The city spent $2.2M for the compactors, including installation and a service warranty. The estimated cost savings from the spend was $13M over 10 years; if they saved $900,000 in the first year that looks like a reasonable estimate.

Assuming they last the predicted 10 years. Someone else linked http://www.philadelphiacontroller.org/publications/other%20r... which says:

"Issue: Contract compactor specifications on expected life of product vague and not supported by data.

"The contract included compactor specifications that included an expected life of 10 years. When questioned about data or research to support this expected life, the BigBelly Solar CEO stated the life expectancy was an estimate based on the type of products and materials used in the compactors. He further stated that the compactor had a life expectancy that was estimated similar to an automobile; in that some repairs and parts may need to be made during the normal life cycle, but full replacement may not need to be made for theyears. When asked if there was any testing data to support the 10 year life expectancy, he indicated that he was not aware of any specific data."

Thanks Dan - didn't see that in the article, but if that's really the return on capital, its a significant one. I'd poke a little bit more as to what the real opportunity cost was here (ie, did you absolutely need to spend the $2.2M to get the fuel savings, or could a more efficient system of checking them with cheap transmitting devices work), but that's context that should have been in the main article.

Our city has these. The price I recall they paid was $4K per unit. I think they are using some really twisted math on these claims. A simple wire trash basket can easily be inspected by sight and emptied when full... it's not like the garbage trucks are dispatched to individual trash cans. They have standard routes at standard times, the only question is whether they need to stop and check each can.

It would seem to me that a better solution would be a simple, traditional open-top can, and a solar powered, high-mounted (ie, out of reach) remote camera with sensing technology would get most of the savings with a much lower up-front and continuing capital cost.

The main savings seems to be from being able to monitor when the cans are full, and only pickup when needed.

To me, it would be simple to write an app that read in a x daily image upload of the can which automatically calculated the level of fullness. It wouldn't require specialised can hardware and could drive most of the benefit (only emptying when required). You could double the size if more capacity was required.

Also, I would expect that these would fall into a very predictable pattern, with some cans requiring emptying 5 times a week, some 8 times a week, some twice a week. It seems some simple modelling would also give you many benefits for a very small upfront-cost.

I have visited many friends places that had compactors installed when their kitchens were new, and over time they just stop bothering with the compaction and use it as a regular disposal. Usually because the compaction is messy, faulty or just too much bother to use. I would expect public trash cans to have the same problems, but magnified.

The whole thing looks like a solution searching for a problem - bludgeoning a simple problem with too much technology.

How about taking the $100 cans, installing sensors in them that phone home when it is full? Power them with 9 volt batteries and replace them on a regular schedule. You get most of the benefits at a much lower cost.

Reading comprehension test: the article claims that the units communicate with a central trash pickup authority--units are only emptied when full. As a result, the trash authority can avoid fuel-wasting trips and optimize schedules on the fly to empty units when they need emptying. Supposedly, this optimization has, or will, yield roughly $1 million in annual savings.

See that's the part I don't understand. I did a separate numerical analysis that the largest cost of garbage pickup is paying people for windshield time not fuel cost. If you skip stopping at every can along the way, you cover the same distance slightly cheaper, but $900K implies an absolutely huge savings which doesn't add up because its nearly their entire departmental labor budget, not, I can drive from the dump to 3rd street in only 7 minutes now instead of 10 minutes because I'm not stopping at every can.

I would have to check my numbers. It might be that garbage men total cost of salary and benefits is a (large) multiple of $30K/yr (getting into software engineer pay territory) in which case windshield time is even more incredibly expensive. Then it might work out that reducing labor cost by 10% or so would actually work out to 900K/yr.

Presumably since the cans get emptied so much less often they can get big savings by skipping 3rd street entirely because none of the cans is reporting full.

Maybe, but they still have to drive past the 3rd street cans on the way from the full cans on 2nd street to the full cans on 4th street. I'm thinking windshield time is much larger than emptying time. Basically garbage truck drivers probably spend a lot more time driving than hoisting trash cans.

Its like the mailman argument. If I don't get any mail today, my mailman can't magically teleport past my house. True he does save the time of opening a box and tossing some sorted mail in, but that action only takes a tiny fraction of the time of walking 100 feet past my property. A registered mail transaction takes a LONG time like 5 minutes and if more people got registered mail (or fewer got junk mail) then the ratio would flip and he'd save huge time.

>Its like the mailman argument. If I don't get any mail today, my mailman can't magically teleport past my house.

No, but he can save a lot of time by not turning down your street at all.