Algeria AgencyIT & infrastructure
Professional or consumer Wi-Fi: what actually changes
The speed printed on the box was never the problem. The problem is how many devices are connected at once.
A business buys a consumer router, puts it behind the counter, and everything is fine for three weeks. Then the service degrades every evening at the same hour, and nobody understands why since "the connection is the same".
The connection is indeed the same. What changed is the number of devices sharing it, and that is exactly the point on which consumer and professional equipment do not behave alike.
This article explains that difference, what ignoring it costs, and how to tell whether your installation is at fault or your subscription is.
The figures quoted are technical and sourced under each table. Where an Algerian measurement exists — the structure of internet subscriptions — we use it; none exists for Wi-Fi performance and we do not invent one.
The advertised speed does not describe your situation
The numbers printed on boxes are theoretical maximum rates: about 3.5 Gb/s for Wi-Fi 5, up to 9.6 Gb/s for Wi-Fi 6. They are obtained in conditions that exist nowhere — one device, no obstacles, no neighbours.
In real conditions with a single client, the measurement is more like 400 to 600 Mb/s on Wi-Fi 5 and 600 to 900 Mb/s on Wi-Fi 6. Already a long way from the label, and still unrelated to your problem.
Because your problem is not one device. It is thirty, on a Friday evening, and at that point these numbers predict nothing at all.
It should be added that these rates are shared, not guaranteed per device. The advertised figure describes what the radio can carry in total, not what each device receives; it is the same difference as between the width of a road and the speed you travel on it on a Friday evening. Almost all commercial confusion about Wi-Fi sits in that sentence.
| Standard | Theoretical rate | Real, one client |
|---|---|---|
| Wi-Fi 5 (802.11ac) | 3.5 Gb/s | 400 to 600 Mb/s |
| Wi-Fi 6 (802.11ax) | 9.6 Gb/s | 600 to 900 Mb/s |
802.11ac and 802.11ax specifications and vendor measurements, 2026 figures
The number that counts: throughput per device under load
Here is the measurement that actually describes a business. In a dense environment with fifty connected devices — phones, laptops, assorted things — a Wi-Fi 5 router delivers an average of 22 Mb/s per device. In the same conditions, a Wi-Fi 6 access point holds above 60 Mb/s per device.
Nearly triple, at identical load and on the same subscription. It is the only gap in this article that corresponds to what your customers feel.
Note what that figure implies in the other direction: at fifty devices, consumer equipment gives each one less throughput than a decent mobile connection. Your customers do not complain about a number, they notice their phone works better than your Wi-Fi and they disconnect from it.
A note on the scope of this measurement: it was taken in a dense environment, which corresponds to a full café, not an office of three people. If your premises never hold more than ten devices, the gap narrows sharply and the investment is debatable. The figure does not say "buy", it says "from what point does this start to matter".
Comparative measurements in a dense environment, 2026 figures
Why the difference only appears in numbers
An access point does not talk to everyone at once: it talks to one device, then the next. Each exchange occupies the channel for a moment, and everyone else waits their turn.
Wi-Fi 6 brings two mechanisms that change that arithmetic. OFDMA lets a channel be split so several devices are served in one transmission, instead of each waiting its turn. MU-MIMO lets the access point talk to several devices simultaneously on separate spatial streams.
With five devices the waiting is invisible and the two are equivalent — which is why the test in a closed shop never shows anything. With fifty, the waiting is the entire experience, and that is where the two separate.
A practical consequence of this mechanism, often counter-intuitive: one old device can slow everyone down. A phone that only speaks an outdated standard occupies the channel for longer to send the same amount of data, and everyone else waits meanwhile. That is why capacity depends not only on your hardware but on your customers’ too.
A consequence for purchasing, often misunderstood: the newest standard is worth nothing if your customers do not speak it. A Wi-Fi 6 access point delivers its gains with Wi-Fi 6 devices; facing an older fleet it behaves much like its predecessor. So the question is not only how many devices, but which ones.
The difference throughput does not measure
Professional equipment brings things no speed test reveals, and which often weigh more than throughput itself.
A separate guest network first: customers on one network, the till and workstations on another, with no bridge between them. On a consumer router everyone is on the same network as your payment system, which is a security problem before it is a performance problem.
Then control: limiting bandwidth per device so one download does not take everything, capping the number of connections, setting hours. And finally visibility — knowing how many devices are connected and which are saturating — which is what turns "the Wi-Fi is not working" into a diagnosis.
Add a capability discovered after the fact that often justifies the price gap on its own: the captive portal. Being able to show a page, ask for acceptance of terms and cap a session’s length changes the nature of the service — and, in a venue open to the public, lets you know how many people connected, which no consumer router will tell you.
Add a capability used one day a year that, on that day, is worth its price: being able to isolate or cut the guest network without touching the rest. During an event, an exceptional crowd or an incident, that control lets you protect the till in thirty seconds, which no consumer router allows you to do cleanly.
An access point is not a router placed higher up
The commonest confusion is buying a second router to cover the back of the premises. Two routers create two separate networks: a device that joined the first clings to it even when the second would be better, because nothing tells it to switch.
The result is the classic case of a customer sitting three metres from the second access point, connected to the first at the other end of the room, on two bars. Adding equipment made the service worse, not better.
Professional access points are managed together: one network name, and the device is steered to whichever point serves it best. That coordination is what you are buying, not transmit power.
There is a reason this mistake is so common: it is encouraged by the sales vocabulary. Packaging talks about "coverage" and "area covered", never about capacity or coordination, because area is what compares on a shelf. The word missing from the box is precisely the one that decides the outcome.
Placement matters more than price
An expensive access point behind the counter, on the floor, against a load-bearing wall, will cover worse than a decent model fixed to the ceiling in the middle of the room. The signal is absorbed by anything dense: concrete, metal, water — and a room full of people is a room full of water.
Two mistakes are universal. The first is placing the equipment where the line arrives, which is to say in a corner. The second is putting it on the floor or behind furniture, because that is where the socket is.
The rule is simple: on the ceiling, in the middle of the area to cover, unobstructed. That often means pulling a cable, and that cable is the investment with the largest difference per dinar spent in this whole article.
One budget consequence deserves stating plainly: between an expensive access point badly placed and a decent one well placed, the second wins almost every time. If your budget does not allow both, take the cheaper hardware and pay for the cable. It is one of the few trade-offs on this blog where the less expensive option is also the better one.
A useful check before having a cable run: ask where the conduit will physically go, and do it while the quotation is still open. A cable route crossing a stockroom, an unreachable false ceiling or a load-bearing wall can double the installation cost, and it is the one item on this job that is hard to correct afterwards.
The number of access points is calculated in devices, not metres
The usual question is "what area does it cover". That is the wrong question: coverage is almost never the limiting factor in commercial premises, capacity is.
The right question is how many devices will be connected at the busiest moment. Count the customers at peak, multiply by the devices each carries — a phone, often a second screen — and add your own equipment, which is more numerous than people think: till, payment terminal, cameras, printer, television, thermostat.
That total determines the number of access points. A hundred square metres with fifteen people and a hundred square metres with eighty do not have the same requirement, and no floor area says so.
A method note for doing this count without spending an evening on it: read the number of connected devices from your current equipment’s interface, at peak hour, on three different days. You will get a real figure rather than an estimate, and it is nearly always higher than the owner reports from memory.
Why your customers care, in Algeria
According to ARPCE’s observatory for the second quarter of 2025, out of 59.10 million internet subscriptions, 88.71% are mobile and 11.29% fixed.
That proportion explains why a venue’s Wi-Fi is not a gadget. For the large majority of your visitors, your access point is the only fixed connection they will use all day: it is where they download, update, and send heavy files.
That means two things. That the service is expected, and that it will be used intensively per device — which leads straight back to the figure in the second section. Venue Wi-Fi in Algeria carries more load per person than it would elsewhere.
A commercial consequence is worth adding, because it escapes equipment calculations: in many venues, the Wi-Fi is the reason a customer stays an extra hour. It is not a side service, it is dwell time, and dwell time converts into spending in a way network hardware never measures.
- Mobile subscriptions88.71%
- Fixed subscriptions11.29%
ARPCE, internet market observatory, second quarter of 2025
When the problem is not the Wi-Fi
Before replacing anything, you need to know whether the bottleneck is your local network or your subscription. Both produce the same complaint and are fixed differently.
The test is simple: plug a computer into the router by cable and measure. If the wired speed is fine and the Wi-Fi is not, the problem is your installation. If the wired speed is bad too, the problem is your line, and changing access points will change nothing.
Redo that test at peak hour, not in the morning. The distinction between the two causes only appears under load — which is also the only time anybody complains about it.
A third case exists, rarer and more frustrating: the line is fine, the local network is fine, and the problem is one device saturating everything — a camera uploading continuously, an automatic update launched in the middle of the day. That is exactly what the per-device limiting mentioned earlier exists to prevent.
What not to do
Do not add a repeater. A repeater retransmits what it receives on the same channel: it roughly halves available capacity while extending the area. It solves a coverage problem by worsening a capacity problem, and capacity is almost always the one you have.
Do not hide the network name believing it secures something: it achieves nothing and complicates connecting for your customers. Security comes from network separation and current encryption, not from discretion.
And do not leave the equipment’s default password, or its administration reachable from the guest network. It is free to fix and it is what we most often find on an audit.
One last thing not to do, and it is tempting: sharing the guest network password by taping it to the counter and never changing it. A public, permanent credential ends up circulating well beyond your customers, and you find yourself supplying free connectivity to the neighbourhood on capacity you are paying for.
One last remark before buying anything: redo this test after every change, not only before. That is how you will know whether the new hardware produced the expected effect, and it is also how you will sometimes discover that the line had become the limiting factor in the meantime.
Where to start
Count the devices at your busiest hour. Not the people: the devices, including your own. That number is what sizes the installation, and almost nobody has ever written it down.
Then run the wired test at that same hour, to know whether you are talking about a network problem or a subscription problem. Those two measurements cost an hour and prevent buying the wrong thing.
Then, if it really is the network: a separate guest network, access points managed together, and a cable to the ceiling in the middle of the room. In that order, and often without replacing everything.
And keep the result of those two measurements in writing, with the date. If you change hardware later, they are the only figures that will let you say whether the new equipment helped — and if you change nothing, they will tell you in a year whether your footfall has grown enough that the question now looks different.
One last thing about hardware, whatever you choose: write the model and installation date down somewhere. The day somebody else works on this network, that line will save an hour, and it is exactly the kind of information nobody can find at the moment it is needed.
What we do, and what we will tell you not to buy
We start with the two measurements from the previous section, on site and at your peak hour: the real number of connected devices, read from your current equipment, and the wired test on the router. Those two figures decide whether you have a network problem or a subscription problem, and they cost one visit.
The answer is often "change nothing for now". An access point moved to the ceiling in the middle of the room, a guest network separated from the till and a properly routed cable resolve a good share of cases with no hardware purchase, and we say so when it is true — it is also what makes the quotation credible when replacement genuinely is needed.
When replacement is justified, we size on the number of devices at the busiest hour rather than on floor area, install access points managed together under one network name, and separate the customer network from the till and workstations. That separation is done as a matter of course, including when it was not asked for.
And we leave behind things you can check: passwords changed, administration unreachable from the guest network, per-device limiting so one download cannot take everything, and the device count recorded with its date. That last point will let you know, a year from now, whether the question needs asking again.
Frequently asked questions
Is Wi-Fi 6 worth the extra cost?
If you have many simultaneous devices, yes: the dense-environment measurement gives above 60 Mb/s per device against 22 on Wi-Fi 5. With five devices the difference is invisible.
How many access points do we need?
It is calculated on the number of devices at peak, not on floor area. Count the customers, multiply by the devices each carries, add your own equipment.
Can a repeater help?
Rarely. It retransmits on the same channel and roughly halves capacity. It extends coverage while worsening capacity, which is the problem you almost always have.
Should the customer network be separated?
Yes, always. A guest network isolated from the till and the workstations is a security question before it is a performance one, and it configures with no extra hardware on professional equipment.
How do we know whether the subscription is at fault?
Plug a computer into the router by cable at peak hour. If wired is good and Wi-Fi is bad, it is the installation; if wired is bad too, it is the line.
Does everything have to be replaced at once?
Almost never. Placement, network separation and one well-routed cable fix a large share of cases before any hardware purchase.
Where we come in
The number of simultaneous devices and the wired speed taken at the same moment almost always settle it, and both are yours to take.
- We take both measurements at your premises, at your peak hour rather than ours.
- We move one access point before proposing three.
- We leave the passwords changed and administration closed from outside.
The answer is often to change nothing for now, and that is the one we give when it is true.
Read next
The guest network: what your customers reach, and what they cost you
A guest network exists the moment a waiter gives out the password. What it should reach, what it takes off the line, and what not to collect.Networks: the cheapest layer, and the one that produces half the faults
A network is a chain: it is worth what its worst segment is worth. That is why it gets bought in the order the signal travels.The line is down: establishing which side, and what comes next
Five minutes decide whether the fault is yours or the operator’s. What to record before calling, how to escalate, and what a second line really covers.
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.