Tuesday, August 31, 2010

Change of Plans

Howdy all!

I've been working on the new report system, and just now realized that the custom sorting is not working right. The 1st custom sort works, but LRO isn't picking up the 2nd custom sort. In addition, I'm only about 30% done with it. Therefore I have decided that the new report system can wait until 1.3.2. I want to do some more thinking and make sure I'm designing the new system in the best way possible. Also, there are too many other good changes (mainly with regards to the S-28 form) that I really want to get out there soon.

One of the new features will help detect erroneous S-28 links. For example, S-14 item 5334 clearly should be linked to S-28 field 5334. Since there's a direct relation, I can detect (and warn you) when this is not the case. LRO will also detect missing links now too. If you have item number 5306 in your inventory, but don't have any links at all on the S-28 builder, LRO will warn you. Because without the links, this inventory item would not be counted! And finally, there is a new report that will show you all of your current links. With all these improvements, LRO is really going to help you make sure your S-28 form is correct! :-)

Remember back then when I said I would consider Inventory Thresholds? Where you could set a minimum # of an item, and LRO would warn you when you get that low? Well, I've determined that this is not the best way to go. It would require much maintenance, and just add to the bells and whistles that are already going off for other things. So...I've decided the best course of action would be to create a report that uses an algorithm to predict what you need to order based on your previous movements of literature. No maintenance on your part! Just keep good inventory records, and LRO will try to help you out with suggestions on what to order. How cool is that??? ^_^

Here's what I have in mind:

You will get to choose between 3 "sensitivities": Low, Normal, or High. Based on your sensitivity, an algorithm involving the Average movement of literature and the Standard Deviation will be used. If the number of stock of an item is LOWER THAN the calculated number, than you will see LRO suggests you order more of an item. Here's the algorithms:

  • Low: The threshold will be the Average movement per month.
  • Normal: The threshold will be the Average movement multiplied by 2.
  • High: The threshold will be the Average movement multiplied by 3.

All three sensitivities take into account Standard Deviation.  If there is a mass movement of literature, the calculated average is reduced proportionally so you don't get a huge threshold.  The Standard Deviation must reach a certain threshold for itself before the Average is reduced.  The higher the sensitivity, the higher the threshold for Standard Deviation.  And so the Average will be more likely to be affected by mass movements.

Let's look at a simple example. Let's say over the course of 4 months, you move the following amount of Knowledge books: 10, 10, 50, 10. Based on these numbers, the Average would be "20" and the Standard Deviation would be "20". So the thresholds would be:

  • Low: 20
  • Normal: 40
  • High: 60

If your current quantity of an item is below that threshold, then that item will show up on LRO's report. LRO will also suggest a quantity to order, while showing you your current quantity.

Other notes on this new feature:

  • When calculating averages and standard deviations, decimals are always rounded up.
  • If an item had a negative movement for a month, that month is calculated as "0". This might happen if someone returned an item to stock, and you didn't mark it down as "received".
  • If an item is marked as "inactive", LRO will not make suggestions on it.
  • If an item is part of an S-28 field with an asterisk in the name (*), then LRO will not make suggestions for it. Since these items (with the asterisk) should be special request items only.

Of course, you shouldn't 100% rely on LRO to tell you what you need to re-stock, but hopefully this will help you catch something you missed (or didn't think of). You might have a "good" stock of a particular item, but based on previous movement, it might be a good idea to stock up again. And that is what this will hopefully catch. As we test this new feature (and I get some feedback), the algorithms may be adjusted. But I think this will do fine.

So...version 1.3.1 is scheduled for this weekend, provided all goes well of course. :-)

No comments:

Post a Comment