Depending on capacity, it could be as simple as running a new service to create a "Crossrail" for a given large(ish) settlement.
Clearly, large cities probably wouldn't work (unless you skirt the centre, to give further out places a higher frequency of services but to get to the centre on the extra services they'd have to change).
However, for somewhere with a mainline through it which has some capacity (just not heading to London, Manchester, or the other large cities) it could be possible.
The example I've given in the past is to run a service from Ascot to Basingstoke via a new grade separated junction to allow free movement Frimley / Farnborough Main.
The lines are fairly lightly used.
The slows through Farnborough max out at 4tph in the peak hours and whist there's some freight, it shouldn't be difficult to add 2tph (maybe 8tph including freight).
The line through Camberley sees 2tph, so this going to 4tph (bearing in mind the extra services wouldn't be using the single track section south of Frimley) shouldn't be impossible.
Whist each place is fairly small, the total population for the Aldershot Urban Area is around 250,000, Basingstoke is around 105,000, that's quite a potential for users.
Where it gets interesting is that there's quite a few places where additional stations could be useful (as an example Fleet station is almost at the eastern edge of the settlement and so a western station could serve other areas, including two large developments built in the last 30 years) this would make rail travel far more attractive for a far larger percentage of the population.
Based on a journey time of an hour (end to end) to run 2tph you'd need 5 units.
Based on the rough rule of thumb of £200,000 in lease costs per coach and that's 1/3rd of total costs, that's £4 million to run that service.
Could that generate enough income to cover it. At 12 hours a day (yes you'd actually run for 15-18 hours, but some of those hours would be lightly used) for 300 days a year (allowing for Sundays, Bank Holidays and engineering works to generate zero) with there being 2 services in each direction (12*300*4=14,400) those costs would be across 14,400 services.
Therfore each service would need to generate around £280 in tickets.
Assuming an average return cost between the local stations of £4 (£2 per person on each service) you'd need 140 people.
However, as context ticket prices going a single stops during the off peak are slightly more than this; Camberley Frimley (£4.50) Fleet Winchfield (£5.50)
That's potentially quite a challenge (although remember that's not necessarily everyone on the train at the same time) and assuming only very local rail travel.
To illustrate the point, if the extra travel was on returns averaging £7 (3 stations) that's 80 people per train
However, the assumptions are conservative by also assuming most of the less busy services are not generating any income, that the only new tickets are for the next station (or maybe 3) and it doesn't factor in for increased longer distance rail use.
Whilst it would be unreasonable to allow the local trains to benefit from the total income of the uplift in passenger numbers for longer distance travel, however some of it could be allocated without harming the financial situation for the rest of the network (for example, an allocation of £2 for each ticket sold with one of the new stations as a starting point or where such a journey was unlikely using another route; for example Winchester to Bagshot - either due to journey time or due to blocking of using routes, for example £30 any permitted route but £23 not via Woking/Ash Vale).
