I can't say precisely for a KIA as I haven't driven enough of them to compare, but taking experience from many miles driving trucks-

The declared mpg in different trucks of the same model can and does vary in accuracy, sometime by a quite large percentage, (that's accuracy of how much diesel has to be put in the tank not the variability of mpg due to load, speed, road type, weather etc)

The mileage and speed are measured by how many times the wheel goes round on the journey. This means that the precise measurement relies on the circumference of the wheel and as the tyres wear out the circumference changes by about 3%, that makes both the mileometer and the speedo 'inaccurate'. For the purpose of comparing onboard and spreadsheet mpg that inaccuracy is irrelevant as the 'mileage' used to calculate both will be the mileage that is declared on the dials. Experience in many vehicles would tend to show that the speedo and mileometer usually read faster/further than reality when timed against known distances or against satnav speeds but that is not always the case, so don't assume that you can have 33 on the speedo in a 30 limit.

Once one has allowed for the inaccuracy of measuring the distance, for speed the other unit of measurement is time, on a modern electonic speedo the time element will be from a digital clock which is generally accurate to within a few seconds a year so not really relevant.

To calculate mpg there need to be some measurement of fuel used and I expect that is done from the cars computer controlling the fuel injection system and a calculation of how much fuel is squirted into each cylinder for each stroke, that has to be done thousands of times each minute in real time and will depend on a flow rate calculated from fuel pressure and injector size multiplied by the small time each injector is open, with those all being very small measurements the inaccuracy of the measurements will come into play much more. When you buy fuel at the garage if you buy 50 litres measured to 0.01 litres then you have an inaccuracy of 0.02%, if you use a spreadsheet for your calculations over 10 refills then it goes down to 0.002%, compare that to the fuel injection measurement/calculation where each shot of the injector is hundredths of a millilitre (less than a single drip from a tap) achieving 0.002% accuracy would require measuring/calculating to millionths of a millilitre, and remember that is done thousands of times a minute in real time as you drive with the mpg then being continually recalculated, it's pretty miraculous that we get an mpg reading at all even if it is a bit inaccurate.

I tend not to be too fussed by the mpg figures of my cars, after all I've bought the car so I'm stuck with it and its mpg, variations in actual mpg are down to me and my heavy right foot, the one figure that I do use is the 'miles to empty' prediction but then I only use it as a rough guide and want to be filling up or know where the nearest petrol station is when it gets down to 50 miles.