End-user innovation: The costs
Then there's the cost. Tools aren't free. When end-users get stuck or find themselves in trouble, they call for support, and that isn't free either. Budgets are tight and getting tighter, and in Maslow's hierarchy of needs, end-user tools aren't needs; they are, depending on your perspective, either wants or desires.
Unlike the risk argument, which is fairly accurate though one-sided, a lot of the cost argument against end-user innovation is utter nonsense. Tools first: Some are free, and a lot more are cheap enough that the out-of-pocket cost isn't worth mentioning.
As for the Maslow argument, Maslow would smack anyone making it upside the back of their heads. Corporations don't have wants and desires. They have nondiscretionary expenses -- needs -- and discretionary expenses. Classifying the latter as either wants or desires ignores the fundamental nature of business -- namely, that expenditures are investments.
The needs/wants/desires formulation says that given a proposal for an unnecessary expenditure (a "desire") that yields a risk-free 387.4 percent return on investment, businesses should reject it for not being a "need." If you work in a company that takes this position, you might want to consider moving to a different company that has more interest in making something I like to call "profit."
That leaves the cost of end-user support, which is real, increasing in proportion to (1) the number of end-users; (2) the richness of the toolkit you provide; and (3) the amount of freedom you allow.
It also might increase in proportion to the level of end-user sophistication, but it might not. Sophisticated end-users won't need your help figuring out how to bold-face text or indent a paragraph. On the other hand, when sophisticated end-users get into trouble, figuring out what went wrong and how to fix it is a whole lot more time-consuming and expensive.
End-user innovation: The return
Risk and cost are easy to justify when there's a quantifiable financial return. That's where the conversation about end-user innovation becomes awkward -- because there isn't one. At least, there isn't anything you can predict in a measureable way.
Time for a brief history lesson: Way back when, after the ancient ferns turned into coal but before the advent of competing private telephone companies, AT&T operated Bell Labs, also known as The Place All Scientists Wanted to Work. At one time, it had as many as 30,000 employees.