You didn’t happen to notice that they were **different **pizzas did you? That’s sort of central to the issue here. Unless you think that trash collection varies by individual taste.
You didn’t happen to also realize that quite of few cities have government control over trash collection did you? Does your neighbourhood have private or government trash collection?
Perhaps you’re confused because everyone on your street orders the same pizza at the same time every week.
As I said before, if you and your neighbour BOTH ordered from Dominos at the same time, would you expect that the delivery driver would take both pizzas at once? Is that how you would do it? Or do you think it would be better for Dominos to send him out, make a delivery, then come back to get the other pizza, then drop that one off?
The question I asked about Fedex was if the driver had two pizzas for you, do you think he should drop them both off while he’s at your house?
You don’t understand why they might leave 70% full? Let me clue you in then.
People don’t put out the same amount of trash in a given week. If a truck has a fixed route, then one of two things is going to happen: either the trash will be lighter than normal, in which case the truck will be returning partially full. Or, the trash may be heavier than normal, in which case the truck will not be able to pick it all up. That means a second truck would have to be dispatched, and it would return mostly empty.
A very common process optimization step for this kind of problem is to figure out how to maximize the capacity utilization for the trucks. But that’s not an easy problem. There are two ways to do it. One is to get a better handle on trash variability. The other is to use dynamic routing. With dynamic routing, trucks keep going until they are full, then another truck pulls in and continues the route. This sounds easy, but it is not. For one thing, what is the second truck doing while the first fills up? Just sitting idle? What happens when a truck is half full but overlaps the route of another?
This is just a tiny example. Routing problems like this are not easy. They can be made damn near impossible if you are hamstrung by union rules, antiquated information systems, trucks without radios and GPS position reporting, and a bureaucratic culture that refuses to consider process improvements that require big infrastructure upgrades.
Oh, so then why did you say government would be at 70% but private would be at 90%?
One thing to note is that they do have a pretty good idea since we sign up for specific volume containers, giving them a sense of maximum load. I pay for the volume whether I use it or not.
but that’s not the point
*All things being equal: *would the problems you describe change when you have 5 companies serving an area vs 1 company serving an area?
It sounds like you’re saying, “it’s complicated so only private companies can do it because they are high tech, government will screw up because it sucks.”
Forget government for a moment and compare 1 large company serving an area with 5 small companies serving the same area.
Now let’s consider how the private industry could deal with this:
They could have a ‘tag a bag’ system where you buy garbage tags per bag, then clip them onto the bag to show you’ve paid. The advantage is that the truck can scan the tag and record how much garbage it picked up from you. This data can then be collected and analyzed to figure out trash patterns per person. They can then send trucks out on specific routes that optimize the fill capacity of the truck.
You can go one better, and request pickup on a web site by entering the number of bags you need picked up. Then they can schedule the right sized truck for your neighborhood because they know exactly how much trash is waiting for them.
If they have a network of trucks picking up from neighborhoods piecemeal, it becomes easier for one truck to request another. Head office can build systems that track the number of bags loaded onto each truck, and do predictive analysis to figure out when each truck is likely to fill. Then they can dynamically reschedule the nearest, emptiest truck to sweep through and pick up the trash the regularly scheduled truck couldn’t take.
They have the option of offering a homeowner a discount for holding their trash for a week if they don’t have a truck available to pick it up.
Because the deal with clients directly, clients notify them to not pick up their trash when they are on vacation, which helps optimize routing.
New startups generally have new computer hardware, the latest scanning hardware and newer, more fuel efficient vehicles, etc. They’re better set up to be able to build optimizations, track data, etc.
New startups aren’t operating under layers of old bureaucracy, union rules, and political considerations that may hamstring a government operation. For example, a government agency may not be able to make optimizations that reduce the number of drivers needed, or use automation to eliminate a second trashman on the truck, because it’s near impossible to fire them anyway so there’s no point.
Even this level if thinking is trivial compared to the work you’d really do if you were trying to re-engineer a trash hauling process. There could be inefficiencies baked into the entire process from dispatching to work scheduling to maintenance.
This really isn’t ideological, you know. It was really my experience with this kind of analysis that made me go, ‘wha?’ when you guys were claiming that it was obvious that the private trash system was less efficient. It may or may not be, but you really have no clue from just your own observations. At best, you have a data point worth investigating.
OK, I won’t take offense, since you put a smiley in there.
I will though note that you have absolutely zero idea about my background, my schooling, or my experience with “even modestly complex system[s]”. As **emacknight **said above, do you really want to measure dicks? I don’t think this is the forum.
On point, and more so since you seem to like the school bus analogy, I’ll reveal the fact that I owned and operated a school bus fleet for a number of years. We serviced all grades preK-12 for more than two dozen schools, morning and evening plus buses for sports and other off campus activities. Oh, and handicapped students, too. We used several different sizes and types of vehicles, and had operations of one kind or another underway pretty much 24/7/365. Plus we maintained our own fleet, and trained all our own drivers. This being a private fleet with a profit motive, the intricacies and the efficiency tradeoffs of routing are more than familiar to me. I lived and breathed routing efficiencies morning, noon and night for almost two decades. So please, cease with the disparaging comments about my familiarity with real world matters of pick-up and delivery.
OK, now we’re quibbling. **emack **and I both seem to agree that we do not know every single intricacy of each of those companies’ operations. But we also agree, with what we maintain is “fair assurance”, that five companies operating over identical geography and providing identical (beyond cosmetic differences) services *cannot *be the most efficient method.
So please stop denigrating what we do know merely because we cannot know everything. You still have to stretch credulity rather far to envision some peculiar set of circumstances under which the situation as **emack **has observed it could indeed actually be the optimum of efficiency.
No, you should relate your conclusion to the model which covers the situation. If you say you have already, then that tells us what we need to know about your training I suppose.
Stop poisoning the well. I kwop we are talking trash, but stilll
You mean the"ideology" where he says your model /data combo is entirely insufficient to draw the conclusions you draw, while also consistently stating that it may be possible to to present such a combo to support your conclusion, and where he takes no position on whether or not you can do it, only that you haven’t yet done it?
Sure, it could work out that way for the reasons already described and many more. Just like City busses run on the same streets, and sometimes they pass each other. sometimes you might see two separate UPS delivery trucks in the same neighborhood or even on the same street. For the same reason you might see two Starbucks across the street from each other, as we do even in our small town. It all depends on what you are modeling and what you are trying to optimize given the resources you identify and how they are constrained.
You act like parallel queues, of which this appears to be a variation, are unheard of. Think about that the next time you go to a supermarket, a gas station, a bank. Are they all crazy to have their queues set up the way they do? Are the inefficient simply because they are parallel?
Me, I’d rather observe the data first, then see what model might be appropriate, and then see how it comes out. But you haven’t presented us with anything other than “there are a bunch of garbage trucks on my street each week, therefore it is inefficient.”. We are simply saying there is a lot missing before you can say “therefore”. Just like your algebra teacher did at one time probably when you first covered proofs.
No need to take this personally, simply fill in the missing blanks, and you might turn out to be correct!
I did NOT say that. I said that you hadn’t made the case that five trucks WERE more efficient than one. To show why you hadn’t proved it, I said that IF five trucks could be filled to 90% and the single truck to 70%, five trucks COULD be more efficient.
I wouldn’t dare to make a claim one way or the other about which one actually IS more efficient, because unlike you I’m not going to make a wild-assed guess based on one piece of anecdotal evidence, based on an observation of a tiny part of a complex system. That would be you doing that. I make no claims of efficiency whatsoever.
And that’s one of the potential optimizations I just pointed out in the message I was writing when you posted yours.
That’s EXACTLY the point. There are potential efficiencies elsewhere which may override the supposed inefficiency of five trucks serving the same neighborhood. And even that may not actually be inefficient. It may just look that way from your window.
All things are not equal. That’s my entire point. You have no idea what drives these decisions. For all you know, your neighborhood is an outlier. Or that you’re just seeing an artifact of a much more complex routing system than you can imagine.
But rather than try to analyze this, there’s a very good proxy for determining if the private trash collection is more efficient - is the overall cost for your trash collection higher or lower?
I never said that either. I do think that there are aspects to government operation that can make it hard to re-engineer, but that has more to do with the fact that these operations have often been around a long time, have not been updated well, have a lot of entrenched special interests, etc. The same thing can happen in private industry, and it’s one of the reasons why big companies often wind up getting their hats handed to them by leaner, newer companies with more drive, ambition, and fewer legacy issues to deal with.
I would not discount the possibility that the inherent competition between five companies would result in them striving for efficiencies that wind up making the overall system more efficient. Or if it truly is not, then what you’ll probably see is adaptation, such as the weakest ones dropping out of the market, or the companies signing agreements with each other to share information or workload data so that a truck from one company could pick up the slack from another to improve both company’s capacity utilization, or something like that.
What I am comfortable with is that these companies will not operate at a loss forever. That drives them to be efficient or to get out of the market. They don’t have the option of just raising their prices when their competitors don’t.
I will say one thing - often when governments outsource to a small group of companies, the result does not improve efficiency because there is still no real market - a government-sponsored monopoly can easily turn into a situation where a company can raise prices without fear of competition, have its risks offloaded onto the government, etc. So it’s not always the case that privatizing government services is a good thing. The critical piece to reform of this type is that the outsourcing is done competitively and the new system operates within a competitive market. That’s often not true.
I said forget government, compare 1 large company vs 5 small companies. How do those issues change?
Are all 5 small companies facing the same issues? Does the large company face the same issues as the small ones?
How do the issues you described hurt the large company more or less than the 5 small companies?
That’s what this is about. Right now 5 small* companies are picking up trash within the same area. Where I consider trash collection a commodity, it’s the same from all companies, either picked up or not. How does that compare to one large company picking up trash within the same area?
*Hard to call Waste Management or Allied Waste a small company.
Doesn’t sound like much benefit of the doubt when you immediately say it is not as efficient as possible withtout everdefining the parameters of the model, let alone what “efficient” mans in the context.
We are not in your head. You might very well have a really good sense of something, but you are not expressing it to the readers here clearly.
Bingo! That’s one point I was trying to make - maybe what the market optimizes is the use of capital, and one way it does that is via marketing. Perhaps effective marketing is far more important to the goals of the company - making money for the investor - then minimizing truck miles (assuming they are not already minimized).
Easy. The first package, and all the others on the first run, paid for morning delivery. The 2nd ones paid for a later delivery, but a bit less money.
Alternatively (and I am sure I can think of alternates all day but what’s the point) either there is a big rush of packages that day that overrode regular capacity to deliver, or regular capacity was diminished via breakdowns, driver absences, errors, etc.
That’s life in a complex system. Sometimes you don’t perceive the longer term maximum and only see what feels like a inefficiency. sometimes if you are striving to run very close to capacity all the time, as FedEx surely is, a temporary breakdown will cause a temporary glitch. You just deal with it and move on.
Same thing happens at airports with bad weather delays. Planes end up in the wrong place at the wrong time, there is not enough excess capacity to simply use a different plane, so people get stuck until it all clears up.
Doesn’t mean the system is globally inefficient just because any of that happens.
I have never been on your block, but it doesn’t sound like anything you said so far. I thought we were talking about multiple trucks from 5 companies doing pickups, not 2 trucks from one company doing deliveries.
Great! you have been thinking about this at least 7 months. Can you share the model you created? After all, you seem quick to accuse others of ideology - lets not talk public or private, lets just look at detailed models.
You yourself described a bewildering menu of services from teh trash company. Whatcha talkin’ bout now Willis?
No of course not. Maybe the pizza come out at different times. I don’t want mine to get cold and soggy just because of the jerk next door, do you? Maybe there is such demand that the drivers are huslting and there is more than one in my neighborhood. They get back to the store, grab a stack, and go. That the very next stack has a pie next door to mine from the previous stack, probably no one knows or cares.
He is working for tips. He will figure out what is best - the answer might not be the same each time.
Certainly the truck packing algorithms will strive for that, but errors happen sometimes. Or, like the pizza, if there are going to be 2 trucks there n the neighborhood, then that gives the packers some leeway as to how to make it fit, they could put two separate shipments in separate trucks to optimize the bigger picture somehow. Nothing wrong with that. Nothing at all.
More flexibility in decisions means they can buy equipment/capacity more closely matched to demand. Maybe smaller trucks in your neighborhood and bigger ones downtown instead of medium or large everywhere. Its the same as how you choose what size glass to drink out of - depends on how thirsty you are, how much ice, etc. Or else you suffer from glasses that are too small (more refills) or two large hard to hold, mostly empty all the time).
But they don;t have trucks to match your paid for volume exactly, since no one will always have full volume, the trucks will always be partly empty if they did. Maybe a lot empty, in which case the trucks are simply too big and capacity is wasted where it could be used across town somewhere maybe. A smaller truck that matches predicted demand, as opposed to paid for volume, could run 90% full all the time with the same trash put out. The day after Christmas, when everyone has a full barrel, then you have to hustle, but the rest of the year it is worth the savings.
If you have unexpected full loads, then another small truck nearby can take up the slack.
For me they would because 1 company would be unlikely to support experiments in the variations of services that you described, at least not simultaneously, and probably never.
Maybe you are the one with ideology on the mind. He is only describing models, and quite well I might add.
Wjhy you guys keep talking about dicks instead of simply describing a model, maybe even one that leads eventually to the conclusion of the OP. We are on the 3rd page already, isn’t it time for the big reveal?
What was the underlying principle of the modeling? What was the definition of efficiency?
Yeah but you are jut talking to hear yourself talk because you haven’t described any of the the terms you are relying on, including what “efficient” or “fair assurance” means to you. Basically you are dismissing everything everyone says because of your intuition. Pardon me while I don’t pay much attention.
LOL talk like that would not have gotten you far in a place where you actually need to justify conclusions.
The pricing and service menu you described indicates the companies themselves do not see it as a commodity, particularly as regards their investors, who they are ultimately responsible too, no one else. In the business world, “commodity” does not mean what you think it means.
Or we could break this down to the main assertion - that it’s always better to have one big organization delivering a commodity to a region than to have five smaller companies delivering the same number of commodities.
Do you really believe that’s true? Would you rather have only one big grocery store that everyone shops at, rather than five smaller stores? Would you rather have one big parcel delivery company? Do you think these monopolies would remain efficient without competition? Would they go out of their way to make better products, treat customers better, or constantly re-engineer their processes to weed out the fat if they were the only game in town?
The main argument for bigness is that it can bring economies of scale. But economies of scale are specific to goods that have declining marginal costs and large fixed costs. Service delivery is not like that. There’s nothing inherently more expensive about having five companies do the job of one then the cost of the job does not decline due to economies of scale.
So we’re left with the perceived inefficiencies of five trucks leapfrogging each other to pick up trash instead of having one truck come through and do it. Your claim is that this is intuitively and obviously inefficient. But you haven’t quantified that at all, and it sounds like you haven’t really thought through the math of it.
Have a look at this cite and it’s references to get a general idea of how complex routing optimization can be. If a city sends one truck into each neighborhood, this could actually be very inefficient if it results in more overall zones, less capacity utilization per truck, more miles driven per truck, etc. It’s not intuitively obvious to me that five trucks operating in a mesh-type arrangement is inherently less efficient than those same five trucks each taking a different zone. I’d have to see the specific details of the route.
Even if the actual curbside pickup was less efficient, you still don’t know how that effects the overall efficiency of system, because you don’t know what percentage of total costs are affected by this perceived inefficiency. It could be that the old system was so inefficent in other ways that five companies could come in, split the market up between them, and still make a profit at current prices because of efficiency gains elsewhere.
In short, we don’t really know much of anything outside of one piece of anecdotal evidence, and drawing sweeping conclusions from that is folly.
I take it neither of you are familiar with The Black Swan by Nassim Nicholas Taleb. You both work very hard to interject the most extreme possible circumstance.
You think my street is an outlier? What exactly do you think happens on all the other streets?
The Fedex driver is suddenly having a busy day? (btw the price structure was the correct answer)
The pizza came out of the oven at a different time?
Weather delay at an airport?
You both are trying very hard to add an insane amount of complexity to an otherwise normal scenario.
Trash collection isn’t new, and I am NOT trying to reinvent it. And as I said, and you both agreed, I had never seen a situation like this before. All my life I have lived with city collection, and simply assumed that was standard.
Then I moved and found it was private. Which to me presented an interesting data point in the comparison between public and private services. We are all very familiar with water, gas, power being a public utility, and more recently becoming private.
So here is a case study were trash collection is privatized, and these are my observations of that system. It is not my final project or a thesis, it’s a fucking internet message board.
It’s also not nearly as complex as the two of you are trying to make it. Right now 5 companies provide garbage pick up. Not just to my street, but to the entire city.
And so it’s funny to see the lengths you guys will go to to try and avoid the obvious answer.
If my city decided that they were going to take control of garbage collection, by issuing a contract to a single company–creating a monopoly–think about the bids each of the 5 companies would submit. Since you both know so much about this, imagine you’ve been contracted to help with a bid.
Do you think any of them would suggest a system in which 5 trucks overlap each other picking up garbage essentially at random?
Like I said, my provider is different from either of my neighbours, so they get a different truck than I do. Do you think a company bidding on that contract would stick with that model? Or are they more likely to just have one truck go down my street picking up trash at each home?
Oh, and since not_alice asked for my model: it is the same as every city that has a single collection company be it government or monopoly. Areas get a trash day, and on that day a truck goes up and down the streets picking up trash sequentially. The size of trucks, and areas they drive, being based on volume/weight. Combined with all the nifty tricks and gadgets that go along with trash pickup.
The current model involves what looks like an ordinary collection system: picture 10 houses on a street, in an otherwise normal subdivision, in an otherwise normal city.
My system has a single truck go up the street, collecting trash one house after another. Eventually it turns around and picks up trash on the other side.
The current system in my neighbourhood looks quite similar. Except that the houses are *randomly **assigned to 1 of 5 companies. Each company goes up and down the streets picking up trash sequentially from their assigned group.
If you were working for one of those companies, would you suggest to the boss, “hey, let’s take our clients, assigned them randomly to 1 of 5 groups, and set up our pickups based on that.”
*I use the term random because each company doesn’t have control over their client list, unless they choose to reject clients they deem unprofitable. I wasn’t aware of this happening.
I will look it up, but what will I find? That seems an unusual title for a treastise on process optimization! And I highly doubt any such treatise will disagree with me or Sam_Stone on this.
We don’t know. That is the point. We don’t have any data, and you aint sharing
We can only tell the about the types of data it takes to model this properly, and we have both just scratched the surface. No way I could make a report to a CEO or investors on a model just on what’s been addressed here. Not in a constrained market with 4 or more competitors anyway
It was one of several correct answers I gave you, yes.
Maybe you should go to a really busy pizza shop that delivers. See if pies come out the same order that calls come in…
Fly much?
Not really. It only seems complex to you. We do this stuff every day (if I may speak for Sam Stone for a moment), probably could do it half asleep. Its nothing really.
Clearly. But the people responsible in your town sure are.
Baseball wasn’t new either when Billy Beane decided that maybe the traditional statistics and strategies in baseball were not really as effective as was commonly believed. He undertook some statistical analyses, and applied them towards building effective and inexpensive competitive teams.
How? By looking beyond what “everybody knows” to see what really is there in the data.
I have never seen a town with such delivery. As an abstract system, it is not far removed from much of my experience. Commercial trash pickup is probably pretty much the way you describe, especially in suburbs where access is good, which probably lends credence to the idea that residential might work well using that approach. I doubt the idea just dropped from the sky. It could have been something as simple as a City Council guy expressing an idle thought with no real preparation - “Why not do it in the residential areas the way we do it out by the malls?”.
Those are structurally different. You are connected to a fixed network that acts as a conduit for the good or service. Where the good or service enters the conduit doesn’t really matter. But you can’t really choose another conduit except in the rarest of cases (maybe you can install sufficient solar energy and drop off the grid, or find a good well or something like that). Trash works like none of these, other than you said you can opt out and haul your own trash - but I bet you can’t let it pile up either!
Have you learned anything? Maybe about not taking about the size of your dick at least? We are all here to fight ignorance after all.
I don’t think anyone here has described anything complex at all. Not conceptually anyway. If you had to actually run the model, well…
Now you are getting into an entirely different area, which is auction economics. And the question there should be, how should the city structure the auction in order to maximize the revenue (or whatever) to the City? And that is not going to be as straightforward as you might thing. Witness the successes and horrible failures surrounding the auctioning of wireless spectrum around the world.
Were I to advise the City, I’d counsel them to get it right the first time. Were I counseling the companies, I’d advise them on if the contract is even worth bidding on as far as their investors go, and if so, then what an appropriate strategy for bidding would be based on the structure of the auction.
Oh hell yeah!
Why do you think the US has as many wireless companies as they do, and they don’t have their own territories? Because it keeps them on their toes for one - good for consumers! Because it maximizes the revenue to the auctioneer when there is more than one winner.
In fact, OTTOMH, I would consider advising the City, depending on its size, to break down the city into districts of roughly equal business demand, and auction each one off individually, quite possibly simultaneously. Since it is illegal for the companies to collude, and their revenues will be maximized by having more scale, they likely will all bid on all the territories. Nice, huh?
Like I keep saying, what are you trying to maximize? They might. Maybe more nimble trucks are more efficient, who knows, we haven’t mentioned that yet. In my neighborhood, the trucks don’t use labor to pick up the can, there is a remote arm. Maybe it is way cheaper to have 5 trucks with 5 guys, than 1 truck with 3 guys as used to be more common. Less spilling, less sweeping, etc. And maybe the trucks are versatile and can be used for trash one day and recycling the next. I think yard waste pickup is not common everywhere, so there is room for growth once you are in…
So it is a traveling salesman problem you are talking about? Notice what there very first sentence n that link says. If you are able to solve that problem, you my friend are going to be very rich and famous indeed momentarily.
If a model suggested that would improve efficiency, and match with the overall business model, not only would I, but I would expect a very nice bonus out of the increase to the bottom line.
Here is yet another possible reason, which you can’t possibly have observed: The trucks you use to implement that break down way les frequently, and any extra cost on the road (if there is any at all) is more than made up for in reliability.
I am sure their marketing groups observe clusters quite carefully. As noted above, perhaps discounts are offered to neighbors of existing customers. I highly doubt that the current distribution of customers is random in any sense of the word. Do you have any statistical evidence of this new statistical claim?
