There are various techniques available to measure human physical activity (PA). Accelerometer based techniques claim to be non-invasive and easy to use. The signal magnitude area (SMA) is the most extended feature used to measure the physical activity. It is calculated by sampling and filtering an accelerometer signal of at least at 50 Hz. SMA has a proven and widely accepted linear relation with the energy expenditure. A novel magnitude called JIM, which is more efficient than SMA, is proposed in this paper. The jerk-based inactivity magnitude (JIM) is also calculated from the acceleration signal, but at a sampling rate of 1Hz, increasing the battery life of the measuring system. This magnitude gives the same information as the SMA (correlation of 95%) and is validated with a group of 39 users in free-living conditions for at least 24 hours.