Generally I make my decisions on what my NEO reads, and also how I feel since I get definite signs when I go into hypoland. I am Hypo Aware, thankfully. The SD is backup / confirmation that the NEO has not misread, a rare occurence indeed.
Now that I am 'fat adapted' on LCHF and have relatively stable bgl ranges, then for me a 'low' reading is not something I panic about so I do not always respond by taking extra carbs. I do take action if it is a 4hrPP result, since this is my bedtime status, and I do not want a hidden hypo while I dream. The other time I may take action is if I need to drive.
Since it is only the Gliclazide that could really give me a problem, and I now understand how it affects me over time, then I can be more relaxed about it than if I was on insulin or a more potent cocktail. Thus I raise the warnings about those meter calibration issues and the ISO since it can be more significant a problem for others, especially where using a single meter only which is Plasma Calibrated: these users may well be unaware that their meter is giving higher result than shown in NICE /DVLA or even the Newbie infopak given out by
@daisy1. If they are not hypo aware then this could be a dangerous situation for them since they would not react to a 6.0 reading
I have on several threads here published graphs I have taken showing that my SD is always higher than my NEO. The two meters track each other quite closely when averaged over time, but can differ wildly on a moment by moment sample. Personally, if I get more than 3 mmol/l difference then i re-test since that indicates a misread, but if both results are both 8 or more then I have to allow greater leeway due to error being percentage of reading