The use of the shipment as a transport measurement unit in addition to the conventional units of tonnes or tonne kilometres is an original feature of these surveys. The shipment is defined as a quantity of freight that is made available at a given time to be transported during a single transport operation from a given shipper to a given consignee and is the natural observation unit for reconstructing transport chains. However, this unit also has conceptual advantages. The shipment is the link between the shipper and the consignee. Likewise, a representation of transport that is closely integrated with the production and distribution process provides information about the economic context of the exchange of goods. In addition, it improves the perception of the practices of shippers and the logistical constraints that affect them. Shippers develop their transport policy on the basis of not only the tonnages they produce but also the number and spatial dispersion of their clients and the frequency at which they send out shipments. There is a great difference between sending 1,000 tonnes of freight to a single consignee and 1,000 shipments of 1 tonne each at different times to different consignees. The number of tonnes is the same, but the economic rationale and the type of transport differ in every respect. Likewise, use of the shipment as an observation unit permits observation of the activity of carriers by the number of operations they perform. This provides a picture of the population of carriers that is close to that based on the total turnover of transport firms. The results from these surveys thus allow the use of a variety of measurement units, not just the number of tonnes or tonne kilometres but also the number of shipments or the number of shippers responsible for the flows in question. Passing from one or another of these units is highly instructive and provides an understanding of the market that is markedly different depending on the unit which is preferred: the relative weight of the different sectors of activity in the generation of traffic changes completely, as do the transport modes. The 1988 shipper survey covered 1,742 firms and tracked 5,118 shipments, including 4,893 for which the transport chain could be reconstructed as far as the final consignee, or the French border in the case of international shipments.
The diversity of logistical situations and the dispersion of the statistical adjustment variables for the numbers of tonnes and shipments led to an increase in the size of the sample, and the reconstruction of transport chains was extended to the borders of the former 15-member European Union. The 2004 ECHO survey was thus based on the observation of 2,935 shippers, 10,462 shipments (of which 24% were international), and 9,742 transport chains. Both surveys covered all metropolitan French sites having 10 or more employees in the wholesale trading and industrial sectors, apart from mining and building and construction. It was extended in 2004 to include mail-order companies, farm cooperatives, warehousing services, and industrial waste treatment centers, which represent a total population of approximately 70,000 sites. Four percent of these are included in the survey. The range of activities covered by the survey had important effects on the measured flows in that they excluded almost all large bulk shipments of raw intermediate goods (such as crude oil products, ores and minerals, and sand and gravel) as well as a large proportion of agricultural products. These types of products were covered by the survey only through wholesale and intermediary traders. Nevertheless, a large range of sectors was covered, in particular those undergoing the most rapid change. This scope represented in 2004 approximately 47% of total national transported tonnage (domestic shipments and that part of export transport which takes place within the national territory) and 84% of tonnage other than the large bulk shipments mentioned earlier. Another characteristic of the survey scope relates to the decision to consider all freight shipments with no other limit than a minimum weight of 1 kg. The selection of this low threshold highlighted the large number of small shipments that are an important aspect of the modern economy. However, the choice was responsible for consider able methodological discussion, because the frequency and size of shipments are closely linked, and the decision made in 1988 to include the last three shipments sent out in the case of each site led to samples with a very low proportion of heavy shipments, making statistical adjustment difficult. The attempt was therefore made in the new survey to increase the representation of heavy shipments and, more generally, modes other than the road, for which the road’s dominance made it very difficult to obtain a sufficient number of observations