There are two simple truths in IT and in life that should be considered:
As long as something is free, there is no reason to limit or manage my use or consumption of it.
If both of these are true, why is it that so many organizations take something that has such a variable infrastructure cost, like SharePoint, and offer it to employees for free? They consider it a "sunk cost" or simply the price of doing business, and utilize an investment or funded model for the service, in which IT estimates what it's costs are going to be for the next year, generally based on past usage or growth trends, and then let employees consume, consume, consume. While I'm all for SharePoint getting used, I'm not "for" the problems this model creates. Specifically, it means encouraging end users to put as much information into SharePoint as possible without limits.
To be fair, I'm not calling SharePoint "expensive"; SharePoint itself is no more or less expensive that any other enterprise solution, such as email. In fact, SharePoint itself can be thought of as a "sunk" cost, usually factored as "cost per head". An end user typically only needs one license to access any amount of SharePoint services, and in many, many organizations that license is provided as part of a cost-effective bundle that includes the Windows, Exchange, Office, etc., licenses that any given employee may need. It might be factored into the employee cost just as a desk, telephone, PC, or internet might be. (I'm ignoring the complications of assets vs. non-assets, capitalization and depreciation, etc., but run with it. :))
So what's different? Consider email, for example: Each employee will typically have a single mailbox, and that mailbox will likely have a specific cap or quota applied to it. It's relatively easy to manage, and relatively static... increases in quotas or requests for secondary or shared mailboxes are exceptions, not the norm. On the other hand, SharePoint, if used as expected, allows not only for the services to be used but for them to be essentially self-provisioned. For example, within Microsoft, I can request a new SharePoint Site Collection whenever I feel necessary. I can create several at once if I feel so inclined. I can move files and content into my newly provisioned sites with impunity, and I never see a penny or problem.
The point is, that "sunk" cost-per-head model doesn't work. The infrastructure costs don't scale with the employee... it scales with usage, it scales non-linearly, and predictions of future growth or utilization are simply bad guesses. This means the investment model doesn't work; It puts a proactive funding model into a platform that must be managed reactively to end-user demand and use. Sure, you can look at the company initiatives and do your best... but that does nothing for the random executive that tells everyone in their division mid-year to move everything out of their file shares (this is not a good idea, by the way). Suddenly your use triples while your funding remains static and you're stuck.
The issue isn't only about users putting information in to SharePoint; It's as much about how rarely they remove or delete anything from it. So, information goes in, information does come out, but information is very rarely deleted. This comes from the fear that someday they might need it back and from simple laziness... the site is no longer used or is no longer necessary, but no one actually attempts to delete or remove it.
This where a Chargeback based model has value. Using chargebacks changes the mentality such that the business user can make an educated decision: Remove things and pay less, or request an increase in quota (at a higher cost) and pay more for the additional capacity. There may still be some elements of the system that are managed as operational costs and do not incur a specific charge to the business, or the model can be designed to recover all operating costs... but there must be an incentive to users to operating efficiently. SharePoint isn't the same as a basic file share, and the infrastructure and operations are appropriately more expensive.
In my perfect world, here is what I might see:
This basic structure resolves nearly all of the problems many customers experience when managing a SharePoint environment:
Governance plays a significant role here as well, and any charging model should be balanced against the governance plan that the environment is ultimately intended to support. Other factors, such as high availability requirements and disaster recovery requirements should also be considered and factored into the pricing. Having a fully redundant infrastructure is going to cost more. Having an identical disaster recovery plan is going to cost much more. Longer backup retention schedules are going to cost more.
The end result is an environment that does get heavily used, but only contains information that is actually important to the business, because they're paying for it. If it weren't important, they could delete it. This also limits the "moving the problem" scenario, where customers blindly attempt to move entire file shares into SharePoint. If they want to pay the higher operational costs, that's their business decision. They're funding you to cover it, and you'll be happy to do so because you can now afford the additional capacity required.