When it works: what breaks next, and in what order
Growth breaks a business before it breaks a campaign. The constraint does not disappear, it moves — and never to where you are watching.
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.