Institutional Memory

Much of putting institutional memory into practice involves documentation, which oftentimes seems either too labor-intensive to do or not valuable enough to implement. After all, we remember lots of things, don't we?

If we're so good at remembering things, answer these questions:

• What was your junior high school locker combination?
• What is your license plate number?
• What color is your next-door neighbor's house? 

If you can answer all three questions without having to think about them, then you probably don't need a program for institutional memory. And, after all, you might think these to be three things you would otherwise remember spontaneously: You used the locker combination almost every day, probably for two straight years; you see your license plate almost every time you get into your car; and you almost certainly see your neighbor's house as often as you see your own.

But we don't remember things like these nearly as well as we think we do. And these are just the personal things -- imagine how much memory we lose of the things that are work-related that we don't see or act upon as often.

Recognizing how frail our personal memories are is a good way to start recognizing just how much help our institutional memories need.

People may resist the notion of documenting their institutional memory, since the process of documentation can take time and effort that don't really seem worthwhile. Considerable evidence exists to show that even things that we think we'll remember well ought to be written down anyway.

Consider the unreliability of eyewitness accounts of crime. The movie "Rashomon" tells the same story from four different angles, each different from the others. The movie may be fictitious, but its essential meaning is actually quite true: Eyewitness accounts of events are extremely unreliable. Even when factors like stress are taken out of the situation, it is thought that we start to forget much of what we've heard or seen within just 20 minutes of the event. This is why police seek to obtain witness accounts of a crime as quickly as possible after an event and why lawyers recommend that drivers involved in auto accidents write down what happened while they're still on the scene. This makes intuitive sense; it might take a moment or two to recall what you had for breakfast this morning, but it would probably take serious thought to remember what you had for lunch a week ago Tuesday. We forget lots of details along the way.

A good program for preserving institutional memory recognizes that an institution has a combination of memories it needs to preserve:

• event memories, which are things like the construction of new facilities or the arrival of new employees
• process memories, which note how things are done in order to save time and ensure their reliable repetition in the future
• decision memories, which explain how the institution chose one path or policy or course of action over another

Event memories are obviously affected by the lapses in eyewitnesses' memories, and it's easy to understand how these can be aided by yearbook-type documentation. Take a photograph of a groundbreaking ceremony, and you'll have a much better idea of who was at the groundbreaking than if you didn't.

But process memories are also deeply affected by eyewitness accounts. Three different people could easily see the same oil change or seal change in a piece of equipment and come away with three entirely different accounts of how to get the same job done. This is why technical writing is such a strange field; everyone knows how frustrating it is to try to use a badly-written operations manual, but nobody seems to be very enthusiastic about writing a better one. Yet doing so is more important than we usually realize -- until something goes wrong and we try to re-establish what we did in the first place. Just try to re-create the steps involved in wiring a home entertainment center, much less tearing down and rebuilding a complex piece of equipment, like a motor or a pump, from scratch.

But even when event and process memories are valued and documented well, it's likely that decision memories are almost entirely neglected. It's easy to understand how it happens -- we don't like to get buyer's remorse, so we convince ourselves that any decision we've reached was the best and we forget about the alternatives. But it's important to document the process of reaching the decision anyway, since those paths not taken may end up being surprisingly useful down the road.

If a capital project has an expected service life of 20 to 30 years, it's entirely possible that people working in an institution in their 20s will be middle or upper-level managers in the same institution by the time the project has to be replaced or upgraded. Unless someone documents the process by which the original decision was made, including notes on the alternatives not taken, the 50-year-old manager who's been with the institution all along will usually be guided more by 25 years of habits and built-in bias than by a fresh look at the available alternatives. And the situation is likely to be even worse if the 50-year-old manager making the decisions came into the institution recently and doesn't even have a memory for when the original project was completed in the first place.

We can always hope that people within an institution will pick up on these memories and this understanding of how things work just by remembering on their own. But just like it's impossible for a student to learn by sleeping on a stack of textbooks and hoping the words work their way into the brain by osmosis, it's impossible to rely on the effectiveness of the human memory without putting pen to paper. As the pen company quotes Confucius as saying, "The weakest ink is mightier than the strongest memory."

A column in the August 2009 issue of Control Global magazine discusses the loss of what the authors estimate as 25% of the nation's industrial-process expertise. It's clearly a rough estimate -- after all, nobody really knows how much expertise is attributable to each individual in an organization. But it's not hard to imagine that the 25% figure is plausible. The stock-market decline of the last year revealed that many Baby Boomers have too little in savings to retire comfortably, and may have to continue working beyond the standard retirement age of 65. But it can't be overlooked that somewhere between 15% and 20% of the working population is over the age of 55. As that working population retires, whether voluntarily or due to health and other factors, they are going to take their knowledge out of the institutions where they work -- unless their organizations make a deliberate effort to capture that knowledge.

Among the "unappreciated skills" cited by the authors that could be used to help enhance an institution's ability to retain that knowledge are two simple but likely under-utilized habits:

• Managers need to give their employees a sense of value and appreciation. People who feel under-appreciated are more likely to take exit opportunities like retirement or other job offers than they would if they were deliberately praised for their contributions.

• People who ask obscure questions and cite potential difficulties in advance ought to be encouraged to do so, rather than discouraged from making life difficult for the rest of the team. The ability to make connections between parts of a system that might otherwise go unnoticed is a characteristic of an expert, and that ability should be openly encouraged. Not only does it take advantage of that expertise when it's available; it also builds a culture of encouraging younger workers to refine themselves and their skills so that they can become experts as well.

Knowledge transfer isn't a simple process; witness the day-to-day struggle of any good teacher in America. But especially as the Baby Boom generation reaches retirement age and largely departs from the American workforce, the need to transfer their knowledge to the people left behind is as urgent as the need to provide the youngest workers with the basic skills to enter the labor force.

A common American experience is the publication of the annual high-school yearbook, complete with photos of students and documentation of the football team's victories and defeats. But yearbooks also serve a role which may go unrecognized: They preserve the "rules" surrounding a school's traditions.

Because most students spend no more than four years in high school, they barely reach the stage where they've learned a set of traditions before they graduate and move on. Four years may seem like an eternity to the student, but in a world where "The Simpsons" has been on the air for 20 seasons, it's comparably just the blink of an eye.

By depositing their memories in a yearbook, though, students contribute to the record and evolution of their school's institutional memory. The human drive to do this may be stronger than we realize: Consider the Wikipedia entry for the BBC's nightly shipping forecast or the website of the Eastern Airlines Retirees Association and its related pilots' association. These websites contain lots of ephemera and cultural memories about institutions that people have, for whatever reason, cared about. But Wikipedia shouldn't be the only place where these memories are recorded.

An institution that stays around for the long term -- like a high school -- has to note what it has done on a regular basis, and an annual cycle makes as much sense as any other. This is not to say that a company or a municipality or a school must publish a yearbook. But the basic principle of appointing an historian and charging that historian with documenting some of the important changes and some of the ephemera of that institution on a regular (presumably annual) basis makes it at least somewhat likely that the culture of the organization will be preserved, even as individuals come and go.

If the culture is healthy and worthwhile, then a yearbook (or something similar) can serve to remind the people involved about what works. If the institution is slipping, then an annual review should help to reveal that as well. An institution has to be aware of and sensitive to its signs of vitality just like a person has to be similarly aware of his or her own physical well-being. Ignoring a serious and recurring pain doesn't make the diagnosis of the problem any better -- it just means that the patient has ignored a symptom and likely made it more difficult for the doctor to help.

An institution could certainly combine its cultural "yearbook" with its annual report, although the two serve unique functions. The annual report documents measurable changes, while the yearbook is mainly a cultural record. But both are valuable -- perhaps it's better said necessary -- to a strong institutional memory.

The Internet offers lots of tools for helping to record these cultural records, though it should not be relied upon unless the institution is also committed to building and maintaining (indefinitely) its own website. Wiki-type websites can be easy to establish, and they are helpful when people are geographically dispersed -- hence, we can use a site like Wikipedia to find out how "You Can Call Me Al" has been used by college marching bands from Washington to Florida -- but when people are closer together, it's most likely better to designate an individual historian (or a small committee of historians) with a designated role.

Every institution has some sort of historical archive, whether by accident or by design. Either there's lots of detritus left around in corners of closets and tucked in distant filing cabinets, or a deliberate effort has been made to record, store, and preserve historical information for posterity. The trouble is that artifacts of every kind, from papers to material samples, are subject to the ravages of time. Thus, for archives to be of any use, someone must have the responsibility to ensure that those archives are safely kept.

A 2008 fire at the Universal Studios video vault was thought to have destroyed 40,000 to 50,000 old movies and television episodes. The company claimed it had backup copies of all of the materials -- and rival studios like Sony claim to have original copies of all of their archives, along with two sets of duplicates, all stored in different parts of the country. That sort of preservation security may sound like overkill, but to a movie studio, those original and duplicate recordings are likely the most valuable things they own.

Institutions looking to protect themselves from the effects of similar calamities ought to assess their own archival policies to ensure that in case of disaster, they have some foundation upon which to rebuild. Proper archives would include both historical and contemporary materials for safe keeping. Those materials ought to include at least the following:

• employee records
• warranty information
• purchase records for major equipment
• duplicates of major contracts
• records of historical significance

Safe keeping of those archives can run from the elaborate -- like the policies followed by the major production studios -- to the easy and efficient. The cost is actually quite minimal to scan an organization's most important archives to digital formats (like PDF), record them to a durable physical format (like a DVD or a portable hard drive), and store them someplace sufficiently far away that they will be protected from harm in case of natural disaster.

Disasters can happen; the city of Parkersburg, Iowa, lost everything in its city hall when an F5 tornado destroyed the building in 2008. All of the city's records went with the building. Protecting the most vital information in a way that can be easily recovered is essential. A simple method may be to ship digital records on DVDs or other memory media to an out-of-town post office box, preferably in some location 100 to 200 miles away from the main site. That helps to ensure that they are geographically isolated from any large-scale disaster close to home, but close enough that they can be retrieved within a day's drive if necessary.

Many people are familiar with annual reports issued by large companies like General Electric. They tend to be a combination of bland fiscal reporting (mandated by the SEC) and cheerleading by management at the company issuing the report. But annual reports have a significant role to play in helping to preserve institutional memory.

One might think that an annual report is too much work for a small organization. But there's no reason why an institution as small as a single person couldn't find value in producing an annual report. Nobody seems to teach a course in "How to write an annual report," but a  simple set of questions provides the outline an organization needs to write its own annual report:

• Who worked here over the course of the year? Who was hired? Who left?
• Who received special training or education?
  
• What equipment or service contracts did you add? Why?
  
• What equipment or service contracts did you remove or cancel? Why?
• What equipment or systems were repaired, changed, or overhauled?
• How much work did you do?
• What circumstances affected the volume of work that you did?
  
• What customers did you serve? Which were new? Which were repeat? Which customers were lost?
  

Every institution needs to assess its own circumstances to determine which metrics are appropriate -- but every organization has metrics from which it could draw. In the case of a municipal water system, for instance, the organization ought to record how much water it pumped and treated, what weather conditions affected its pumping requirements, and the size of its geographic service territory.

A charitable organization might not have "customers" in the business sense, but it has members, donors, and charitable recipients. Knowing the ebbs and flows in those interactions -- and why they occurred -- can tell future leaders what worked in the organization and what did not. Expertise isn't driven by innate genius but rather by deliberate practice and analysis of past successes and failures. The world's most-renowned investor writes legendary annual reports reporting as often on his failures as on his successes. But had these lessons gone unwritten, their long-term value would be anything but obvious.

An organization's annual report doesn't have to be long or complex -- even a single page will do. But unless year-to-year changes are documented and kept somewhere everyone can find them and refer to them, an organization easily risks losing the basic understanding of its own historical performance that it needs to remain viable.

• Sample annual report

People know that experience matters, but sometimes we understand neither how much it matters, nor how to measure it. It is said, for instance, that true mastery of a subject takes a lot more than just time: It requires deliberate practice and attention to improvement. Even those we think of as prodigies often required years of focus and practice to achieve mastery in their fields.

An excellent discussion of the very notions of "expertise" and "mastery" from Legg Mason Capital Management suggests that the differences between experts and run-of-the-mill "old hands" at a task are many:

• Experts are practiced in the art of pattern recognition within their fields. They can see when patterns have emerged that would otherwise be missed by those who have less experience or who have spent less time deliberately learning the patterns of their work.

• Experts look beyond the superficial to recognize deeper problems. While it might be easy to diagnose a problem, it may be much harder to see the cause behind the cause; that is, to see what circumstances are causing a particular problem to emerge. Second- and third-order causes are difficult to recognize if all of a person's cognitive attention is devoted to figuring out the symptoms.

• Experts are especially good at knowing the limits of their own competence. Warren Buffett may be regarded as the world's greatest living investor, but he frequently refers in his annual letters to the importance of staying within his "circle of competence". Experts know where their circles of competence end -- and they regularly test themselves to learn from their failures.
  

What's difficult to tell is just how easy (or hard) it is to transfer expertise from one individual to another. If all expertise were merely a matter of apprenticeship, then anyone could become an expert simply by spending a sufficient amount of time with a skilled professional. That is clearly not enough -- but it does point us in the right direction.

The more expertise we can find "bottled up" inside an organization, the more important it becomes to find younger individuals -- perhaps even a full generation younger than the experts -- and encourage them to practice deliberate expertise-building in their daily work. From studies of expertise, it appears that a deliberate emphasis on pattern recognition and self-testing for success and failure may be the best way to develop the successor generations of experts.

Perhaps a good place to look for guidance about developing experts is the demolition industry -- it's at least as much art as science, and there's a colossal difference between a demolition team that knows what it's doing and one that doesn't. Because it might be funny to watch a building go rolling down a street instead of imploding, but that certainly wasn't the safest outcome. Expertise would've made a lot of difference to the outcome.

Peter Gabriel -- yes, the same guy who formed Genesis before it became a vehicle for Phil Collins -- is backing a social-networking website for the dead. Called TheMemory.com, it doesn't contain anything now, or at least not anything accessible from the front page. But the intention is reportedly to provide a wiki-style site for documenting and sharing memories of the dead. This isn't entirely new as ideas go -- Facebook memorials to the dead have already become commonplace, if not mainstream.

The concept, though, is a reminder that we often don't know what we have until it's gone. Memorial tributes to the deceased, though, are only a symptom of a more generalized condition that we aren't very good at preserving and documenting what we know. The purpose of a program for preserving institutional memory is to deliberately discover and record the knowledge of an organization before it's gone.

It's not enough, though, to simply gather data and allow it to sit in a pile. Google's Book Search service is scanning and digitizing huge numbers of old books and magazines, but in and of itself, a big pile of data isn't especially useful. The value in that data emerges when people are able to draw new connections from the data and record those observations as useful narratives. For instance, one can use the book-search tool to find that Isaac Asimov wrote a series of articles under the title "Futureworld" for the house magazine of the Boy Scouts of America in the early 1990s. Those individual articles don't stand for much, but taken as a whole, they paint a picture of how the future looked from the perspective of 1990 and 1991. Someone has to recognize, though, that a pattern exists before it can be usefully recorded.

According to an article in the July 6th edition of the Air Force Times (available only behind a subscription firewall), the US Air Force has been fighting a long decline in the popularity of its on-base membership clubs for officers and enlisted airmen. The story notes a variety of factors leading to the decline, including fallout from the 1991 Tailhook scandal and an armed forces effort to de-emphasize alcohol consumption. But it also quotes a retired lieutenant colonel who says, "You don't see as many members of leadership inside the clubs anymore. That hurts. Before, you'd see the general heading to the club so you'd tell your wife: 'We're going to the club. We'll have a good time.' It was important."

Base clubs encouraged servicemembers to talk with one another, encouraging the cross-pollination of ideas among people who might branch out more in a social environment than they would normally do in the formal on-the-clock environment. It's no different for a civilian institution: It's easy for people to pigeonhole themselves into doing one kind of job, and talking only to the people who do related tasks. But too much individual myopia can lead the entire organization to forget what it does well and to fail to develop new innovations that would make it work better. Like Les Nessman and his imaginary walls of masking tape, people often voluntarily restrict their interactions at work with people who don't seem necessary to their tasks.

As the retired colonel notes, though, leaders understand that talking across those masking-tape lines makes the organization better and more resilient. They have to set the tone, though, whether it's going to the base club or hosting a company holiday party or just inviting employees out to lunch. Every institution needs conversation across those imaginary lines in order to preserve what it does well and to find new ways of doing things better.

When people hear an unfamiliar phrase like "institutional memory", they may find it confusing. We can define institutional memory as how an organization preserves and enhances the knowledge of its people, but that's not necessarily much more helpful. A better way is to illustrate with the quintessential American game of baseball.

Most Americans grow up to have at least a passable knowledge of the game of baseball, many times without even knowing how they learned it in the first place. You may not play in a rec league or manage a fantasy team, but if you take your average American to a ballpark, they'll have a basic understanding of the purpose of hitting, fielding, pitching, and the like. That's because baseball is around us -- even if only in the background -- from spring training in late February all the way through the World Series at the end of October. Minor-league baseball teams are everywhere, and tickets are cheap enough that it seems almost impossible to escape an American childhood without having spent at least a couple of days at the ballpark.

But if you ask someone from a non-baseball-playing country, the game will probably seem like a mystery. The world plays and watches soccer and is even giving baseball the boot from the Olympic Games. So how do Americans learn a complicated game so well that we use its jargon as shorthand for everything from dating ("We made it to second base") to politics (where candidates are "on deck" for meetings and rallies)?

It happens because Americans have an institutional memory for baseball. But it's so ingrained in our habits that we generally don't even notice it. But by taking a step back from a game we know well, we can learn how to apply the lessons of baseball's institutional memory to the other things we do:

• Knowing the rules. Baseball has a long list of formal rules, many of which are well-known (three outs retire the side) and some of which are a little more complicated (like the infield fly rule). But all of the rules are written down. They have been debated and revised many times over, but the entire code of baseball has been written down. Similarly, any institution that wants to preserve its institutional memory has to write its rules on paper. There's no way to avoid it. If baseball's rules were left solely to the memories of the people playing it, the game wouldn't stand a chance. Neither does any other institution hoping to survive for the long term without committing every rule, policy, and procedure to paper somewhere. Nobody's saying that you have to lug out a copy of the rule book for every pick-up game of baseball in the park any more than you have to carry an operations manual around for day-to-day work. But if it isn't written down, it simply cannot be expected to be retained.

• Testing and measuring with a purpose. Baseball is a beautiful game in part because it has a colossal history of statistics. Teams play 162 games a year in the major leagues, plus spring training and post-season play. Every pitch and every play have been recorded. This mammoth warehouse of data has attracted economists and statisticians to baseball like no other sport. There's even a society dedicated to baseball research. The result is that statistics and measurements aren't feared; they're loved, even by non-sabermetricians. There are, of course, hundreds of decisions made in every game that come from gut instinct or trained reaction. But every manager knows his batting averages, ERAs, and pitches thrown, and uses those to his advantage. One of the great achievements in baseball is the no-no (no runs, no hits) -- a statistical measurement of what didn't happen, as much as of what did. Institutions outside of baseball should be similarly enthused about measuring, recording, and analyzing their statistics. How long has equipment worked without failure? How much time has been spent on maintenance? How much fuel has been used? These sorts of measurements should be recorded and reviewed. There's always going to be room for instinct and intuition, but those are for individuals. The institution needs a form of memory for what works and what doesn't. By measuring and analyzing data, the institution may be able to "learn" to correlate run-time measurements with that moment when Bob the maintenance guy thinks it's time for bearing replacement on motor #3. Some day, Bob won't be there anymore. But the records should be.

• Marking the field. When players are in the middle of a game of baseball, they don't have to guess whether a hit has gone fair or foul. The foul line is well-marked and extended vertically with the foul pole. Nor does anyone have to guess how far a home run ball has to travel to be out of the park. The outfield distance markers tell exactly how far. Much of what a player needs to know to make the game happen is stored outside his brain: It's recorded in the physical environment around him. We encounter this in our daily lives, too, particularly when we drive: Speed limit signs, painted crosswalks, and passing/no-passing zone centerline markers store and deliver necessary information at the moment we need that information, instead of forcing us to memorize every bit of knowledge that we might need. Building knowledge into the environment is often easier than storing it inside people's heads. And doing so makes the knowledge institutional, rather than something that stays (and leaves) with the people who have it. A good system for institutional memory seeks out ways to store knowledge inside the physical environment. A simple but extremely valuable example is the lock-out, tag-out system. A simple scissors clamp with room for multiple individual padlocks can form an extremely simple but reliable means of ensuring that the safety of individual workers is built into the environment itself. Instead of relying upon human memory or someone calling out to get an "all clear" before starting machinery, a lockout mechanism builds in a simple but reliable means of ensuring that nobody is accidentally overlooked. Building knowledge into the environment frees up people's mental capacity for bigger things.

• Keeping a box score. Wrigley Field has one of the nation's most recognizable icons: The big green manually-operated scoreboard. The line score for every major-league game appears for the entire audience to see, just as it has for more than 70 years. And on the flagpole atop the scoreboard, a simple "W" flag is run up after a win, originally to tell passing train passengers of a home victory (an "L" flag tells the obvious story to the opposite). Line scores (the boxed parts of box scores) are clear, simple, distilled depictions of the most important factors in a game: Runs scored per inning, total hits, and total errors. A quick glance at the linescore tells the reader everything essential in a single look. Institutions need their own forms of box scores, as well. Institutions need simple "box scores" and "line scores" for their essential performance data. In addition to telling how well things are performing, a routine built around checking the essential details of the institution's performance on a daily basis helps ensure that the most important things are never missed. A baseball line score with a blank space in the top of the seventh inning would immediately signal something wrong. Similarly, an institution needs to know what parts of its performance are so important that any hole in the data stands out immediately upon even a cursory glance.

If institutional memory is a matter of preserving the institutions that are important to us, then it can hardly hurt to take some lessons from one of the longest-standing and most popular institutions in American culture.