Let say you want to know if people are having less sex (well, it may come out from the search or browsing history but maybe not) and you want to use this information to target advertisements of services or products.
So how would you do that using cloud connected sensors?
Easiest method is to use CO2 sensor in the bedroom (assuming that this is the place where usually people do have sex). Rate of CO2 production is correlated with the rate of energy usage and spike in CO2 levels around certain time may mark people having higher bodily activities (and when located in the bedroom it usually means sex). Change in the frequency of these spikes may mean the change in the frequency of having sex. As ads do not have to be that precise, this is probably enough data.
Another method is to use a temperature sensor. As higher bodily activity would consume more energy, it should then rise the room temperature that could be detected with sensitive enough temperature sensor. This method is more evolved as the spikes in the temperature could be also caused by the rise in outside temperature or the temperature difference might be very small when windows are open, or may be caused by fall of temperature outside that heating system tries to over compensate.
But it is likely that there exists distinguishable change profile and it is likely possible to mitigate outside effects with additional sensors.