I'm sure I've mentioned before that most of what my department does is create a new form, process the input, put the results in a database, and write reports against the data. We're just automating the paperwork.
The only place performance tends to be an issue is in reporting. There, a bad query can mean the difference between two minutes to run and 2 hours to run. But even then, we "solve" the problem by installing another slave DB server and run all reports against that. It doesn't require interactive response times, and doesn't affect performance of the master DB server.
We're still early on the exponential curve, but we're still able to double our reporting capacity by throwing hardware at it for less than the cost of an entry-level programmer. We could double again twice before paying for a real DBA.
Given that, what is our company's justification for hiring people who are really good, and paying what they're worth? (Note the sound of heads bumping against the glass ceiling of what "we" are willing to pay for programmers.)