Trying to gauge economic performance from electrical use is like trying to read tea leaves. It is a very crude and judgement based approach and different people can make the same data produce polar opposite conclusions based on the assumptions they make.
Basically, you can make the numbers say what you want since there are so many factors to consider and eliminate.
For example, the power consumption measure does not really take into account efficiency savings and consumer habits. Using less power could mean businesses are retiring old, inefficient machines and production methods and switching to new, more efficient and productive machines and techniques that yields more output for less power consumed. Hell, if you upgrade the transmission lines and/or built new power plants closer to where power is used so you don't have to transmit it as far as before, there is less transmission loss as a result, so you can have the same amount of power usage yet be able to cut back on power production significantly.
With electricity consumption, it is also all but impossible to separate commercial and civil usage, so we have little way of telling if an increase in power consumption is because factories are making more products, or because millions of more homes now have air conditioning compared to the comparative year. With the hybrid and electric car industry about to take off, that will further skew any attempt to analyse the raw data.
If you want to look at supply side indicators for economic activity, single purpose raw materials like iron ore, cement and coper would be far better bets, but even there it is hard to factor in efficiency savings and what those raw materials are ultimately used for.
As the old analysis saying goes, trash in trash out. If the raw data you are using has too much 'noise' to 'signal', your results are unlikely to be sound.
Supply side observations are only useful if you can get precise data on exactly what inputs are going in to where as well as factor in as much of the related variables like technology improvements and changing production methods etc. Supply side observations are thus only really good for focused analysis on one or a small group of select factories or locations, which is why intelligence agencies often use it. You can get a good idea of how many tanks a tank factory is making without having to have eyes inside the factory by keeping taps on how much iron ore the factory is using for example, but there is no way to work on that level of detail system-wide to monitor an entire economy.