Testing methods for freeze–thaw scaling of concrete in presence of de-icing salts are developed in the form of normative documents to assess the frost resistance of concrete, aiming to unify the testing standards. These tests consist of exposing cured concrete specimens to several freeze–thaw cycles inside a climatic chamber. The amplitude of each cycle may reach 40 °C with considerable freezing and thawing rates that may reach 10 °C/h, depending on the temporal cycle length (24 h vs. 12 h). When comparing different testing methods, it appears that the imposed experimental conditions are significantly different and may lead to conflicting results on a same concrete mix. Moreover, there is a lack of agreement between the test methods and the real environmental conditions. Yet, the assessment of the representativeness of a testing method for freeze–thaw durability should start with the measurement of the concrete response to its natural environment and comparison to its response under laboratory conditions. Therefore, a first experimental campaign was held in which temperature measurements and scaling monitoring of concrete blocks were carried out at Mont Aigoual in France. This paper is reporting the conditions of measurements, the meteorological analysis as well as the scaling results of concrete specimens exposed in real freeze–thaw conditions with a comparison to other specimens, with same concrete mixture, tested in laboratory conditions. All scaling concrete specimens were prepared and cured together; they had the same age before the beginning of the exposure to freeze–thaw cycles. We show that the applied thermal cycle of [+ 20/− 20 °C] does not correspond to real conditions, even in regions with a harsh winter environment. Furthermore, a considerable dispersion was found in laboratory thermal conditions when applying the normative cycle. Therefore, a reflection about the relevance of the applied temperature cycle and its representativeness of field conditions is proposed.