The picture below (not great, but hopefully visible) shows the difference in temperature between a probe taped to the outside of the glass carboy (insulated with silver, reflective bubble wrap) and a probe in a thermowell in the middle of the fermenting beer.

In this case, I was fermenting a Kolsch at 60F. The "Cool" Ranco controller is the probe in the thermowell (running the freezer). The "Heat" Ranco controller is the probe that was attached to the outside of the carboy (for comparison).
In this case, the freezer was running, to knock the temperature back down to 60F from 61F. In the process of cooling the beer, the first thing to cool is the glass carboy, like a bottle of beer in the fridge. In this picture, there is a 5 degree difference between the two probes.
When the freezer isn't running, the two temperatures are usually very close or the same. When the freezer kicks on though, the "outside" probe reaches the desired temp (60F in this case) about 20 minutes before he internal probe. If I used that to control the ferment, it wouldn't be quite as accurate. The liquid wouldn't be cooled to the desired temp, and could get quite warm over time. I actually had problems with fusels in the past before I moved to the thermowell.
As always, your experience my vary, but I've found the thermowell to be more accurate and to produce better beers, in my set up. If you use a refridgerator to control temps (as Jamil does), I don't think it would be as much of an issue since the liquid would be cooled slower. If you ferment in buckets or plastic carboys, I also don't think it would be an issue, since plastic is a better insulator. For glass carboys fermenting in freezers though, I would encourage everyone to test for themselves. You might see something similar to what I have experienced.


