"
Difficult to lose weight when you don't have a scale
".
This is how Sylvain Duranton, of the Boston Consulting Group (BCG) summarizes in a vivid way the study he conducted, published on Wednesday.
This establishes that less than one in ten companies (9%) estimates that they measure their greenhouse gas emissions precisely, while 85% say they seek to reduce their climate footprint.
The survey was carried out among 1290 companies from nine major sectors, spread over twelve countries.
When it comes to measuring CO2 emissions from the direct activity of a company and its energy consumption, it remains relatively easy.
As soon as we want to take into account the carbon footprint of suppliers, it becomes a puzzle.
"
It's even dizzying, when you think of the tens of thousands of references from a hypermarket, for example,
" comments Sylvain Duranton.
Read alsoThe variable geography of the carbon footprint of the electric vehicle
The difficulty of the exercise is such that the companies questioned by the BCG estimate their margin of error in their measurements between 30% and 40%.
This leaves one wondering when some undertake to reduce their pollution by the same order of magnitude.
The survey also reveals that 86% of companies still fill out Excel tables by hand to report their CO2 emissions in their annual reports.
If the BCG is interested in the question, it is because it is a provider of solutions.
Sylvain Duranton, a sociologist and mathematician by training, heads BCG Gamma, the consulting firm's artificial intelligence and data management unit.
A service created six years ago which already has 1,400 employees around the world, many of them young "data scientists".
A hundred databases
BCG Gamma has acquired a hundred databases providing masses of figures on various sectors. Its know-how consists in crossing these databases wisely. With what result? Sylvain Duranton takes the example of a drink producer. Until now, the company has calculated its average CO2 emissions related to the manufacture of glass bottles, which are very energy-intensive. Artificial intelligence, based on databases, makes it possible to refine these estimates by taking into account the hundreds of different models of bottles, whether produced in France, India or elsewhere. The solution developed by BCG is already used in the cement industry or in mass distribution. It operates by a very fine simulation rather than by direct measurements of the activities,which would also be long and delicate to carry out. While artificial intelligence still arouses a certain mistrust among the general public, one of Sylvain Duranton's challenges with his clients is to convince them that "
a good simulation is better than a bad measurement
”.
