I have seen many people do this wrong.
Printing 1 calibration cube is not enough

This test is the correct way to calibrate 3d printer flow rate.

The instructions for how to run this test are in the excel document

Here is a summary of what you will do:

1. print the cube using both your normal settings and again in vase mode ('spiralized outer contour' in Cura)
2. measure the 4 sides of both cubes and input the data into the excel file along with the flow rate, and your acceptable error
3. iterate this with the new flow rate until the test passes in the excel spreadsheet, then you will have the optimal flow rate

I have entered the equations that I derived and tested into the excel spreadsheet, however here is a summary of how I got them:
When printing there is the average thickness of the flow line and there is also some extra squish on the outermost wall than extends out beyond the average wall thickness (you see this as layer lines on the print). From this I derived 2 equations 1 for a cube with 1 wall line and another for a cube with multiple wall lines. so, with the 2 cubes there is a system of 2 equations and 2 variables (outer wall squish and flow thickness). From this the optimal flow rate can be calculated.

The problem with the wrong method that many people use is that they print 1 cube and assume that the outer wall squish is 0, if this were true there would be no layer lines.