Algeria AgencyBranding & design
Branding: which decisions are irreversible, and in what order to take them
This pillar ends in objects made by somebody else. There is no undo, and that is what settles the order.
Two things with almost nothing in common are sold under the word "branding": decisions, and objects. Decisions cost nothing to take and everything to change; objects cost money to produce and are the only part anybody will ever see.
Most businesses buy the objects without having taken the decisions, and end up with a very tidy set of files expressing no choice at all. The opposite mistake exists too, and is rarer: months of positioning workshops from which nothing is produced.
But this pillar has a peculiarity that lies elsewhere, and it is material: it is the only one whose production leaves your hands. A page is corrected, a campaign stopped the same day, software patched the following week. Five thousand labels in the wrong colour are not corrected — they are five thousand labels.
This article will not tell you what your brand should look like. It will give you the order in which to take the decisions, the three tests that cost nothing at all, and the things we refuse to deliver even when they are ordered.
What this pillar contains: decisions and objects
The first half is decisions: the name, what you are and what you refuse to be, who you address, and the protection of all that. They are taken in a room, with no production budget, and they are nonetheless the most expensive part to alter once everything else has been manufactured around them.
The second half is objects: the identity files, the interface of a site or an application, the printed matter, the packaging, the sign. They cost money to produce, they require outside suppliers, and they are the only part your customers will ever meet.
The usual confusion is to order the second half hoping it will produce the first. That never happens: a designer can give a decision a form, they cannot take it for you, and a palette justified by nothing but taste will change with the next taste.
The mirror mistake is rarer and just as costly: weeks of positioning work from which no object emerges. A decision nobody can see is not a brand, it is a document, and a document does not sort customers on your behalf.
The rule joining the two halves is easy to check: every object should trace back to a decision, and every decision should show up in at least one object. Put your current identity through that test and you will know which half you are missing.
What makes this pillar different: there is no undo
Everything else we do can be recovered after the fact. Text is rewritten, a campaign is cut the same day, software is fixed in the next release. The cost of a mistake there is the cost of the correction, and it is nearly always small.
This pillar ends in objects made by somebody else: a printer, a label supplier, a sign maker. The cost of a mistake is not the cost of the correction but the cost of the entire production run, plus the stock that has already gone out to customers.
So there is a gradient, and it is the same everywhere: a screen changes for free, printed matter costs a reprint, packaging costs a reprint plus the stock on the shelves, and a sign costs a hoist and a day of installation.
That gradient is the only dependable way to order this work, and it is exactly the inverse of the order people instinctively want. A director investing in their brand wants to start with the sign, because the sign is what they see; it is the last object to make, for precisely that reason.
The practical consequence fits in one sentence: nothing irreversible gets made until the decisions behind it have survived a month of use on the reversible surfaces. A month on screen costs nothing and reveals what no presentation reveals.
The order: from least reversible to most visible
First the name and its availability, because a name you cannot keep cancels everything built on it. It is the only step whose failure makes the others pointless, and it is the cheapest of the eight.
Then the core: a drawn mark, two typefaces, a palette whose contrasts have been checked rather than chosen. The article on visual identity sets out what that core contains and what serves no purpose in it; we are not repeating it here.
Then the screen surfaces, which cost nothing to revisit: the profile picture, the site, a document template, an invoice. That is where an identity gets worn for the first time, and one month of real use will tell you what does not hold.
Then the printed matter, whose mistakes cost a reprint. Then, last, the packaging and the sign, whose mistakes cost a full production run and are visible from the street for years.
That order has a useful side effect: by the time you reach the most expensive object, the brand has been used for a month by people who did not draw it. Its weaknesses are known, and you are not cutting them into a medium you will never be able to take back.
The name is the one decision design cannot rescue
An excellent identity placed on a name somebody else registered before you is a spend lost in its entirety, and it is discovered at the worst possible moment: when the sign is up and the letter arrives.
The first test is the telephone one: your name has to be spelled aloud without being repeated twice. A name requiring "no, with two Ls" on every call costs you a friction at every contact, every day, forever.
The second is the typing test. A name whose spelling cannot be guessed will not be found again: whoever is looking for you types what they heard, and they arrive somewhere else. A brand half its searchers cannot find has a naming fault, not a search fault.
The third is the three-script test. A French pun dies in Arabic, and a joke that only works in one language becomes an awkwardness in the other two. A name that means nothing anywhere is sturdier than a witty one in a single language.
What we refuse here is plain: we do not design an identity on a name whose availability has not been checked. The check happens before the design rather than after, because afterwards it is no longer availability that decides but the money already spent.
Trilingual is not a translated logo
A brand working here is read in three languages and written in two alphabets. That is not a translation job: the name in Arabic is not the French name transcribed, and the drawing carrying it is not the Latin drawing flipped.
Latin letterforms have no Arabic equivalents. There are no capitals, the forms join, and the writing line and the rhythm of thick and thin are not the same. An Arabic logotype is drawn; it is not converted, and no automatic setting produces that result.
Both versions have to be equally good, because neither of them is the real one. The mark of a rushed job is recognisable from a distance: one confident version and one that looks like a subtitle added at a client’s request.
In practice you do not equalise measured heights but optical weights: at the same size, Arabic setting looks lighter and wider. Look at the two versions side by side at forty pixels, not in a presentation where each of them occupies a page.
The shortcut we refuse is the commonest in the market: setting the Arabic name in a system typeface beside a custom-drawn Latin mark, then delivering the pair as a complete identity. That is not a bilingual identity, it is an identity and a caption.
Where your identity is actually seen
There is a material constraint that decides half the choices in this pillar, and it is about size rather than taste.
The internet market observatory published by ARPCE counts, for the second quarter of 2025, some 59.10 million internet subscriptions in Algeria, of which 88.71% are mobile and 11.29% fixed.
The consequence usually drawn from that concerns advertising. For a brand it concerns dimensions: the first appearance of your mark is neither a sign nor a letterhead, it is a forty-pixel profile thumbnail and an image a platform has recompressed.
A mark needing air, fine detail and a secondary line of text does not exist at that size. It becomes a grey smudge, and that smudge is your identity for the majority of people who will ever meet it — including those who see the sign later.
The test costs nothing: export your mark at forty pixels, put it on a phone beside two competitors’, and ask somebody which one is yours. If the answer hesitates, the problem is the drawing and not the platform.
- Mobile subscriptions88.71%
- Fixed subscriptions11.29%
ARPCE, internet market observatory, second quarter of 2025
Print: what your printer will change without telling you
The colour you approved on a screen is not the colour that will come off a press. A screen emits light and ink absorbs it: some bright screen colours simply have no printing equivalent, and the printer will pick the nearest one without calling you.
The substrate changes it a second time. The same ink on uncoated paper is flatter and darker, on recycled board it dulls, on plastic it saturates. A colour does not exist on its own: it exists on a material.
The safeguard is old and nobody asks for it: require a proof on the real substrate before the run, look at it in daylight, and sign it. That is the only comparison that means anything, because it compares an object with an object.
We will quote no figures here on printing costs or colour deviation in Algeria: we know of none that is dated and checkable. That is why this pillar carries almost no charts, and the article on visual identity explains it at length.
The habit that replaces the figure is free: keep one printed, signed copy, physically, in a drawer. Every later run is compared with that rather than with a screen, and the drift that settles in over three years becomes visible in three seconds.
Packaging is judged on a shelf, at a metre, next to a competitor
Packaging is never looked at the way it was designed. It is seen at a metre, in shop lighting, at an angle, flanked by two products in the same trade as you that chose their colours without consulting you.
The matching test costs the price of one print: print the label at real size, stick it on the real container, put it on a shelf between two competitors and step back a metre. Whatever you cannot read from there, nobody will read.
Then there is the part mock-ups leave until last and the only part that is compulsory: the weight, the origin, the composition, the contact details, the declarations specific to your product category. Find out about those obligations before the layout rather than after it.
A layout that can only fit the compulsory text by shrinking it past legibility is a failed layout, whatever its quality elsewhere. That text is not a constraint added to a design: it is part of the design brief.
Finally, one thing changes on a shelf and never changes on a screen: the neighbours. Your colour has to work beside theirs, which makes the choice a decision about them as much as about you — and that is why it is taken in a shop rather than in a meeting.
The sign: the only medium that has to work at night and at a distance
A sign is read from twenty metres, at an angle, often at night and sometimes from a moving car. No other medium in this pillar works under those conditions, and none of them costs as much to redo.
The website logo enlarged nearly always fails. Thin strokes vanish at a distance, tight letters merge, a gradient becomes a mass, and a secondary line that works on screen becomes a grey stripe nobody deciphers.
A sign therefore usually needs a simplified version of the mark: fewer details, heavier strokes, open spacing. That version is part of the identity and is delivered with it; when it does not exist, the sign maker improvises it, and they will improvise it in their workshop without you.
There is also a hierarchy people often invert: at a distance a name is read faster than a symbol is decoded. If something has to be sacrificed, it is the size of the symbol and not the size of the name.
The test is free and nobody does it: photograph the shopfront at night with your phone, then look at the photograph instead of looking at the shopfront. That photograph is what a driver sees, and it is more honest than your own eyes, which already know what is written.
What can be measured on an identity, and what cannot
Three things can be measured in this pillar, and none of the three needs a tool.
The first is recognition: show your mark for two seconds to ten people who are not your customers and ask them what trade it is in. Their answers are worth more than a specialist opinion, because they reproduce the real conditions.
The second is the number of versions in circulation. Count, today, how many variants of your logo exist across the sign, the invoices, the social thumbnails and your salespeople’s presentations. Above three, your problem is discipline rather than design.
The third is the number of reprints: how often you have reprinted something because it was wrong. It is the one figure in this pillar that turns an aesthetic problem into a spend, and therefore the one a board can act on.
What cannot be measured: beauty. It is a genuine benefit, it does not justify a spend on its own, and a supplier building their case on it is selling you their taste rather than your result.
What to check before signing
The first check is about the files: you have to receive the editable sources, in vector form, and not only exports. A flat image cannot be corrected, cannot be scaled up for a sign, and forces you back to whoever holds it.
The second is about the typefaces and their commercial licence, in your name. A typeface used without a licence is an invoice that arrives later, and it arrives at the moment your brand is too widespread to be changed quietly.
The third is about ownership of the artwork, and it has to be written down. It is the same rule as for software source code, and it is discovered the same way: at the moment of separation.
The fourth is about Arabic: is the version included or billed separately, and was it drawn or merely set in a system typeface? The answer to that single question sorts a large part of the market.
The fifth is a test: ask what they would refuse to produce. A supplier who accepts everything does not have a method, they have an order book, and in this pillar an order wrongly accepted is paid for in manufactured objects.
What we do, and what we will refuse to do
What we will refuse: designing an identity on a name whose availability has not been checked. Clients often treat that check as secondary, it happens before the quotation, and it has already lost us projects we were glad to lose.
We will refuse to deliver a Latin mark on its own as a complete identity to a business trading here. We would be billing for finished work while delivering half of it, and the missing half is the one part of your customers will read.
We will refuse to start a print run with no signed proof on the real substrate, and we will refuse to redo an identity whose only problem is that six versions are in circulation. That second case is an inventory, it costs a fraction of a redesign, and saying so costs us the difference.
What we do: the decisions written down before any object; the mark tested at forty pixels and at a metre before it is cut into anything; Arabic drawn rather than set; the sources, the licences and the ownership in your name.
And what you should do without us this week: show your mark for two seconds to ten people who do not know you, and count how many versions of your logo are circulating in your business today. It is free, it takes a day, and those two figures say whether to redo or merely to tidy.
Frequently asked questions
Should we start with the logo?
No, with the name and its availability: it is the only step whose failure makes all the others pointless, and it is the cheapest. The drawing comes next, and the irreversible objects — packaging, signage — come last.
Can our current name stay?
Put it through three tests: can it be spelled on the telephone without repetition, is it written the way it sounds, and does it hold in all three languages without becoming an awkwardness? If it fails the first or second, the cost is paid daily.
Is an Arabic version really necessary?
Yes for a brand trading here, and it is not a translation: Latin letterforms have no Arabic equivalents, so an Arabic logotype is drawn. Setting in a system typeface beside a custom mark is not a bilingual identity.
Should we redo our identity or repair it?
Count the versions in circulation first. Above three, your problem is discipline rather than design, and an inventory costs a fraction of a redesign. Redoing is justified when the mark fails the size and distance tests.
What should we require from a printer?
A proof on the real substrate before the run, looked at in daylight and signed. Keep that copy physically: every later run is compared with it rather than with a screen, and drift becomes visible immediately.
How long before an identity has an effect?
It does not produce sales by itself; it reduces a friction at every contact. What does show quickly is the reverse: a mark illegible at forty pixels or at twenty metres costs you immediately, every day.
Where we come in
The two free measurements hand you a symptom, not a cause: ten strangers hesitate, seven versions of the logo are in circulation, and neither says where to start.
- We sort the objects by how reversible they are, from a file to an injection mould.
- We date what has to be remade and what will simply run its stock out.
- We write down the exact name of the problem: drawing, discipline or name.
When the answer is to remake nothing this year, that is the answer you get, and it is not invoiced.
Read next
Rebranding: what you already have, what has to change together, and what can wait
Almost nobody creates a brand — most inherit one assembled piece by piece. A rebrand is a replacement operation, not a design project.Having a sign installed: the quotation, the permit, the maintenance
The design is approved; what remains is three comparable quotations, a permit, and lighting that still works in two years.A brand in service: what everybody has to be able to say
A brand decided in a meeting and a brand spoken on the telephone are two different objects. How to make the second one hold.
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.