That's a very interesting question.
From the earliest days in 1965 it was realised that some parts of the metrication process woukd be harder than others and re-doing all the road signs would be expensive. It would have to get past the treasury to be implemented.
In the late 90s, in the run-up to the last round of metrication legislation, there was reckless talk of how much it might cost to do the roads and the figure of one billion pounds was mentioned somewhere. I did a Freedom of Information request to find out what studies the Department of Transport had done on the subject, only to find that they hadn't done anything at all. They had no idea what it would cost.
The Irish switchover to metric roads was done in two stages, several years apart. First the main routes, which were in any case built by EU infrastructure funding, had distances marked km, so no cost there. Then smaller roads were done over a period of years.
For many years they seemed happy to have the speeds in mph and the distances in km. In high tourist season you could have a line of cars behind a foreign tourist doing 50 km/h (30 mph) on a 50 mph section of road and they realised they would have to do something. I remember being at how low the eventual cost was. I think it was something like €7 million. The UK road network is quite a lot larger than the Irish network but I think we're talking about a few tens of of millions of pounds. Certainly nowhere near a billion.
From a Keynesian point of view this is exactly the sort of thing that government should do in a downturn. All the mony is spent in the UK and benefits the UK. More jobs, and the multiplier effect spreads it through the economy. Well worth doing, in my opinion.