Rather than trying to solve real business need, even though they think that's what they're doing, many business end up complicating their organization through a mish-mash of mix-matched technologies that aren't really solving a real business need.
I remember when I learned this lesson: that not everything is worth solving through new software development (writing code). The company I worked for at the time has SLA agreements with their customers for hosted-software uptime and availability. To compliment this, they produced PDF reports on a monthly basis showing the particular clients availability, or the lack there of, for the previous month. It was a very manual process, to say the least. Data had to be aggregated from many different sources, and manually assembled into Excel spreadsheets which would be used to populate the PDF files.
Being the young and idealistic programmer that I was, I offered to automate the process for them. Luckily, I was shot down rather quickly, albeit very nicely I must say. It was explained to me that even with all the aggregation that took place, the man hours per month to aggregate and compile the data added up to about 0.5. I had estimated it would take me at least a couple days to a week to complete the project. Therefore, even at 2 days of development time, it would have taken them at least 16 months to begin to realize a return. There were far more important things to work on at the time.
That was definitely an eye-opening experience for me. I realized that it's not enough to write great software for the sake of writing great software. There has to be a demonstrable return on the investment. To proceed with my proposed project would have complicated, rather than complimented the business.
I was recently party to this same line of thinking, though this time from the other side of the fence. A product-based company I was at had some rather old and mundane tasks that had to be performed as part of their business. Folks would complain, and suggest the processes be automated. But the same realization held true. The time and effort required to automate those efforts would not help any more product get on the shelves. Further, while it might save in administrative costs, those costs were very hard to prove or quantify.
The last example I will share, is the business user who happens to read, see, or hear about the latest and greatest technology to ever hit the universe and send an ultimatum down the pipeline that this is what everyone should focus on. I think I've summed it up above, but regardless of the emotional response, there needs to be a real, quantifiable business case for adopting and new technology, or developing any new code. Too many businesses are adopting technology for technologies sake, without appreciating any real business value.
I ultimately fall back to my realization that a proper process is of the utmost importance in upholding an organizations best interest. Through proper use case modeling and risk analysis, most of these types of endeavors can be nipped in the bud before they turn compliment into complication.