When I take my own measurements, I take five measurements back to back without moving and see bigger swings that you did. I've seen that using three different home devices. I've seen it in clinical settings being measured by many, many professionals. I remember once seeing a thirty point swing between two measurements taken immediately back to back by a nurse. The sheer amount of downright superstitious behavior I've seen in nurses surrounding blood pressure measurements is pretty incredible, too: lots of retaking measurements until the number is what they expect, repositioning, or trying something random like you did. I find it all very reminiscent of the 2 4 6 Puzzle -- a lot more confirmation and explanation than investigation.
I'm not trying to say this sort of thing is statistically invalid at a population level. On the contrary. I'm pretty convinced in my own case that breathing exercises are worth about ten points. But I strongly recommend developing a healthy respect for the sheer variability of the number before drawing strong conclusions from a low number of measurements.