Getting accurate measurements with IR thermometers can be a lot more
complicated than people think, for two reasons.

The first is simply a matter of spot size and placement. The gun will
display the average temperature of the entire spot, and on the cheapest
ones the spot is about two inches across at eight inches distance. Fancier
ones may go two inches in twenty or twenty four inch distance. And of
course at short distances the laser pointer is offset from the measuring
area. A combination of those factors accounts for many low readings.

The second reason -- an excellent example is my aluminum dryer duct. With
the dryer running the IR gun shows ~86F; but the spot of gray paint two
inches away that I sprayed on the duct shows ~127F, which is the same as I
get using a surface temperature probe. But if I adjust the the gun so the
bare aluminum reads correctly, then the paint spot shows around 275F.

The reason is a property of materials called emissivity, which is to say
how efficiently they radiate infrared. Most materials have high emissivity
ranging from 0.9 to 0.98; and cheap IR thermometers are set to a fixed
emissivity value of 0.95 (fancier ones like mine allow you to adjust the
emissivity correction).

However, metals, particularly shiny metals, have very low emissivity, often
below 0.1. This causes the measured temperature (hot or cold) to be closer
to room temperature than is accurate, with the difference increasing as
temps get farther from room temperature. In order to get my dryer duct to
read correctly I have to set the emissivity on the gun to 0.22.

Also, emissivity varies with surface finish and with the temperature being
measured. If you want exact results for a given material and condition, you
really have to measure the actual temperature with a surface probe, and
adjust the gun to match. Emissivity tables like the one below give a
starting point which is usually close enough for non-critical work -- but
the lower the material's emissivity, the more difference a few hundredths
makes. 0.95 to 0.90 isn't a huge difference; but 0.10 to 0.05 is.

http://www-eng.lbl.gov/~dw/projects/DW4229_LHC_detector_analysis/calculations/emissivity2.pdf