Why can't we get an accurate volume projection for our sort?

[wilberforce15]

I'm going to assume this isn't a problem that only my hub has.

Every night, volume is forecasted for our shift - say 70k as an example. Our forecast is regularly off by 10-20% in either direction. We may end up with something in the high 50s or low 80s. There doesn't appear to be any rhyme or reason for it being off one way or the other. It's about right half the time, and not even in the ballpark the other half. This means we don't even really know what our staffing needs are, and it makes every sort a big guessing game.

It's the 21st century. How can we not know how many packages are at our hub, or are coming to our hub, by start time?

We should know a piece count within hundreds, at most. Instead, we are often off by a dozen 100% trailers. Shoot, even if we didn't have scanning technology, couldn't we at least know how many drivers are coming and the percentages of their loads? Even using that, with nothing being scanned, would give us a better forecast.

I'm befuddled, and would appreciate some input.

 

[Catatonic]

I'm going to assume this isn't a problem that only my hub has.

Every night, volume is forecasted for our shift - say 70k as an example. Our forecast is regularly off by 10-20% in either direction. We may end up with something in the high 50s or low 80s. There doesn't appear to be any rhyme or reason for it being off one way or the other. It's about right half the time, and not even in the ballpark the other half. This means we don't even really know what our staffing needs are, and it makes every sort a big guessing game.

It's the 21st century. How can we not know how many packages are at our hub, or are coming to our hub, by start time?

We should know a piece count within hundreds, at most. Instead, we are often off by a dozen 100% trailers. Shoot, even if we didn't have scanning technology, couldn't we at least know how many drivers are coming and the percentages of their loads? Even using that, with nothing being scanned, would give us a better forecast.

I'm befuddled, and would appreciate some input.

Your observations make no sense unless you are a secondary hub where volume is diverted to another close hub with the same service levels.
​Are you a hub sup?

 

[UPSGUY72]

I'm going to assume this isn't a problem that only my hub has.

Every night, volume is forecasted for our shift - say 70k as an example. Our forecast is regularly off by 10-20% in either direction. We may end up with something in the high 50s or low 80s. There doesn't appear to be any rhyme or reason for it being off one way or the other. It's about right half the time, and not even in the ballpark the other half. This means we don't even really know what our staffing needs are, and it makes every sort a big guessing game.

It's the 21st century. How can we not know how many packages are at our hub, or are coming to our hub, by start time?

We should know a piece count within hundreds, at most. Instead, we are often off by a dozen 100% trailers. Shoot, even if we didn't have scanning technology, couldn't we at least know how many drivers are coming and the percentages of their loads? Even using that, with nothing being scanned, would give us a better forecast.

I'm befuddled, and would appreciate some input.

What are you smoking ??? " we are often off by a dozen 100% trailers.."

 

[PiedmontSteward]

I'm going to assume this isn't a problem that only my hub has.

Every night, volume is forecasted for our shift - say 70k as an example. Our forecast is regularly off by 10-20% in either direction. We may end up with something in the high 50s or low 80s. There doesn't appear to be any rhyme or reason for it being off one way or the other. It's about right half the time, and not even in the ballpark the other half. This means we don't even really know what our staffing needs are, and it makes every sort a big guessing game.

It's the 21st century. How can we not know how many packages are at our hub, or are coming to our hub, by start time?

We should know a piece count within hundreds, at most. Instead, we are often off by a dozen 100% trailers. Shoot, even if we didn't have scanning technology, couldn't we at least know how many drivers are coming and the percentages of their loads? Even using that, with nothing being scanned, would give us a better forecast.

I'm befuddled, and would appreciate some input.

Late trains, trailers arriving early/late, etc.

On my sort, our downtime is dictated by the time the last package car driver gets in from his route. If the "regular" loads have been processed by that time, then they start processing volume for the other sorts instead of having 60+ unloaders/sorters standing around doing nothing.

It doesn't bother me because they (management) usually screws up and that's how I get my hours.

 

[Floridacargocat]

"Why are we off by +/- 10 - 20% on the number of incoming units".
That was one of my early observations and question a few years ago, when I was hired, and unfortunately it has not changed. As it was explained to me, the concept is based on historical data and forward projections. It is not based on actual information, accumulated for product in transit. For Next Day Air, I do understand that. But for shipments, which are in transit for 3 - 5 days, management and Information System (for one reason or the other) is stuck with this concept. UPS is so great in accumulating numbers on each and everything that is measurable, but aspects such as number of unit of works, weight, volume per piece, I do not see any inkling that they would go into a short-term planning system. All these data are made available when it comes to the Air side (when you see all the scanners at the Worldport in Louisville).
Misdeclarations by customers on weight and size of box are not uncommon; they will be picked up by the Revenue Recovery guys and girls, when they make their rounds or are notfied by personnel on the ground.
Effects of a variance (especially when it is + 5% or higher) can be predicted very easily. Just add one team in Unload, push it through the system (come hell or high water), and make the numbers. That is what counts. Quality and safety are secondary, but are made primary when something happens (damages, incidents, accidents, leakers etc).
As profitable as UPS is, in the end these "minor" things like variance in numbers of UPW coming into a distribution facility, thet state it arrives and leaves, and the effort to make it appear halfway acceptable look like just as being the cost of doing business.
Quality??? What is quality? It is the sum of all the characteristics that a customers sees in a product/service that he/she/it wants to pay for (and not only once).

 

[wilberforce15]

Your observations make no sense unless you are a secondary hub where volume is diverted to another close hub with the same service levels.

​Are you a hub sup?

I was a hub sup for a few years through college, then left the company for a great full-time job, and then had to scrounge for work after the financial crisis a few years later. So I started my own business and things were/are great except for benefits. So I came back to UPS and have been an hourly for 4 years, while running my business. With a family now, the benefits mean more than the pay, and I like the lower stress as an hourly, so I'm not going to be wearing the collar again.

These issues were there when I was a sup, and they're still there.

Your observations make no sense unless you are a secondary hub where volume is diverted to another close hub with the same service levels.
​Are you a hub sup?

We do divert some volume on occasion to another hub that is an hour away, but that's rare. And when we do, we know about it. It wouldn't explain a surprise an hour before down-time about how far ahead/behind we are.

Floridacargocat is right on the money about how they get projections. It's based on historical data and forward projections. But that's exactly what I don't understand - we have the data to make better predictions, to the point that it's almost 100% accurate at start time. But we don't do that, and get surprised after we've already determined staffing. Or, sometimes, getting a shock when we look at the numbers after the shift is already over.

 

[Anonymous 10]

I'm going to assume this isn't a problem that only my hub has.

Every night, volume is forecasted for our shift - say 70k as an example. Our forecast is regularly off by 10-20% in either direction. We may end up with something in the high 50s or low 80s. There doesn't appear to be any rhyme or reason for it being off one way or the other. It's about right half the time, and not even in the ballpark the other half. This means we don't even really know what our staffing needs are, and it makes every sort a big guessing game.

It's the 21st century. How can we not know how many packages are at our hub, or are coming to our hub, by start time?

We should know a piece count within hundreds, at most. Instead, we are often off by a dozen 100% trailers. Shoot, even if we didn't have scanning technology, couldn't we at least know how many drivers are coming and the percentages of their loads? Even using that, with nothing being scanned, would give us a better forecast.

I'm befuddled, and would appreciate some input.

Who cares just go with the flow it doesn't really matter anyway.

 

[curiousbrain]

Yeah, it's historical, but it's also fed from other sources (inbounded loads? not sure).

As an example, when I first walk in, the volume friend&A will be N; wait twenty minutes, and N will jump by some arbitrary metric - so, causality kind of determines that it is not all just historical, there is some "active" (or foreground, if you prefer) number crunching going on.

A few things I've noticed.

Watch the percent-to-match on each days sort; if it is 80 percent or something "not close" to 100%, expect to get really whacked the next day, regardless of what some tie-wearing in IE says.

Also, the feeder guys and hub sups lie constantly on the trailer percentages - for different reasons I'm sure; honestly, the reason is immaterial from my perspective, as the functional reality is the same. From what I've heard, if a feeder guy pulls low percentage trailers enough, the route will be axed, combo'd, or something I don't fully understand; but, basically, some of them fudge the percentages as a sort of "job security". I'm sure it's not an advisable practice, as if they ever got caught and the company disliked them, it'd be a slam dunk termination.

Another thing is that not every shipper in the known world does a proper EOD with Worldship (or whatever it is called), so the data is not always correctly transmitted to UPS and it's subordinate systems.

I've been around the building I work in long enough to have a feel for the volume on specific days, and the trailer count/percentages (even if they are wrong in the absolute sense, I know what a "65% long box on Monday" really is versus a "65% longbox on Tuesday"), that I only use DMS/PKG as a guideline; I have an internal barometer that says volume will be "this" - unless things are totally crazy, I'm usually within a hundred or two packages.

Granted, I don't deal with 70k a day, more like 10 - 13k during the normal year; so, good luck having a barometer in a major hub I would imagine.

 

[cosmo1]

.......regardless of what some tie-wearing in IE says.

Regardless of what anybody else says, I like your perspective!

 

[Speedy Cerviche]

They could get the numbers accurate, but it would cost more money than it would save. Therefore, you're not getting accurate numbers. You can accept it or quit.

Change isn't an option.

 

[LongTimeComing]

brownbaggin, you are more correct than I think you realize.

He summed it up about as good as it gets. A prediction is a prediction, and should be taken as only a enthusiastic suggestion, rather than a end-all-be-all standard for planning your sort. There is no 'issue' here unless you make it more than what it is. Sometimes the crap IE spits out is crap...and they know it. These predictions also have to be done well in advance of every shipper knowing exactly what they will have. They do get adjusted as information flows in, but it's still a general guideline. And considering how long we have been around, our 'historical' data is pretty solid.

Basically....it's not a big deal. If your hub is failing to perform because of these volume predictions, there are way more underlying problems that need to be addressed than some stupid IE reports.

 

[wilberforce15]

They could get the numbers accurate, but it would cost more money than it would save. Therefore, you're not getting accurate numbers. You can accept it or quit.

I don't wear a collar anymore, so it's more of an intellectual curiosity to me than anything. I've been getting 25 hours a week on my sort because of this, and I know good and well that's not going to change. In that sense, it's fine by me.

And no, it's not cheaper to have bad numbers. And no, the 'historical data' are not solid.

The bad forecast costs thousands and thousands of dollars every time it's wrong on the low end. Why? Because we break packages by forcing things through without enough people. And we didn't have enough people because we sent them home based on a bad forecast.

UPS pays for that bad forecast to the low side with sup-working grievances, injuries, busted package claims, and lost customers. Then the next shift (preload, in my case) starts late (then all those bad things happen again on preload) and drivers are out late.

​It is NEVER cheaper to do business the wrong way. And breaking bodies and stuff is the wrong way. It doesn't even matter what it would cost to get the right system in place. There is no possible way for that to exceed the long-term problems that this stuff causes. That, and I don't buy for a second that we couldn't get very, very good data for almost no money anyway. Even if I'm wrong on that, it doesn't matter.

UPS, as a privately-held company that valued the long-run, might have figured this out a few decades ago. Now, we certainly won't.

This same old thing could be repeated on plenty of other issues - about half our outbound doors have power extendos. Half don't. Even if a power extendo would cost $1mil over its life, it is a slam-dunk case that it would save oodles more than that in productivity, injuries, and damages over its life, especially in a hub that does our volume on three shifts. The only reason I can imagine for why UPS doesn't immediately install them in all our doors is that huge cash layouts look bad on quarterly reports.

 

[TxRoadDawg]

Personally I always thought IE just stuck their finger somewhere and guess a number by the smell. pkg says 9000 stops, dms sits at 9200, them dms pulls epd at 9500, rinse repeat. you would THINK with all the tracking and scanning we do by the time the trailers roll to the centers we would KNOW how many pieces are on them....