Measure to Measure: Data, management, and the reliability of the mail



Reliable.  It seems like a simple and straightforward word.  It conveys, if not a very specific meaning, at least a pretty concrete sense of satisfying expectations.  In some contexts, like certain forms of statistics and technical applications, reliable and reliability refer to the overall consistency of a measurement or the accuracy of the data being measured.

But when we’re talking about the postal system, the word reliable means much more than mere consistency.  The reliability of the mail goes to the heart of what we expect of the postal system, especially a public postal network based on a promise of universal service.  But what does reliable mean in this context?

For example, if for years it took two days for your mortgage payment to be delivered to the bank, but now that service standards have changed it consistently takes five days, would that be considered reliable?  The Postal Service seems to think so, and it has produced marketing surveys showing that customers care less about the speed of delivery than they do about knowing the mail will arrive when the Postal Service says it will (even if that’s days later than in the past).

Or, to take another example, the news is filled with stories about people complaining that their mail is arriving much later than it used to.  We might all agree that such anecdotes aren’t as reliable as scientific data, but if a system designed to measure performance ignores customer feedback or shunts complaints off into some nebulous customer satisfaction index, aren’t we missing an opportunity to identify problems?

The issue of reliability in the postal system involves a host of related questions:

If we choose a narrow and technical means to measure performance, is it enough to say the network is providing reliable service, regardless of other evidence?

Should we just look at a statistical measure and check to see if the results have the proper p value and demonstrate some acceptable probability of being accurate, and then conclude that the network is reliable and give ourselves a pat on the back and move on?

If we replace a longstanding, well accepted, relatively straightforward, and externally conducted system for measuring on-time performance with a new system that’s conducted internally using complex statistical sampling, how will it affect customers’ perception of the reliability of the mail?

Can the success or failure of a public postal network be defined by simple statistical measurements?  What happens when the measuring becomes an end in itself?  


A shift of focus

For almost ten years now, we have been defining the success or failure of the Postal Service by an arbitrary concept of profit and loss that is based on the false premises of the 2006 Postal Accountability and Enhancement Act (PAEA).  These values may be suitable for a corporate entity, but they are not appropriate for a public infrastructure.  Now the chickens have come home to roost.

In response to the mandate of PAEA to act like a business and focus on the bottom line, the Postal Service has focused on a two-pronged strategy — cutting costs and shifting the focus of the institution.

To cut expenses, the Postal Service has reduced service standards, closed facilities, eliminated hundreds of thousands of good jobs, and become more and more dependent on part-time workers and cheap temporary labor.

As damaging as these cuts have been, the shift in the institution’s focus has had even more destructive consequences.  We have witnessed the transformation of a public network with a mission of universal service into a corporatized delivery company that treats this mission as an unwanted burden.

Lately many members of Congress have been raising questions about the Postal Service’s on-time delivery performance.  What percentage of the mail is being delivered within a particular service standard (two-days or three-to-five)?  Are rural areas getting worse performance than urban and suburban areas?  What measuring systems are being used to determine the percentages?

Rather than focusing on such questions, we should be talking about what the reliability of the mail really means.  Rather than worrying about the numbers, we should be looking at how the mistaken assumptions and false premises embedded in PAEA have undermined the public mission of the Postal Service.

But here we are.  If we can’t fix the ills of PAEA, if we can’t acknowledge that constant cost cutting and abandoning the basic principles of universal service together lead logically to deteriorating service, then maybe we can at least learn something from the recent focus on service performance measurement.


Modern service standards

The law that governs the service performance of the Postal Service appears in Title 39, Section 3691: “Establishment of modern service standards.”  This section says that the Postal Service, with consultation with the Postal Regulatory Commission, shall establish “a set of service standards for market-dominant products.”  The law goes on to enunciate a set of objectives that the standards are designed to achieve.

These objectives are described as follows: (A) enhance the value of postal services; (B) preserve regular and effective access … in all communities; (C) reasonably assure Postal Service customers delivery reliability, speed, and frequency consistent with reasonable rates and best business practices, and (D) provide a system of objective external performance measurements for each market-dominant product as a basis for Postal Service performance measurement.  The section also states that with the approval of the PRC an internal measurement system may be implemented instead of an external system.

The ordering of these objectives in the law under PAEA is interesting.  The fact that “enhancing value” comes first — before preserving regular and effective service or issues of delivery reliability — demonstrates in a nutshell the shift in focus that lawmakers mandated for the Postal Service with PAEA.

Gone is the concept of value embedded in a universal postal network designed to serve as a basic public infrastructure.  Instead, value is measured in the terms of the marketplace, where it’s all about profit and loss.


Two PRC dockets

The question now being discussed in the proceedings of the PRC and elsewhere is whether the systems that are being used to measure the performance of the postal network are giving us an accurate picture.  There are two PRC dockets looking at this issue.

The first, docket number PI2015-1, examines the Postal Service’s request to discontinue the current system for measuring on-time performance, EXFC — the External First-Class Mail measurement system.  In this system, participants mail bundles of letters across the country, and an external company (IBM) measures the time it takes for the mail to be delivered.  The Postal Service wants to replace EXFC with an internal system that relies on statistical sampling instead of tracking individual pieces of mail.

The second docket, PI2016-1, was spurred by the GAO report “Actions Needed to Make Delivery Performance Information More Complete, Useful, and Transparent.”  That report, with its mouthful of title, expressed a desire to have more of the mail measured in a census manner rather than by sampling and to provide data that could help determine if rural areas were suffering lower performance than other areas.  (The GAO also encouraged the Postal Service and the PRC to make the performance reports more available and accessible, which the PRC has already done.  You can find all the reports here.)

As the notice establishing the docket says, the focus of PI2016-1 is “to address how the Postal Service may improve the completeness of its service performance data.”  Stakeholders are invited to comment on “deficiencies with respect to the accuracy, reliability, and representativeness of the current service performance measurement data.”


The visibility of the mail

A basic premise underlying both dockets is the idea of the visibility of the mail.  Thanks to the advent of cheap data storage and scanning technologies that capture information-rich barcodes, we have come to expect that we should be able to know where every piece of mail is at every moment on its entire route.

That’s not just a matter of being able to locate a piece of lost mail.  It’s about seeing everything as a means of control, in the same way that large postal facilities (like prisons and other institutions) have balconies from where management can watch workers.  For many in the industry, this panoptic vision of the mail has become the Holy Grail.

Technology is certainly a wonderful thing, but it should be seen as a tool, not an end in itself.  As Thoreau warned over a century ago, “We do not ride the railroad; it rides upon us.”

The very astute media scholar Neil Postman made much the same complaint in books like Technopoly: The Surrender of Culture to Technology and Amusing Ourselves to Death.  As Postman shows, technology comes with costs — monetary, human, and cultural — and too often we mistake what are little more than neat parlor tricks for great advances.

The idea that we could track every piece of mail from end to end has a certain attraction.  The accumulation and analysis of such data would presumably give us a concrete sense of the performance and reliability of the postal network.

But would this information really tell us anything more than we already know?  Even if it did, would this new information be useful or, to borrow a fancy term from the intelligence industry, “actionable”?  And perhaps most important, if the information were truly actionable, how likely is that postal management would act on it in the appropriate way?


Ignoring what we know

When we look at surveys about what mailers expect from the system, we find that expectations are pretty basic.

If I put a letter in my mailbox for the carrier to pick up, I want to have some reliable sense when the letter will arrive at its destination.  If the publisher of a county newspaper puts her papers out on Wednesday evening for local delivery on Thursday, she needs a high level of confidence that those papers did indeed get delivered.  An advertiser needs to have a sense of when and where (how deep into the network) to drop mail in order to get delivery on a date that corresponds with the upcoming sale.

Knowing where a particular piece of mail is in the system at any particular moment may have some value in identifying bottlenecks, but the truth is that in most cases the best a totally visible system can do is tell us where something went awry, where and possibly when something got lost. The truth of the matter is that we already have a good deal of information about how mail moves through the system.  We just don’t listen or follow through.

The problem isn’t one that can be solved solely by technology.  It’s a human problem, and it lies with management.  I’ll give you an example from my own experience as a postmaster handling the local newspaper, The Sylva Herald.

The paper was dropped in Sylva on Wednesday evenings for dispatch to the plant in Asheville (now closed), with delivery to post offices in the county on Thursday. If the papers didn’t come to my post office, I knew one of two things had happened.  Either the sacks weren’t prepared properly (e.g., they were mislabeled, sometimes by the publisher in a rush to get the papers out on time) or they were mishandled at the plant (which usually happened when we got a new plant manager and the system got changed).

Both of these problems are basically human and system problems, and to the extent that a problem occurs over and over again, it is almost always due to someone not listening, not paying attention.  Tracking every piece of mail end to end doesn’t really address such problems. We can scan every sack of mail, every single piece of mail, all we want and all the technology allows us to, but if the culture that supports the system fails to listen, then failures will occur.

Look at the minutes from MTAC (the Mailers Technical Advisory Committee), or read the trade papers of the mailing industry, and you will find repeated complaints and frustration over the Postal One, the information management system used by business mailers when they enter their mail.  One thing becomes very apparent when you look at the complaints — a consistent failure by management to listen and pay attention.

The Postal Service has built a fancy system for collecting customer complaints. The PRC also has a means that allows people to make rate and service complaints.  Collecting, collating, combining, and tallying all these complaints are relatively easy.  Computers do most of the work. They can look at the data, manipulate the data, even determine where to assign blame for problems.

None of that matters, though, if the underlying institutional culture is not built on a foundation of listening and problem solving. Scanners, barcodes, and all the tracking technologies are wonderful, useful, and even essential tools, but ultimately the mail is a human system serving human purposes.

There is already a tremendous amount of feedback built into the system.  The problem is that the feedback is largely ignored.


Not the count but the culture

In docket PI2015-1, the Postal Service is asking to do away with the current external system for measuring single piece mail and replacing it with an internal system based on statistical sampling and scanning.  The new system is called Standards Performance Measurement or SPM.  The way it works is pretty complicated, but the basic idea is that rather than having an external company track specific pieces of mail end-to-end, the way EXFC does, the Postal Service would do the measuring itself and do statistical analyses of mail at various stages in the process.  (For more about the new system, see this previous post.)

For a number of reasons, what the Postal Service is asking for authority to do is very problematic.

First, there’s the issue of allowing the Postal Service to do the measuring itself.  The Postal Service argues that the law allows shifting from an external to an internal measuring system, provided the PRC gives its approval.  That’s true, but it also seems clear from the language in the statute that Congress preferred an external system.  If nothing else, it gives the measurement a greater sense of being objective and reliable.  In its various filings in the docket, however, the Postal Service largely ignores this preference.

One characteristic inherent to the concept of reliability is trust. For a system or a measurement to be considered reliable, we have to trust those who do the managing and the measuring. That’s a big problem with the Postal Service’s proposal and one that is unlikely to get addressed head-on in the PRC proceeding.

A second problem involves the Postal Service’s claim that EXFC is too expensive and the new system will save money.  However, as a 2012 OIG audit on EXFC points out, the contract for the external provider (IBM) was never competed.  Moreover, it appears that some of the costs associated with EXFC are actually costs to transfer the information from the program for internal use. It’s also not clear that the new system the Postal Service is offering will be any cheaper since its proposals fail to discuss the costs of having delivery employees scan mail on the route.

In any case, EXFC costs the Postal Service only about $41 million a year.  That’s 0.06 percent of the annual budget — obviously not very much in the scheme of things, especially if the system provides a consistent and generally accepted measurement of delivery.

EXFC is admittedly not a perfect system.  For years, there were complaints that the system was being gamed by postal managers who seemed to be taking extraordinary measures to ensure pieces that might be included in EXFC were given special handling.  As noted in the OIG audit on EXFC, preferential treatment is often given to mail collected from some blue boxes (which is included in EXFC) at the expense of mail picked up from individual mailboxes (which is not measured by EXFC).

But the Postal Service is not claiming that SPM will be an improvement on EXFC because it will eliminate the opportunity for gaming.  Instead, the Postal Service argues that the new internal system will be better than EXFC because it will provide even more information and give management a better chance to solve problems.

But the Postal Service has most of this information already, and it has shown little interest in acting on what it knows.  In any case, the reliability of the mail is not simply a matter of counting how much mail arrives on time.  It’s about the culture of the post office itself.


Managing the measures

“If you can’t measure it, you can’t manage it” is an oft-quoted line by management guru Peter Drucker.  There may be some truth to the idea, but as other management experts are quick to point out, it’s largely a myth.  Managers are too easily captured by measurements, and they often ignore feedback and information that is pertinent and useful but that can’t be put in statistical form.

Unfortunately, the management of the Postal Service has taken Drucker’s line as gospel.  They don’t manage the postal network, its people, its equipment, and its processes.  Instead, they manage the measures.  That’s how they’ve been trained, and it’s part of the culture of the workplace.

I’ve written many times about the failures of postal management (as in posts like this one).  Red Laws and Zero Bundle Drills, routines that are ostensibly designed to monitor processes that support EXFC, turn out to be nothing more than exercises in bureaucratic facility.

EXFC is supposed to measure how the postal network is performing under normal operating procedures. It never really did that, but the problem wasn’t with EXFC or its design. The problem was that postal managers don’t establish and manage normal operating procedures.  Instead, they design extraordinary measures to try and max the measure – something that is both costly and dishonest and often stupid.

If collection mail arrives after a dispatch and a carrier is forced to make a hundred mile round trip to the plant, an EXFC failure may have been prevented, but have we truly measured the effectiveness — the reliability — of the system?  And if we don’t collect and publish information on those failures, especially when they occur with some regularity, then can we reasonably say the system or the measurement is reliable?

During the PRC proceeding on the proposed shift from EXFC to SPM, several commenters raised concerns about situations like late collections or how postal employees might game the new system. In its reply comments the Postal Service insists that postal employees and especially postal managers would never, ever attempt to game the system in order to make higher scores.

For example, in response to a commenter’s concerns about late collections, the Postal Service defended the integrity of postal workers. “Postal employees at all levels of the organization,” the Postal Service says, “strive to provide quality service for the sake of providing quality service.  For all of its products, the Postal Service has a long tradition of field managers and employees going beyond the call of duty to improve customer satisfaction, irrespective of whether their actions are recognized or rewarded, or whether the results are captured by service measurement systems that generate reports submitted to the Commission.”

It’s true that most postal employees are dedicated to providing good service, although that may be becoming less true with an increasingly temporary and overburdened workforce.  But the fact is that nearly every postal employee could cite examples of USPS procedures that assume that employees are incompetent, dishonest, or just plain stupid.

Many managers operate on the same assumption.  It is a message that is built into the culture. Talk to a hundred postal employees and at least half of them will have a story about the utter failure of common sense — or worse — in the culture of postal management.


Managing the managers

Quis custodiet ipsos custodes is a Latin phrase that is translated as “Who watches the watchers?”  With the Postal Service, we can ask, “Who manages the managers?”

The Postal Service’s Board of Governors is supposed to serve this role, but it is a toothless political construction that serves no real function. The Office of Inspector General does regular audits, but its recommendations are frequently challenged or ignored by postal management, and they are primarily focused on finding ways to save money.  The PRC is supposed to watch over the Postal Service too, but it has limited powers, like determining whether a particular statistical technique is valid.

Given the absence of true oversight, the management of the Postal Service is able to act without being held accountable.

But accountability isn’t just a matter of achieving certain levels of on-time performance.  The issue isn’t whether we should make every piece of mail visible from deposit to delivery.  The issue isn’t whether the Postal Service, rather than an external company, should be permitted to manage the measures of performance.  And the issue isn’t making data and information about delivery performance and standards more readily available to the public on a website.

The systems already in place, the information currently being generated, the feedback from customers on performance — all of this is sufficient to design and manage a reliable public postal network that satisfies a robust definition of universal service.

No, the issue is the failure of PAEA to recognize the value of a public post. It lies with the utter lack of accountability of the current design.

Mail is becoming less reliable. The postal network has deteriorated.  These are obvious facts.  Arguing around the edges about performance measurement systems is ultimately futile. The problems are within PAEA, a law that fails the American public, and with a postal management that has faced little real scrutiny or accountability for forty years.

Mark Jamison is a retired postmaster.  He can be contacted at