Back in the early 90's, we had a Mac only product that retailed for about $20. If a customer called in for support, in general, it meant that we made no profit on the product.

When we released a DOS / Windows 3.x version of the product at the same price, our sales shot up 10x the mac version - but the support calls we started getting usually ended up being support for DOS or Windows, not related to our product. So, effectively we were spending twice as much on support.

Of course, since we were selling so many more copies of the software, we still made money.

As a side effect, the support calls for the OS led us to develop quick scripts to get people back on their feet once we recognized the smell of their problem, which meant that they were more likely to buy our other software.

Customer loyalty is an intangible, but valuable asset, and is usually borne from direct customer interaction. When things go wrong, how is it handled?

If google did employ people to "fix the interwebz", I bet they'd find more customers for their paid services.

So far at least, all of Google's for-pay services are relatively niche/targeted at power users. I use Google products EXTREMELY heavily and I've never had occasion to consider a paid service of theirs. I think the overlap between "fix the interwebz" people and potential customers for paid services would be approximately zero.

It is possible however that what you're saying will open the door to charging for less niche services, but that's a whole 'nother can of worms.

At ISP contact centers, calls take 8 minutes on average. 75% of all customers never seek support, so average time spent on support is 2 minutes per customer.