It’s Friday, eight o’clock. The bar is three deep. The kitchen is firing on all cylinders. And then it starts: a guest waves down a server because their card payment isn’t going through. Another can’t load the menu QR code. A third is asking your duty manager for the WiFi password for the second time, convinced they’ve typed it wrong.

You pull out your phone, check the signal, and everything looks fine. Speedtest runs at 200 Mbps. The router lights are green. And yet, by the bar, ten people are staring at spinning wheels.

I’ve walked into this exact scene more times than I can count. It’s almost always the same handful of problems, and almost never “the WiFi is down”.

The phrase “WiFi is down” almost always means something else

When a guest tells you the WiFi isn’t working, what they actually mean is “my phone isn’t doing what I expect”. That could be a hundred different things, but in a busy hospitality venue on a Friday evening, it’s usually one of five:

  1. The access points can’t handle the number of devices in the room.
  2. Everything is on a single flat network, and traffic is fighting itself.
  3. The DHCP server has run out of addresses to hand out.
  4. The internet line itself is saturated by something other than guests.
  5. The radio environment is so noisy that nothing can get a clean signal.

None of these are “WiFi being down”. All of them feel exactly the same to a guest. And all of them are fixable, but only if you know what you’re actually looking at.

Root cause 1: consumer-grade kit in a commercial venue

I once went to look at a 120-cover restaurant in Fitzrovia where the entire guest network ran off a single high-street router tucked behind the till. The owner had bought it because the box said “great for whole-home coverage”. On a quiet Tuesday lunch it worked fine. On a Friday night with a queue out the door, it fell over by half eight, every week, like clockwork.

Here is the maths nobody does before they buy the cheap box. A busy bar with 80 people in it doesn’t have 80 devices on the network. It has somewhere between 150 and 220. Each guest brings a phone, plenty bring smartwatches that join automatically, and a chunk of them will have a second device in a bag. Add staff phones, the POS terminals, the kitchen display, the music streamer, the CCTV recorder, the card machines, the booking tablet on the host stand, and you are well past 200 connected devices in a room.

A consumer router is built to handle perhaps 30 to 50 concurrent clients before it starts misbehaving. Not because the radio can’t see them, but because the CPU and the internal client tables can’t track that many sessions at once. It doesn’t crash. It just quietly stops responding to new associations and starts dropping older ones. From the floor, it looks like the WiFi is “going slow”.

The fix is proper commercial access points, more of them than you think you need, sized for the actual device count rather than the floor area. For most central London venues that means at least one AP per 30 to 40 concurrent devices, mounted properly, on power over Ethernet, not plug-in.

Root cause 2: one network, everything fighting

The second thing I see constantly is a single flat network with everything on it. Guest phones, staff phones, the POS, the kitchen printer, the CCTV, the office laptop, the card terminals, the music. All on the same VLAN, all on the same SSID, all able to talk to each other.

This is bad for two reasons. The first is performance: chatty devices like printers and CCTV recorders broadcast constantly, and on a flat network every guest device has to listen to all of it. The second is security. A guest phone with a dodgy app on it is now sitting on the same network as the till. That is not a position you want to be in, and it is the first thing we look at when we do a cyber security review for a hospitality client. Guest isolation, rogue device detection, and segmentation aren’t optional any more.

The fix is straightforward but fiddly: separate SSIDs and separate VLANs for guest, POS, staff, and management. Guests can never see each other or anything else. Staff devices live on their own segment. Payment terminals are isolated entirely, which is also what your card processor expects under PCI DSS.

Root cause 3: DHCP exhaustion

This one catches people out because the symptoms are odd. The WiFi works fine for the first few people, then partway through the evening new guests can connect to the SSID but never get an IP address. They sit on “obtaining IP” forever, or they get a self-assigned address and nothing works.

What’s happening is that the DHCP pool, the range of addresses the router can hand out, is too small for the number of devices passing through, and the lease time is set absurdly high. A typical home router ships with a /24 subnet (around 250 usable addresses) and a 24-hour lease. That means a phone that connected at lunch is still holding its address at midnight, even though the guest left hours ago. By eight on a Friday you’ve run out, and every new arrival sits on the spinning wheel.

The fix is a bigger pool, a much shorter lease (we usually run guest WiFi at 30 to 60 minutes), and monitoring so you actually see when you’re approaching the limit instead of finding out from a complaint.

Root cause 4: the upstream is the bottleneck

Plenty of venues have decent kit inside the building and one tired FTTC line going out of it. That single line is shared between guest WiFi, the POS uploads, the office, the CCTV pushing to cloud storage, the music streaming service, and whatever the manager is downloading in the back office. On a Friday evening, the CCTV is backing up the day’s footage at exactly the moment the guests want to scroll Instagram.

You can have the best access points in London and it won’t matter, because everything is queueing for the same 70 Mbps uplink.

The fix is two things. First, put a proper business circuit in, ideally a leased line or dedicated fibre, with a second connection from a different carrier as failover. Second, use a router or firewall that can actually do bandwidth management and quality of service, so payment traffic and POS sync always get priority over a guest streaming Netflix at the bar. This is the kind of thing we build into every managed network we run for hospitality clients, and it’s the single biggest difference between a venue that copes on a Friday and one that doesn’t.

Root cause 5: the airwaves are full

If you operate in Soho or Covent Garden, or honestly anywhere in central London, you are not the only WiFi network in the building. You’re one of forty. Each of those networks is broadcasting on the same handful of channels, and on the 2.4 GHz band in particular the situation is hopeless. There are only three non-overlapping channels, and every consumer router defaults to picking one of them and shouting over the top of everyone else.

The result is a noisy, congested radio environment where even a perfectly configured AP struggles to hold a conversation with a phone six feet away. Guests see drop-outs, slow page loads, and intermittent disconnections that have nothing to do with your kit.

The fix is partly hardware (commercial APs that prefer the 5 GHz and 6 GHz bands, where there’s much more room), partly design (channel planning, transmit power tuning, careful AP placement so coverage cells don’t fight each other), and partly survey work. You cannot fix this from a desk. Someone has to walk the venue with a spectrum analyser at the time of day the problem actually happens.

How to design guest WiFi that survives a Friday

If you want a network that doesn’t fall over when the venue gets busy, the work happens in this order:

  • Site survey first. Walk the building, measure the existing radio environment, count the concurrent devices on a real busy night, find the dead spots and the interference sources.
  • Design for device count, not square metres. Size the access points around 200-plus concurrent clients in a busy bar, not the marketing brochure’s “covers 3,000 sq ft”.
  • Segment everything. Separate SSIDs and VLANs for guest, staff, POS, and management. Guest isolation on by default.
  • Resilient upstream. Two circuits from different carriers wherever possible, with automatic failover and bandwidth management so business-critical traffic always wins.
  • Captive portal for guests. Branded, simple, with a sensible session timeout and short DHCP leases. Optionally a marketing opt-in if it suits your brand.
  • Continuous monitoring. You should know about a problem before a guest does. We watch client counts, channel utilisation, DHCP pool usage, and uplink saturation in real time, and we get alerted before the floor starts complaining.

What we actually do

Most of our hospitality clients call us after a few too many Friday nights like the one I described at the top. We come in, do a proper survey, design a network that fits the venue, install it cleanly (cable runs done out of hours, no disruption to service), and then we monitor and manage it from there. When something goes wrong we usually know first, and we fix it before the duty manager has noticed. That’s what we mean by IT support for hospitality - not a number you ring after the problem has cost you a service.

If your guest WiFi is the thing your team dreads on a Friday night, it doesn’t have to be. Get in touch and we’ll come and have a look at the venue.