You know more than me on this subject.
They give before and after power and torque figures, but my guess is they have not run the car on a dyno.
They probably use Vauxhall stats for the 'before' figures and there own 'think of a large number' figures for the 'after' power and torque figures.
Torque is king. The only thing that matters to me is peak torque, how flat the torque curve is from tick-over to red line, and what red line is.
Torque is basically governed by how much air you can get into the cylinder. On a normally aspirated engine, that's down to cylinder head, intake and exhaust efficiency. There isn't much you can 'chip'. On a turbo engine, the same things matter, but the turbo boost also forces more air in. 1 bar(14psi) boost means you've got double the pressure of air in the inlet manifold, so you get (roughly) double the volume of air going in. Your 2L N/A engine can therefore behave like a 4L N/A engine at 1bar boost. Or a 3.6L twin turbo can behave like a 7.2L N/A with 1 bar of boost.
Torque increases more or less linearly with boost. The boost is normally electronically controlled, so you can up the boost by chipping the engine. Torque is usually fairly constant with RPM until you get to the top end of the RPM scale when it tails off due to reducing inlet efficiency (valves not open long enough to get a full charge in).
Power is just (Torque * RPM). So given that torque is constant-ish up till probably 75% of red line, power increases to a peak and then tails off again. The power curve is very rarely anything like a flat line - it'll have a peak in it somewhere just above where torque starts to tail off. Cars with highly tuned inlet manifolds (like the old N/A F1 cars) typically have a very peaky power curve. They produce loads of power when on song, but only over a very narrow RPM range.
So on a turbo car, you basically just up the boost to get more power. And hope the rest of the components can handle the extra pressure, fuel flow, heat generated, stress etc.
What these guys do is anyones guess, but to get more than a few % increase they must be fooling the ECU into allowing more boost. I can see how it could work using the method MarkDTM says, but I'd like to see a proper torque curve before accepting any power increase figures. You can tune a car to produce silly power, but if the result is a narrow torque band then the car becomes undrivable in the real world.