The support conundrum, or why a smaller company can have better support than a larger company. If you're an ad supported service, I'd say you're lucky if your company makes $1/year per user. A back of the envelope calculation for a friend's ad supported company puts that number closer to $0.04/year per user, so $1/year may be high.

Providing support is expensive. A support person is going to cost minimum $40k/year in the Bay Area, though maybe cheaper elsewhere. That means, one support person can support 40k people. We'll say that a support person working 40 hours a week, 50 weeks a year works 120k minutes, so they can spend 3 minutes on every user. In practice, only a fraction of users will need support, but they'll need way more then 3 minutes of it.

When a company is small, they'll start out with a small number of support people, say 1-4. They'll also start out with a small number of users and a small number of support questions. Those support people will work closely with the engineers, and problems will be quickly solved. As the number of users grow, they'll quickly out-grow the ability of a small staff, and the support department will grow... but eventually, someone realizes that support doesn't scale.

This is a fallacy, of course, support does scale linearly. What they're actually noticing is that support costs quickly overrun their business model. A single user without a real problem, just unable to understand the product and want severe hand-holding, could take 30 years to make back the support investment. A single real or perceived bug can result in a backlog of support tickets that can take days just to find and close as "known issue".

Or, they're noticing that even with linear scaling, they'll need 10k support personnel to support a million or 10 million users, and they object.

Its at this point, that changes are made to the support organization. Attempts are made to make even lower cost support, switching to email, hiding phone numbers, grading support on how many responses they can do instead of quality, trying to force users to figure out their own problems by reading the FAQ or having forums where they can pow-wow with other users to get the help they want. They'll switch from email support to web forms where they can try and force the user to read the documentation, or at least provide all of the information necessary to debug their problem in one shot, instead of a bunch of "please provide this" replies.

Often lost in this is how to get real bugs reported back to engineers, especially since companies as they grow are less and less likely to want to admit to the possibility of bugs, much less be open about their existence and fixing them.

Eventually, users will start to complain, they'll say there's no support available, they'll grown that they can't get someone on the phone, they don't have a way just to email their problem to someone... and they're right. They don't understand the support conundrum, they think the ads pay for everything.

The odd thing, to me, is that companies which do make real money off of customers, companies selling >$100 boxed software, or charging you $30-$60/month, are also incapable of providing support. That's probably due to another effect, which is the idiots drown out everyone else. They aren't trained to give support to people with real problems, they're giving support to people who think the mouse is a foot switch, or who can't be told that something isn't plugged in, they have to ask them to unplug and plug it back in.