I wrote a python script to do this for him, wrapped a gui around it, and compiled it into an exe (this is how I taught myself python by the way). We sold it for $200 a piece to a couple of the departments at our school. When we went to the oracle rep to talk to her about automatically integrating this into the system's workflow, she told us that this feature already existed. Maybe it did, but apparently nobody knew about it or how to use it. We probably could have pursued it further, but then I graduated and moved across the country.
One of Joel's better articles for those that haven't read it explains why: http://www.joelonsoftware.com/articles/CamelsandRubberDuckie...
I think the point here is that that is rarely the case. Mainly because the "customer" that buys is disconnected so much from the "customer" that will use the software that both the pain and the ability of the application to solve said pain are woefully misunderstood.
If I make an investment, say in a mutual fund, I am "using" my money even if the mutual fund is "using" my money.
The "use" in each case is different, but the value needs to align with the buyer.
If a child wants candy, it's up to the candy manufacturer to realize that they have to sell "healthy" candy, because the buyer is the parent. But the parent "uses" the candy in giving the child a treat, making them happy, and not hurting their health. That has to come across in the marketing.
I have been on both sides of the enterprise software sale many times and have concluded that a) it always sucks and b) it's rarely in anyone's best interest.
So instead of examining the current model and making suggestions for accomodating or improving it, I prefer to suggest an alternative.
I believe the best way to crack the enterprise software market is the same way to eat an elephant: one bite at a time through the soft underbelly...
Find a critical business function being done in Excel and provide an alternative web app.
Find a "business within a business" and automate it with modern technology. (Examples are small independent business units, warehouses, job shops, sample shops, anything a user has set up that can be autonomous.)
Provide a modern satelite system to augment and integrate with an existing enterprise monster. (A separate module for one function like payroll or fixed assets, special processes for marketing, engineering, manufacturing, etc.) The possibilities are endless. Somebody is not getting what they need out of SAP, Oracle, or whatever.
Provide a separate business unit with everything they need. This may be cheaper than the customer adding more licenses to their ERP system.
The key to this approach is staying under corporate IT's radar. The way to do that is by keeping your prices below your customer's boss's threshold.
How do I know this can work? Because it has, many times. I have implemented dozens of apps in enterprises that they thought they could never have because of the existing software and sales model.
And I remember history. At one time, IT departments were very threatened by PC's. They challenged their ivory tower with a mainframe and dumb terminals. So users just bought their own PCs from their expense budgets and forced IT's hand.
Lightning can strike twice. Users are once again tired of waiting 18 months for a fix and are ripe for a custom 37signals type of solution. Let the app rush begin.