There is method to his madness...
The whole concept comes from large companies seeing services as the new cash cow. However whether this can be applied to consumer products is questionable.
if you are buying say a jet engine, there is a lot of value in the monitoring of that part so that you can analyse performance over time. You can use this information to advise on maintenance schedules, more efficiently schedule servicing, and better analyse potential failure modes. The idea here is that by using the same techniques used by companies such as google to analyse big data, you can better support your customer, producing a win-win scenario. This is already being done by companies like RR and GE
This makes sense with a big ticket item, but it is questionable about whether it scales. You could argue that it would work with cars, since these are big purchases and require constant maintenance. So you could argue that by constantly analyzing performance via the cloud, the customer can be given better servicing advice and advised on possible upcoming failures, before they actually occur(for example by measuring battery life-cycles you could predict when it was likely to fail) . Another possible usage would be AC units or Central heating boilers. Again it could be used to better schedule maintenance.
However would be of any use for something like a fridge or a TV? Seems unlikely. In these situations the benefits to the customer seem to be outweighed by the benefit accrued by the data collector. Here we have a situation akin to extended warranties, where customers would be made to pay for a service which they accrue very little net benefit.