The different stages of the survey ensure that the process meets the requirements for statistical production:

• stratified sampling
• reminders and controls of data being collected
• data cleaning
• correction of overall non-response
• calibration
• correction of partial non-response
• the questionnaire was tested in 2013 upstream of the collection, with 97 establishments.

The survey shows a declining response rate: 80.4% (88.8% in the previous edition) due to the additional sample surveyed for 2015. The latter implied a higher sampling rate in small groups, where the response rate is always lower.

Before any exploitation, it is also useful to assess the quality of the variables through their partial non-response rate and confidence intervals provided in the survey quality indicators.

Thus, for certain variables the partial non-response rates are around 5 or 6%, with a maximum of 6.4% (ER_QAUP - other so-called first family gases, and KR_QAUP - commercial heavy fuel oil).

The variable ER_QAUP also contains 47.9% of modified values and should therefore be analyzed with caution, especially since it is one of the energies that concern few units.

In terms of accuracy, the most used energies have the following coefficients of variation:
BR_QAUP - quantity purchased in electricity : 0.7%.
DR_QAUP - quantity bought in natural gas : 1.0%
JR_QAUP - quantity purchased in butane/propane : 5.5%.
LR_QAUP - quantity purchased in domestic fuel oil : 2.0%

Accuracy degrades when one focuses on finer dissemination domains. In particular, the following groups of staff have a coefficient of variation exceeding 7% and should be analysed with caution:

(for JR_QAUP)
Less than 50 employees
500 or more employees