I think the first step would be to buy a second monitor, ideally from a different manufacturer, and verify that your readings are actually correct. My experience is that cheap monitors are basically random.
If your reader raises it's reading as you sit in a small room, and lowers when you open a window... It's probably not at risk of being random. Inaccurate maybe, but still sensing in the correct direction.
I haven't seen anybody take any of the cheap CO2 sensors and demonstrate that they are anywhere in the range of the readings of a lab grade CO2 sensor.
