its not so much that they aren't functional, as it is that they are difficult to read in many cases, or difficult to have reliance on the reading that shows up.
testkits like NH3-4, NO2, are very effective in that as long as you get the bottom coloration provided on the color chart, you are in good shape. NO3 however, once you get over 30-40ppm, is very difficult to read and not even worth looking at. For that you would have to go by the fish health to determine if you are adding way too much NO3. I have had up to 80ppm NO3 (as far as I can tell anyway) with no signs of fish stress, so have decided to just not test at all.
Testkits like PO4 and Fe are commonly said to be defective from the start and give false readings. Not worth buying.
pH testkits (other than digital) - although essential - has a .2pH margin of error. Is that effective? Within reason, yes, but not for an exact co2 reading.
KH - assumed to be very effective, or at least I trust mine.
GH - Good to an extent, but as Tom Barr pointed out, do you know the ratios of Ca/Mg within the GH reading you get? It is just common to believe that there is enough of both in readings of GH > 5dH. I don't trust it, so I still dose Mg after each water change.
So, I wasn't necessarily implying that testkits are worthless, just that 100% reliance on testkits is obviously not "effective". I have a testkit for most of the variables (didnt bother trying K or Mg - if there is one). I just don't want to rely on them for my dosing. If I ever see signs of fish stress, I will get them out and it should point me into the correct direction though.
450g Planted Discus Tank - Work in Process!