Don't forget the four ground rules; they are implicit throughout this brief example. Don't confuse "preference" with "need" when you work your own system integration question through the "rule-book" either.
We will put our COBOL/SCRIPSIT/Profile through the seven point rule-book.
1) Seek Direction. The problem is that, our six or seven salesmen seem to be good at getting their foot in the door, so to speak, and they are very good with the big customers and problem accounts and so on, but we seem to be missing the undoubted opportunities we should have, in getting our middle of the road customer up to be a big customer or preventing them from becoming problem customers. We ought to be able to provide our salesmen with the right type of information, soon enough, and conveniently enough, to increase our sales substantially.
2) Prepare Study. Our salesmen are spread over two or three states, and they do some fine work with Mod II Profile to help them keep track of qualifying prospects. They seem not to have the time to update Profile from the Accounts Receivable (Mod II + HD) reports that we give them every so often at the sales meetings here at Head Office. It's a miracle if they track any more than their ten best and ten worst accounts. Their secretaries in the local offices do a good job with SCRIPSIT/Profile merge for these cases, and for prospective customers, but they are so busy on the telephone that they don't have time to type in the sales details for a good 90% of their customers. Not until we get our sales up substantially can we afford much in the way of more staff, but a few thousand dollars isn't going to hurt us, if we can see it'll get us somewhere. See the problem? Nobody has the time to get the information to the place where it will be of most use, because, as we're setup, nobody has the time to type all that stuff in again.
3) Analyze Findings. The fact is that we have a Head Office Order Entry/Accounts Receivable database, and it has no pre-defined means of interfacing with the sales offices Profile databases. In any case, they are at different locations. We check a few more facts here and there, and yes, it should be feasible to get the relevant data from here to there without adding any more data entry clerks. Watch out for one thing though. As we grow, we may need to get a little more organized here at Head Office, which is the only place we do the telephone Order Entry and Accounts Receivable. It should be OK if we watch it carefully.
4) Formulate Ideas. At various meetings, with various people, the following ideas are thrown into the arena:
Get the local offices a Telephone link to hook into Head
Get the Sales Analysis program.
Get AR/OE source code and write (or have written) some custom programs.
Upgrade to TRS-Xenix multi-user operating system.
Get each office to run its own AR, etc.
Get a mini-computer that will allow on line use of ten or so terminals.
Modify the sales bonus scheme.
Hold regular sales meetings.
Get the Profile Plus program—more features.
Get the needed information to the salesmen on disk and look at ways to dump that data to the salesmen's databases.
No violent objections to this last suggestion, even from our Accountant.
5) Finalize Ideas. The telephone link, the mini computer, and everyone on Xenix are presently discarded as options because of the cost/benefit ratio, and some of the other options get dropped for other reasons, such as "throwing away the baby with the bath water," or we will lose a great deal of our present efficiency in the process of change.
6) Formulate Solution. The system outline we end up with?
With the help of the AR source code and the COBOL compiler, a custom program will go into AR's files, select the information required, and spin it off to disk files (say one for each salesman).
FCOPY from HD to one floppy for each salesman. Hand out the floppies at sales meetings each 1st and 3rd Friday of the month.
Write a BASIC program to be run by each salesman on his own machine.
Get hard copy with "warning flags" and dump data into Profile at the same time.
Profile can search for flags and spin off data for SCRIPSIT/ Profile merge.
Personalized letters get written to the appropriate customers, and a pretty well up to date database is available for everyone.
It all looks good! We seem to have a solution which meets the constraints.
7) Verification. We believe we have solved the problem, and fulfilled the aims of the exercise. Is it the only solution to the problem? Not quite. The Xenix alternate is very attractive still, so we decide to go that way when we have the secure customer base and increased sales that we want, making as certain as we can that everything we do now will be possible under Xenix. A two stage implementation plan in other words.
Resources? The Software is now on order and having delegated some of my responsibilities, I have the time (and fortunately the expertise) to get the programming done. Our machine at the head office is not very often used before 11:00 in the morning anyway; disk space is there, so we are away!
Was this article helpful?