The Phantom Bar Problem
You're on a video call in the far bedroom. The router is two floors down, forty feet away, and your laptop is showing three confident Wi-Fi bars. You assume you're fine. Then the call freezes, your face tiles into a mosaic, and you spend thirty seconds staring at a spinning indicator that has the audacity to appear over your own home network.
Those three bars lied. Signal strength and usable throughput are not the same thing, and your device latched onto the only access point it knew about (the one in the basement) and is now squeezing data through walls, a floor, and a load-bearing column. A traditional range extender would nominally fix this. In practice, it usually makes things worse in ways that are harder to diagnose.
Mesh networks do fix it. The reason comes down to a single architectural difference that most explainers skim past.
One Network vs. a Chain of Compromises
A range extender receives your router's signal and rebroadcasts it. Simple enough. But that rebroadcast eats bandwidth before your device ever sees a packet.
A typical dual-band extender grabs the incoming signal on one radio and retransmits on the same radio, which immediately halves your available throughput. That's not a bug or a cheap-hardware quirk. It's physics. The radio can't listen and shout at the same time, so it alternates, and you pay for every alternation with a throughput cut that compounds with distance like interest on a loan you didn't know you'd taken out.
Some extenders use both bands to sidestep this: receive on 5 GHz, rebroadcast on 2.4 GHz. Better, technically. But now your devices are on a different subnet with a different network name, and your phone won't hand off automatically when you walk from the kitchen to the garden. You reconnect manually. You've solved coverage but broken continuity, and that trade is a bad one.
Mesh systems are built around a different assumption entirely: every node is a full peer, not a relay.
Take a three-node mesh setup, say an Eero Pro 6 or a TP-Link Deco XE75. Each node shares the same SSID and password. Each node talks to the others over a dedicated backhaul channel, often a separate 5 GHz or 6 GHz radio reserved exclusively for node-to-node traffic. Your devices never see that backhaul conversation. They just see one network, everywhere, behaving like a single coherent thing rather than a series of increasingly desperate handshakes.
When you walk from the living room to the garage, a process called band steering and fast roaming (the technical standard is 802.11r, if you want to look it up) hands your phone from one node to the next without dropping your session. The call doesn't freeze. The handoff takes milliseconds.
A Tale of Two Houses
Picture two neighbours, Marcus and Yemi, who both move into identical 2,400-square-foot Victorian terraces on the same street. Both have routers in the front room and dead zones in the rear kitchen extension.
Marcus buys a dual-band range extender. Setup takes ten minutes. Speed near the router: 400 Mbps. Speed at the extender, one wall away: 190 Mbps. Speed in the kitchen, behind the extender: 55 Mbps. Each hop slices the bandwidth roughly in half. His smart TV buffers on 4K streams, and his phone clings to the wrong access point whenever he moves between rooms.
Yemi spends more and installs a two-node mesh kit. The second node sits in the hallway, connected to the first over a dedicated backhaul radio. Speed in the kitchen: 310 Mbps. His phone hands off without a hiccup. The TV never buffers.
Same house shape, same ISP, same distance from the router. Different architecture, different outcome. The hardware gap between those two scenarios is maybe eighty dollars.
The Misconception That Costs People Money
The most common mistake is treating mesh networks as expensive range extenders with better branding. They're not, and the distinction is worth being direct about. The price gap reflects genuine engineering: dedicated backhaul radios, centralised network management with one app and one firewall, and firmware that actively monitors which node each device should connect to and nudges it if it's clinging to a weaker one.
That last part matters more than people expect. Devices are sticky. Your phone will hold onto a distant node long after a closer one would serve it better, because switching costs battery and the device's firmware is conservative about it. Mesh systems solve this with client steering, where the network itself tells a stubborn device to move. Range extenders can't do this. They're passive. They broadcast and hope.
And here's something the marketing copy tends to bury: wired backhaul beats wireless backhaul, every time. If you can run an Ethernet cable between your mesh nodes, even cheap ones, you eliminate the backhaul bottleneck entirely. A modest mesh kit with wired backhaul will outperform a premium kit relying on wireless backhaul in almost every real-world test. Powerline adapters, which send Ethernet signals over your home's electrical wiring, can approximate this where cable runs aren't practical.
Mesh isn't magic. A node placed behind a thick concrete wall or inside a metal cabinet will still underperform. Placement still matters, just as it always did.
Checking Your Own Setup
Most mesh apps (Eero, Google Home, Deco) show you a per-node signal quality score and a device list that tells you which node each device is connected to. Open it. If you find a phone connecting to a node two rooms further away than the nearest one, that's client steering failing, and repositioning the closer node by a few feet often fixes it.
If you're still running a single router with a range extender bolted on, run a speed test right next to the extender, then run one on the far side of it. If throughput drops by more than 50%, you're experiencing the half-bandwidth tax in real time. On a typical single-radio extender under real conditions, that number lands somewhere between 40 and 60% loss. Every time.
A phone that starts the day at full signal near the router and hits frustrating slowdowns in the back bedroom isn't showing you a battery problem or an ageing-device problem. It's showing you an architecture problem. And unlike a battery, the architecture is fixable without buying new hardware, just different hardware.
Range extenders solved the coverage problem of a decade ago. Mesh networks solved the network problem. The distinction is worth the price difference, and that price difference has narrowed enough that the extender is now the expensive option once you count the hours you'll spend troubleshooting it.