AO, it isn't actually daft if you look at the big picture.
The network airlines generally sell flights up to 360 days out. Based on histories as each flight comes on-line they will have estimated how many 'local' sales and how many 'returning' passengers they expect. In each class of service, the difference between the aircraft capacity and the expected passenger load, is an opportunity.
They expect all of those remaining seats to be empty, so any revenue they can get for them, for example from a passenger originating in BRU, is gravy. Of course they have to factor in the demand for the 'married' sectors (in your example BRU-LHR and LHR-BRU) to ensure they are not displacing passengers on those flights. They also need to look at if other origins (for example GVA) might give them higher total revenue. A transatlantic flight ex LHR could be 'sold-out' for BRU originating passengers, but available for GVA originating passengers if the numbers are right. This does happen.
So when they are looking at pricing BRU/GVA-wherever via LHR they need to set it at the highest value that they believe, on the day, will still attract a sale. This will often be lower than the prices to/from their hub - but remember this is all gravy. In most markets they can't set it at, or higher than non-stop flights offered locally, they need to be lower. How much is dependant on local market conditions.
The game is to sell every seat on every plane. And selling beyond hub for less that the hub is irrelevant to the airline. It's gravy!
Oh, I mentioned they do this analysis when the flight first comes on-line. It isn't just then. They constantly analyse and reanalyse ever flight in their network multiple times every day right up to departure time. Almost all of it is automated, with some human intervention, but the objective remains: fill that plane with the maximum amount of revenue ...