Why placement matters more than the router
Wi-Fi is radio, and radio gets weaker with distance in a way that is easy to underestimate. Double the distance and you lose a large slice of the signal before anything else has happened. Then put a wall in the way and you lose more again, and the amount depends on what the wall is made of. A stud partition costs you a little. A solid brick chimney breast costs you a lot. A concrete floor costs you more than either.
That is why two homes with identical broadband and identical routers can have completely different Wi-Fi. It is also why the most effective upgrade in most British houses is free: pick the router up and put it somewhere better.
The four things that matter, in order
1How central it is
The router radiates outwards in every direction. Sat against the front wall of a terrace, roughly half of everything it transmits goes straight out into the street. Moved into the hall, the same router is covering the house rather than the pavement. Central beats powerful nearly every time.
2How high it is
This one gets repeated a lot, so it is worth saying what the evidence actually shows. Radio engineers describe a device as clutter embedded when its aerial sits below the average height of the furniture around it, and clutter elevated when it sits above. A router on the carpet is embedded in the sofa, the sideboard and the skirting; the same router on a shelf is not.
The effect is not a fixed number of decibels. It changes the rate at which signal falls away with distance, which is why the difference is small in the same room and large at the far end of the house. Work by Kacou, Guillet, El Zein and Zaharia, presented at the European Conference on Antennas and Propagation in 2018, looked specifically at transmitter height in a multi room home between 800 MHz and 6 GHz and found path loss falls as height rises. The size of the effect used here is a calibration choice within the range that work implies. One much cited comparison of low and high aerials found the distance exponent moving by about 0.38, but it was measured in a factory at a far higher frequency than home Wi-Fi, so it fixes the mechanism rather than the number. The full working is on the methodology page.
So yes, lift it up. Chest height on a shelf or a table, out in the open, with the aerials upright. Just do not expect it to fix a room three walls away on its own. Switch the router between the four heights in the panel and watch which rooms actually change.
3What is next to it
Metal reflects and absorbs radio. Large metal objects near the router make a measurable difference: a fridge, a washing machine, a filing cabinet, a mirror with a metallic backing, a wall mounted television. A running microwave transmits in the same band as 2.4 GHz Wi-Fi. A router inside a media unit behind the telly is in about the worst place a living room can offer.
4What is between it and the rooms you use
Count the walls between the router and the room where the Wi-Fi annoys you. If the answer is three, no amount of new hardware in the same spot will fix it. Signal upstairs mostly arrives through the stairwell rather than through the joists, which is why a position near the bottom of the stairs works so well in a two storey house.
What about mesh, extenders and powerline?
Try placement first, because it costs nothing. If the best available position still leaves part of the home short, the order to work through is: run a longer cable and move the router, then add a mesh point, then wire that mesh point back to the router if a cable can be run, then consider powerline adapters if it cannot.
One thing worth knowing about extenders: a repeater can only pass on what it can hear. Put one in the weak room and it repeats a weak signal. The right place is roughly halfway between the router and the problem, where it still has a strong link back. You can try this on the plan: add a mesh point, drag it around, and watch the number next to it. When its link back to the router drops, everything it serves drops with it.
And before buying any of it, check what a new deal would give you. Most providers now bundle a Wi-Fi 6E or 7 router, and some a mesh disc, with a new contract. If you are out of contract you may be paying more than a new customer for an older box; comparing deals can turn the router upgrade into a saving.
Neighbours matter too, and in a different way. Their networks do not weaken your signal; they raise the noise it has to shout over. That is why a flat can show a strong signal everywhere and still feel slow. Switch the neighbours setting to a busy block and watch the colours stay put while the speeds fall.
Where the master socket comes into it
Most routers end up wherever the socket is, which is usually by the front door or in the corner of the front room, because that is where the line came into the house. Neither is where you would choose. A telephone extension lead, or an ethernet cable from the socket to a better position, is a few pounds and is very often the entire fix. On full fibre, the box on the wall is fixed but the router can still be moved with an ethernet run.
How this tool estimates coverage
The model starts from the transmit power of a typical domestic router, subtracts the loss over distance, subtracts an amount for every wall the signal passes through based on what that wall is likely to be made of, and subtracts more again for each floor it crosses. Doorways are treated as real gaps, because signal spilling through an open door is a large part of how a house is covered. The distance term follows the log distance indoor model described in ITU-R P.1238 and the wall by wall approach follows the multi wall model published as part of COST 231.
The threshold drawn as a dashed line on the plan is minus 67 dBm, which is a widely used design target for reliable Wi-Fi rather than a hard cut off. Above it, most things work. Below it, video calls start to stutter first and everything else follows.
The figures are estimates of a typical home of that shape, not a survey of yours. Use it to understand what is happening and where to try, then confirm with a speed test in the actual room.