Ok, as usual, the press gets it half right. For that matter, Annise Parker isn’t exactly on the money either.
The project was supposed to cost $50 million and be complete in 2003. Instead, it’s now costing approximately $75 million and won’t be complete until 2008.
“The failure rate is beyond anything that we should have experienced,” Houston City Controller Annise Parker said.
She’s right about that much. How bad is it? Bad. Very bad. Is it as bad as it’s said to be? No, not hardly.
Lets get the worst out of the way: The program was ill-conceived, poorly planned, terribly executed, over budget, and has never worked as advertised. Inasmuch as it has caused numerous problems, resulted in many bill estimations, and caused hundreds, maybe thousands of man-hours per month to be devoted to correcting bills, and resulted in a great decrease in the public’s confidence level even before the story finally broke, it should be probably regarded as the single biggest failure in city administration over the last ten years, short of outright corruption. (Which is not to speak of any specific incident, just to say that I put malfeasance in a class seperate from mere incompetence.) Well, ok, it’s in the top two, anyway.
It will be made clear that the problems stem from decisions made as far back as the Lanier administrtion, but the key mistakes were made during the Brown era, and that the managers involved have since taken their retirement pay and run for the beaches. The current team running Public Works and the section responsible for meters (Utility Customer Service) is all-new, having been shuffled in after the Pension Massacre of 2004. They have inherited a very bad hand, and are trying to play it as well as possible, after about eighteen months of learning that their hole cards are deuces. They’re dealing with the administrative issues well enough, but this is a fundamental city policy–a decison to abandon this program wouldn’t be made in Public Works, not even by Marcotte. Whether they should just fold and ask for an expensive re-deal is a matter of public policy and high-level decision-making by the Mayor and Council–something we may now see if this story becomes a major public issue.
The good news is that, while it’s bad, it’s not nearly as bad as it looks when you see “47%.” If you think it means 47% of the bills are wrong, you’d be way off. You could take that on faith or my word, but I suspect that you, as the reader, want a bit more to go on. To explain why it’s not as bad, I have to explain exactly what is failing and why. It’s not very technical and I’ll avoid jargon as much as possible. In this article, I’m addressing just the technical aspects necessary to show why the 47% figure is highly misleading. History of the decisions and the people repsonsible will have to wait for a future article.
Ok, obviously there is a water main either in front or back of your property (or rarely, side). A line to your property branches off from it and enters the meter box, where a valve (the “city valve”) controls the water flow. The water passes through the meter into the customer’s private lines, which usually have their own master valve. (Note: the meter is the end of the city’s involvement; any problems downline from the meter, including the connection to the customer’s line, are the customer’s responsibility.)
If you want the technical stuff about meters, this site has a lot of very good information understandable by a lay person, though it doesn’t discuss the transmitters much. The short explanation is this: all meters used by the city (for residences and small commercial/apartments) have three basic parts.
- The meter case, explained in detail here
- The meter register
- The transmitter (or ERT as it’s often called)
Water only passes through the case– not the register, which is a seperate unit bolted to the top. The transmitter is usually bolted to the meter box lid, and connected to the register by a short cable. I’ll skip the discussion of the magnetic coupling between the case and register; see the link above for details. For the meter to record water use accurately, only the case and register need to work correctly. Frankly, as a rule they do; it’s a very old and reliable technology, and while the materials have changed slightly over the years, meters have only gotten more reliable, not less. I’m not exaggerating when I say that 95%+ of customer complaints have zilch to do with the meter (You’ll see for yourself in a moment.) The problem is the the third part, the transmitter, and a lot of it was the crappy first-generation technology we got stuck with from one of our vendors. Who, I am happy to say, became a former vendor because of it. (Sometimes the system actually works….)
As long as the first two parts work, a meter reader, an inspector, or the resident can open the box and look at the meter. It reads just like an odometer on your car, except that for approximately 80-90% of the residential meters out there, the “speedometer” is actually a needle recording 0-10 gallons of usage. Every time it makes a complete sweep of the dial, that’s ten gallons and the first number rolls over. Ten times around the dial is a 100, and 100 times around is 1000 gallons. And that is the first time Public Works will take note. Meters are read and billed only in thousands of gallons, anything less is ignored. (For reasons of both human and mechanical engineering, it just isn’t practical to bill all the way down to discrete gallons.)
Where the system has been fubared is that the third part, the transmitter. All this item does is transmit the existing read from the meter to a handheld unit carried by a meter reader walking by. Now look at this the way an engineer would do a fault analysis. Pre-transmitter era, you have the meter, the meter reader, and recording unit s/he input the reads into. Now you’ve got the meter, the transmitter, and the recording unit. Supposedly, you’ve eliminated the human error factor, making things faster and more accurate. That was the whole point of the plan. But this is true only if the transmitter is more accurate (or rather, reliable) than the human. Otherwise, all that has been done is to add in a new “failure point” where something can go wrong.
And it did, in spades. All too often, the transmitters gave screwy data, or no reading at all. This has no, repeat no, effect on the meter itself; the reading is there, and opening the box is all that’s needed to obtain the correct reading. (Why this isn’t done will be part of the later article.) The best figures I have been able to obtain indicate something in the area of 10-12,000 meters are failing to give any reads during a given month; I have no data on how many are giving bogus reads (because until we find out, we don’t know they’re bogus of course). However, the number, based on the corrections issued when they are discovered, is probably only a few thousand a month and it’s very rare for any of those to be anything other than transmitting the last available reading (which might not be the exact reading at the time the reader passes by). Often the register and transmitter will “re-synch” later. This is less common now, but it used to be a constant issue with the first-generation transmitters.
Still, it happens enough to poison the atmosphere, and keep us from saying “it can’t happen.” (Which in theory, anything can happen, but the chances of you being the only one it’s happened to in years is pretty miniscule. Never stops anyone from claiming their meter just decided — all on its own — to record a whole lot of water. “Obviously it’s incorrect.” Riiiiiiight. Nobody told me the city invented “perpetual motion” meters that can violate Newton’s Laws!)
Let me point this out: any City of Houston customer has the right to read his or her own meter and compare the reading to their own bill. Frankly, the city would much rather the customer did it than called us to demand that someone be sent out RIGHT NOW to read their meter. (Ok, take a number, buddy. Just like the other 100 that called this morning. No, I’m not kidding.) Remember that the City bills only in thousands of gallons, so the last few numbers on the meter will be dropped (usually the white numbers on a black background are not read, but other color schemes exist). So if you have a city bill and think your transmitter is screwy, open the meter box and compare it to your bill. Allow for the fact that it’s been a few days since Public Works read the meter; your reading should be equal to or a little higher than the one on your bill. (Please be careful not to damage the wire. I’m told that wire damage is the leading cause of transmitter failures now.)
Twelve thousand meters sounds like a lot. It actualy works out to 2.72% out of the city’s 440,000 meters. This goes to show just how accurate the City needs to be each and every month, when a success rate of over 97.2% isn’t good enough.
The transmitter is also known as the Encoder Recorder Transmitter, or ERT. Early models had a less than stellar performance. To avoid dependancy on one vendor, the city insisted that two interested bidders work out a compatible system, and it would buy from both of them. The vendor making the remote reading system was also required to cooperate. It took a lot of head-knocking to get them to do it, but scores of other utilities out there may have benefited from the city’s foresight; this forced cooperation resulted indirectly in an industry standard, simplifying everyone’s logistics and reducing costs. (Actually, in early trials, it was a little too standard. Readers picked up some HL&P/Reliant electric meters!) Since these transmitters were going to be in “pits” (meaning in-ground), they had to meet a certain standard for waterproofing; NATO military standards were adopted as the best available reference.
Of the two vendors, the second’s register and transmitter were extremely complicated, requiring a careful (and touchy) multi-step setup, involving a special tool. An average high-school graduate with a day or two of training could do it. Whether he’d do it right, in the mud, on a weekend, in lousy weather, for eight hours a day, is another matter. (And not lose the tool, as so often happened, resulting in improvisation.) I believe I’ve noted the problems with hiring qualified and motivated help at the City’s low pay scale? Let me put it this way: on meter reader job postings, it reads “GED or HS diploma preferred.” (The questionably bright side is Public Works isn’t scraping the bottom of the barrel — that would be Solid Waste. Those postings should read “Applicants with no outstanding felony warrents or drug habits are preferred.” Again: No, I’m not kidding.) So anyway, this careful and touchy setup often got done wrong. By the end of the first three years of the program, it had been done wrong over 6,600 times (based on the number of such register/transmitters that had to be replaced.) Additionally, the NATO milspec turned out to be woefully inadequate for electronic items constantly immersed in water and often run over by kids, bicycles, cars, trucks, and the occasional large lawnmower. Anywhere a meter can be put, it will be put. And the transmitters have never been up to that level of abuse.
Even in the best of conditions, the transmitters supplied by the first vendor (the one eventually kept) experienced some problems, mainly in the connection between the register and transmitter, and in moisture penetration causing the mis-reporting problems mentioned above. However that vendor proved to be much faster and more cooperative in improving their hardware. Eventually, it was decided the problematic second vendor would be dropped; Utility Customer Service would rather be stuck with a single vendor than knowingly install defective registers/transmitters. But of course, by that manufacturer’s view, they weren’t defective, it was our incompetent installation. “No tickie, no shirtie, sowwy!” So the city took a bath on the already purchased registers, and they got tossed in the dumpster.
Until Wayne Dolcefino famously crawled in after them on camera and made a big stink, that is. Apparently, he either never thought to ask why they were being thrown away, or he got snowed by the management at that time. Probably the latter. As much as I dislike him, I can’t see him being that stupid. Of course, since he’d “scored” with a story on waste already, maybe he was just lazy? (Needless to say, the first few years of the program, he and Marvin Zindler were said to be familar faces down at 4200 Leeland. Yet they never got to the bottom of the problem.)
As for the installation, just like the article says, a lot of damaged water lines resulted, usually from torquing (twisting) the customer’s PVC lines. Customers were a little torqued themselves by that, y’know? So, the private sector got called in to rescue the project. Hey, everyone knows the private sector works better, right?
Wrong. (One day I’ve got to finish that series.)
Naturally, the private sector went out and hired the cheapest help available; essentially the same level of employee the city had. And since they didn’t answer to city supervisors who were getting burned whenever things went wrong, they were even less motivated to get it right. A common lament in 2001-2002 among rank and file Public Works employees was “These contractors are killing us! When are we getting rid of them?” (Fair warning: they’re coming back soon.)
As can be seen from the above, there was a huge failure rate among the transmitters. The important point to remember is that except for a few thousand supplied by the eventually-rejected vendor, the failures didn’t affect the consumption recorded on the meter. Only one of several models they supplied had a design flaw that rendered the register itself unreliable, and replacing those became a high priority. Except for those, it only affected the consumption reported by the tranmitter to the meter reader, which was bad enough. However, I don’t believe the figure of 47% given is accurate. I can’t pin down a substitute number without hard figures that my sources don’t have access to, but many meters were replaced not because of a problem with the transmitter, but because the customer complained.
Here’s another truism that the public simply doesn’t believe: Meters “slow down” as they age. It’s an article of faith for most of the public that the meters get “looser” and “spin easier” as they get older. Well, even if small meters (often called “cold-water displacement”) were made that way, the customer would still be wrong. See, when a meter has set in the ground for 10, 20, or even 30 or more years, a great deal of water has passed through it. Water that, for the last 20 years, has become increasingly harder as the City converts to surface sources. Boil a pot of tap water until it’s gone (but do it in a cheap pot you don’t care for). Look at the whitish crud coating the inside of the pot. You won’t be able to get rid of it short of heavy scrubbing, if then. That’s calcium carbonate, and it eventually coats the inside of everything, including your meter. Which means the meter starts to slow down as the disk drags across those deposits. Eventually, it starts sticking intermittantly, and can cause up-and-down bills. Finally, it gets stuck entirely, leaving the city no choice but to estimate bills until it can replace the meter. (Yes, it’s legal. And the city has accelerated its replacement guidelines in the last several years to deal with this.)
This whole process takes several years; sometimes a meter will just stop but usually the “terminal phase” is anywhere from six months to two years. So, let’s replace the old, dragging meter with a new one that isn’t “under-registering” consumption. Guess what happens? An irate caller starts screaming at Utility Customer Service, “That new meter is defective! I never used that much water before!” Somehow, they never listen, even when confronted with records from several years back, documenting the gradual decline of the meter, and a consumption that now that exactly mirrors the consumption back before the old one wore out. They look at that one minimum bill for $8.50 and declare that is their average.
In fact, thousands of them were so stubborn about it over the last few years, that the city tested their meters free of charge, on site or at its shop. And that’s the problem: to test the meter at the shop, it had to be removed and replaced by another meter, which required a city worker, a vehicle, tools, and time (including travel). These are resources that had to be diverted from elsewhere; they weren’t sitting around waiting on a call like the Maytag repairman. Replacing the original meter afterwards would tie up more resources that weren’t available, and result in a “used” meter being in stock. Since all meter changes have to be carefully documented to ensure the right meter is read for each customer, putting the original back would not only require more man-hours; it would require as many as six meter readings over the course of a month to be compared to determine the correct consumption, and perfect documentation of each change. (The city doesn’t have the option of just ignoring the usage through the ‘loaner’ meter; by law it must bill for all water sold to the customer.) That’s asking for trouble, so the city doesn’t bother; it just leaves the replacement meter out there. Until recently, the tested meter was simply discarded.
In short, to satisfy skeptical customers, the city was eating the cost of several hundred meter changes per year. (That the Department’s reputation is so low as to give the customers a reason to be skeptical cannot be denied.) Additionally, after about 2002, the sole remaining vendor remedied one of the problems with its cable by agreeing to a city demand: to manufacture the register and transmitter as one unit; no assembly required (except for bolting the register to the case, which was simple). Cases, on the other hand, are delivered seperately. (Edit: When replacing an existing meter, the city now generally replaces the case only if it’s over six years old.) So now, if a transmitter is defective, both the register and transmitter have to be changed. But! The meter’s serial number is on the register, not the case–so it is recorded exactly as if the entire meter has been changed. (I know, that sounds stupid to me too. Legacy reasons, I’m told.) With transmitter problems still being suffered (though now most are damaged cables), every replacement of a transmitter is recorded as a meter replacement, inflating the figures tremendously.
Of course, this also means that if the cable between the register and transmitter is damaged, a complete replacement appears to have happend, because again, the whole assembly has to be replaced. (This happens far too often, and accounts for most of the 12,000 meters per month that can’t be read automatically.)
So 47% of the meters being defective is incorrect. Now 47% of the registers, transmitters or wires. . . that’s possible (over time), but all but the tiniest fraction of cases (we’re several places to the right of the decimal here, ok?) the problem wasn’t the registers and did not affect the actual meter reading. Of course, that’s not to say that the customer’s bill didn’t go through some “interesting” gyrations before the Department got around to opening the box to get the correct reading. And maybe again afterwards for that matter.
For the City of Houston, the problem is this: From a certain point of view, it doesn’t matter if you’re right or wrong–if the customer has no confidence in the bill he or she is getting, you might as well be wrong for all the good you’re doing.
Well, that will have to be it for now. I have a great deal more information on the sordid history of this program; the reasons, both political and bureaucratic, behind its inception and its failure.
There will be familiar names in that article…
(Edited for clarity at 12:06 a.m.)