Workflow Institute

...promoting the understanding of real-time learning in the workplace.

About Us | Blog | Home | Newsletter | Reports | Contact Us

Thursday, October 20, 2005

Workflow 2.0

SQL Designer

A full end-to-end "thinkflow" from visual representation to machine code using standardized tools and conventions. This is a step toward wyswyg data structures only now we can also imagine applying the wiki way to data structures and let end-users edit the data structures collaboratively, in real time.

Monday, August 29, 2005

Volume 44 Number 8 September 2005

In Her Own Words: Gloria Gery on Performance

by Tony O'Driscoll and Jay Cross

Fortunate are we who have been inspired by a true visionary. Gloria Gery profoundly shaped the beliefs and work practice us both. As Gloria moves on to developing schools in Nepal and tending failure-to-thrive babies in Romania, we want to acknowledge her work and share her a few of her insights.

Gloria has a knack for diving into a complicated performance issues only to point out what should have been obvious to the rest of us with concise, provocative, and often humorous language. Whenever we heard Gloria speak over the years, we've taken notes, and those notes are the source of the quotations that follow.

Our First Exposure

Tony: I ran across Gloria's book, Electronic Performance Support Systems in early 1994. The first fifty-one pages obliterated all of my paradigms regarding the role of training in organizations. My synapses were rewired and my mental model of learning and performance was forever altered. Throughout my career, Gloria's insights on performance-centered design and electronic support have continued to be invaluable. Had I not been exposed to her insights, I would not have had as much success in helping organizations perform more effectively.

Jay: The first time I heard Gloria's name was a dozen years ago when my company's chairman showed me a copy of Electronic Performance Support Systems and announced that EPSS spelled the death of the training industry as we knew it. Of course, that didn't happen. The ideas were right but ahead of their time. Now, at long last, technology is catching up with Gloria's vision. Her concept of intrinsic EPSS was the forerunner of Workflow Learning, and I was delighted when Gloria accepted our nomination to become the first fellow of the Workflow Institute. The first time I heard Gloria speak, seven years ago, she provided the mantra of my efforts, "Training will either be strategic or it will be marginalized."

Now it's time to hear from Gloria, in her own words. Our comments are italicized, the rest is pure Gloria.Systems Design, Training and Performance Support

In her early days at Aetna, Gloria saw workers struggling with arcane, data-centric mainframe systems. The default solution to their frustration was training and documentation. Training Band-Aids designed to camouflage poor interface design. Ironically, the training often cost a lot more than designing the application for performance in the first place.

Most of our existing systems were designed to function in a paradigm of scarcity where each organization unit developed process and applications based on its own history. This parochial approach to work system design has yielded an increasingly disjointed and unintuitive work context for the employee.

Most of our training is compensatory for bad system design and help desks are the balloon payment on poor system design. If we have to teach people how to use a system, it wasn't designed right in the first place. Why do we have training that teaches useless jargon? Why should we have to live with error messages like 'File sharing illegal error?' Look at the evolution of a program like TurboTax. Simplify, simplify.

Learning must be reconceived to influence the primary purpose of organization: to perform effectively and efficiently. We must give up the idea that competence must exist within the person and expand our view that whenever possible it should be built into the situation. What workers need to do their jobs – information, rules, and knowledge – is often spread all over the place. Good design puts these things within easy reach and shows how to use them to optimize performance.

The emergence of a new discipline such as electronic performance support often starts when a few people are frustrated with the mismatch between their needs and traditional approaches to filling them. The purpose of performance support is to help people do what they need to get done, we need to provide whatever is necessary to generate performance and learning at the moment of need.

We don't need new technology, we just need new thinking. We must fuse learning and doing to enable immediate performance with minimal external support.

On Getting to the Performance Zone

One of Gloria's key concepts is the performance zone.

The performance zone is the metaphorical area in which things come together. It is the place where people get it, where the right things happen, where the employee's response exactly matches the requirements of the situation.

In any learning experience, there is always that moment where you 'get it.' How do we accelerate people's arrival at that moment? There are two contexts for doing this: in courses, or while working. Courses lack authenticity as they are separated from the work context. In too many organizations, users are bouncing between multiple systems to get one task done. How can we configure the interface layer to structure the processes and provide in-context learning because THAT is the teachable moment we are always looking for?

The goal of performance-centered design is to institutionalize best practice on an ongoing basis, all of the time, by the least capable of performers: to enable people who don't know what they are doing to function it as if they did.

First Know the Work

About 80 percent of what people learn to perform effectively happens on the job and yet we continue to dismiss it as informal learning.

People don't deal in subjects, they deal in work. The unifying schema or context for performance-centered design is work.

We must learn to look at the reality of people trying to get through the day. We must reflect deeply on the way work presents itself to the user and build our systems on the metaphors that are connected to the work context itself. The context is the workflow, and the content is what the user needs to perform work within that context.

Today our analytical approaches yield a sterilized view of work, not a real one. We have to understand the work that people do. Most of all, we have to be able to sit in the learner's chair, to find out how the work comes at them. We need to understand what really goes on.

We need to put the real truth into our training. Courses are necessary but not sufficient. We must have a strategy. Architecture is a part of it. Courses are a part of it. But we must understand people, how they learn, how they collaborate, how inquiry teaches, how we learn from observing models.

Performance Support focuses on work itself while training focuses on the learning required to do the work. Integrating resources in the workplace is inevitable, and the need is urgent. Filtering resources so people get the tools and resources they need while actively working is the goal. Work process and roles are the primary filters. The mechanisms vary: portals, performance-centered workflow interfaces, enterprise applications, integration projects, etc, but what's important is that performer be able to name that tune in one note, to perform in exemplary fashion.

The common thread for the learning and performance support communities is this: How do we get people what they need at the moment of need, and what form should it be in?

Learning's New Role in Enabling Performance

As learning and performance come together to address the pressing issues of the enterprise, we must challenge our conventional wisdom about how we ply our trade.

We conceive of learning as an event in which we fill people up in advance with enough information to survive on the job. Instead we must emphasize learning as an outcome of performance, not a precondition to it, and we must strive to limit the amount of learning as a precondition to doing.

To do so will require that we act not on what we know, but on what is known. We must avoid defining the performance problem too narrowly to tackle what we already know how to do. We should focus on how we design a job for day one performance, not how we leverage technology to automate training

In our pursuit of solutions we have assumed that our future should be an extension of our past. What's wrong with this scenario is that we are applying radically different technological alternatives to old frameworks without reexamining their underlying assumptions and structures.

If the effort to learn is greater than the time available at the moment of need, you will lose the employee. Instead of making an effort to learn, they will make it up.

We need to leverage technology to enable new learning structures, not automate training.

We should not default to prior mental models, but instead give up on the viability of the old point of view. The goal of establishing day one performance is not hard to do it is hard to get done. It will live or die on the political issues within the organization.

Workflow Learning

Many people have equated EPSS with Workflow Learning. While they are certainly kin, they are not twins.

How the context has changed… is changing… will either render us irrelevant or make us more critical. How can we proceed to have more leverage in what we do? Workflow is one way for us to better integrate what we do with people's lives. The computer-mediated context IS the workflow context. People are willing to accept less at the moment of need if it is focused and relevant.

One of the questions I hear is, 'How is Workflow Based learning different from performance support?' Well, this is performance support on steroids – magnified, with a much higher impact. The workflow is the context, the magic filter through which we will be able to filter content, against which we have to compare default tactics. There will always be instructor-led training, but there will be far less of it than the workplace learning resources.

Here's a definition: Workflow is a sequence of activities that a person has to do to achieve defined desirable goals and results specific to the condition. Deliverables, solutions, decisions… Filters are needed to screen out the irrelevant and bring to the fore the things that are relevant. The workflow is the best default filter for all data. A fusion of learning and doing is on the way

JAY: A little while ago, I blogged that humankind is awakening to the realization that everything's connected. The point of learning is to prosper within our chosen communities, to optimize the quality of one's connections to one's networks. However, many people have failed to change the default settings their personal firewalls came with, even though their factory-installed settings haven't been upgraded since 1 million B.C.

Gloria thoughtfully replied, "Almost worse than the default settings that are millions of years old are the cultural, political, ethnic and religious settings we were given in our early lives. They, of course, reflect the biases of prior generations and, in my experience, no longer fit in a globalized world. They limit us from more than learning. Rather, they limit us as people interacting as humans with other people. Our networks must go way beyond the filters that sift out important other people -- or have us judge them by trivial attributes."

TONY: As we were chatting at the Workflow Symposium, Gloria commented she really believed that new technologies such as second generation portals and business process modeling finally provided us with the ability to enable the integrated performance at the workflow layer as she had had originally envisioned it more than 15 years ago.

Thanks to technology, the promise of Gloria's performance centered vision moves ever closer to becoming reality. But the change management issues are where performance-centered design will live or die. My own goal is work tirelessly on these issues to make Gloria Gery's performance-centered vision the status-quo in creating workware for the On-Demand Enterprise.

We sincerely hope that Gloria inspires you as she has us.

To sponsor a child in Romania, contact Global Volunteers at

Friday, August 12, 2005

Great Wave off San Francisco

DSC00400This painting appears on the wall at the garage under Japantown. Coincidentally, Gary Dickelman will soon be touring Japan and India proselytizing workflow software.

Friday, August 05, 2005

Workflow Learning Day at Geo. Mason University

Want a peak at the future? Here's my presentation at George Mason University a couple of months back (34 minutes). I'm the opening act for lots of interesting speakers. A panel session with Ben Watson, Harvey Singh, Duane Degler, Gary Dickelman, and IBM's Michael Littlejohn discusses what comes next. (46 minutes).

Michael Littlejohn discusses the future of learning. (57 minutes). Gary Dickelman and Harvey Singh show examples of workflow learning (58 minutes). Ben Watson, Duane Degler, and I factor collaboration and meaning into the workflow equation (52 minutes).

Godfrey Parkin was in the audience and had this to say on his blog, Parkin's Lot:
I have just spent a couple of days at a small highly-focused symposium titled “Innovations in E-learning.” It was put together by the US Naval Education and Training Command and the Defense Acquisition University (DAU), who have among the best and brightest training minds that the American taxpayer’s money can buy. They are not short of budget, manpower, or technology, and they get to mess with lots of experimental stuff. I decided to participate because the future of learning matters to me, and because a couple of my virtual colleagues were pretty much dominating the presentations in one stream.

For several years now, Training Departments have been transfixed by the evolving internet in the same way that dinosaurs were probably awe-struck by the approaching comet. So what does the future hold? I’m happy to report that learning will thrive, but trainers will have to merge back into operational roles. Oh, and Training Departments are dead, at least as we know them. As are Learning Management Systems and any other relics of centralized distribution of learning. Learning that is informal, collaborative, contextual, real-time, and peer-generated, will be the mode of tomorrow.

I recommend reading all of Godfrey's post, including the numerous comments from others.

Wednesday, July 13, 2005

New Workflow learning paper

Workflow and web services is a new paper by Scott Wilson at CETIS. The paper does a great job of describing the state-of-the-art and assessing the requirements for workflow in education. The strictly defined Web Services of yesterday have morphed into more flexible lower-case web services. Wilson seems to feel that workflow learning, especially with multiple feeds, is quite primitive in the workplace and holds no interest for academic institutions. One thing I don't understand is Scott's statement that one of the criticisms of workflow learning is that discrete training activities create problems of transitioning experiences successfully from learning to actual tasks (the “forgetting problem”). Compared to what? Merging work and learning largely does away with the forgetting problem.


Monday, June 20, 2005

Old wine in new bottles?

Gary Dickelman was the only speaker to appear both days at the Innovations in eLearning Symposium earlier this month. Gary's a fellow of the Workflow Institute; he's also an adjunct faculty member of George Mason University. His latest newsletter from EPSScentral begins with the answer to a perennial question here at the Institute.

"Workflow Learning: Old Wine, New Bottle?" This question was posed to me by a faculty colleague at the Innovations in eLearning Symposium 2005 on June 7th - an event sponsored by George Mason University (GMU) in partnership with the Defense Acquisition University (DAU) and the Naval Education and Training Command (NETC). "No," I replied. "Workflow Learning is not just repackaged EPSS." On the other hand, if the common understanding of EPSS were performance-centered systems and tools then the answer would have been "yes."

Workflow Learning, as articulated by Jay Cross and documented by the Workflow Institute, is what many of us have referred to for years as process-centric performance support (or simply process support). If workflow represents a current of tasks and activities and interrelationships, then Workflow Learning is how we are enabled when dropped like a cork into the flow that represents our work. We are carried along to receive the data, information, knowledge, tools and context such that tasks are inevitably completed. From another perspective, the things we need to make us smart (competent and productive) come to us at the time of need - whether artifacts or collaboration or handoffs.

So why do we need a new phrase like "Workflow Learning" if the concept is a well understood instance of performance-centered design? Because what has become commonplace "EPSS" is only slightly warmed-over eLearning. If the business or organizational problem at hand is a performance gap, then a learning solution will likely not close it. Yet we continually see learning solutions relabeled "EPSS." Wrong, wrong, wrong.

I am intrigued by the extent to which the phrase Workflow Learning resonates with most people as an instantiation of our notion of process support. "EPSS on steroids?" I guess so. So much so that the Workflow Institute is alive with activity, including inquiries, workshops, assessment, certification programs and more. The best and the brightest who are tackling problems from service oriented architecture to the semantic web are finding themselves at the Workflow Institute participating in panel discussions, writing articles, applying these principles to their work, arguing and conducting research. It has been a long time since a performance-centered idea has generated so much activity from such a bright and eclectic crowd. Great stuff!

P.S. Last time we checked, Moet & Chandon was still putting vintage Dom Perignon champagne in new bottles before inserting the corks.

See EPSScentral for more.

Friday, June 17, 2005

Looking back at the Symposium at GMU

While we await MP3 recordings from the Innovations in eLearning Symposium, you might be interested in Godfrey Parkin's summary on his blog:

Learning innovations
I have just spent a couple of days at a small highly-focused symposium titled "Innovations in E-learning." It was put together by the US Naval Education and Training Command and the Defence Acquisition University (DAU), who have among the best and brightest training minds that the American taxpayers' money can buy. They are not short of budget, manpower, or technology, and they get to mess with lots of experimental stuff. I decided to participate because the future of learning matters to me, and because a couple of my virtual colleagues were pretty much dominating the presentations in one stream.

For several years now, Training Departments have been transfixed by the evolving internet in the same way that dinosaurs were probably awe-struck by the approaching comet. So what does the future hold? I'm happy to report that learning will thrive, but trainers will have to merge back into operational roles. Oh, and Training Departments are dead, at least as we know them. As are Learning Management Systems and any other relics of centralized distribution of learning. Learning that is informal, collaborative, contextual, real-time, and peer-generated, will be the mode of tomorrow.

It seems counter-intuitive that military types whose culture is defined by command and control hierarchies would advocate devolution of learning to the swab on the deck-plates or the grunt in the foxhole, but that was the gist of what was being said. Admittedly, it was not being said by the JAG look-alikes or their entourages, but by the civilian gurus who write their white papers for them. And devolution of learning does not necessarily mean relinquishing control – in fact there are some very scary big-brother systems being deployed that (allegedly) will tell anyone with access pretty much what any individual sailor anywhere in the world had for breakfast last Tuesday and, to five decimal places, what his or her competency rating is on any given skill. It is hard to reconcile what they are saying with what they are doing, until you realize that, because these systems are so vast, they take a long time to build and deploy. So at any point in time the military are rolling out systems and policies that have long since been abandoned for something new – which may not see the light of day for a decade.

I was mainly interested in hearing what folks like Jay Cross, Clark Aldrich, Harvey Singh and Ben Watson had to say about workflow learning, collaboration, and simulations. However, in amongst their sessions was a real eye-opener from a VP at IBM. IBM used to be a blue-suit red-tie operation as monolithic as a bank, but it has been doing a lot of shape-shifting in recent years. These days any organization that is unwilling or unable to do that is unlikely to be around very long. It's Darwinian – those who can adapt most readily are most likely to survive in times of rapid change. IBM's consulting wing, adrenalised a couple of years ago by their acquisition of Price-Waterhouse Coopers consulting, is doing what big consulting firms rarely do – they are advocating unique solutions that they don't already have parked in a truck around the corner.


Sunday, June 12, 2005

Beware of Imitations

It's both exasperating and fun to watch clueless vendors misappropriate the term Workflow Learning. A case in point: ICUS, an "international elearning consultancy" headquartered in Hong Kong. ICUS offers a free seminar entitled Workflow Learning - The Key to System Implementation Success.

One of ICUS's selling points: "Find out what is Prairie Dogging and how has it become the accepted social norm for workflow learning." Ha, ha, ha, ha. Sam Adkins described prairie dogging as a head popping up over the wall of a cubicle to ask a question. This is a worst practice. Half the time, the prairie dogger gets no answer. When an answer is provided, it's often wrong. The workflow solution involves identifying expertise and facilitating collaboration.

This evening I came upon Tactics, purportedly an independent consulting outfit that is wholly Australian owned. Tactics partnered with ICUS in 2001. The partnership extends to copying ICUS's mistakes. Someone at Tactics is not so hot with the space bar, e.g.:
  • what is Prairie Dogging and how has it become the acceptedsocial norm for workflow learning
  • how to use technology and procedure based, role (or task)methodologies to formalise this informal learning environment
  • a practical demonstration of the Smart@ss technology forcreating, distributing monitoring and administrating informal learning
formalise the informal? administrating? And did they really name their software smartass?

We'll soon post mp3 recordings of the recent Symposium at George Mason University.

Tuesday, June 07, 2005

Innovations in eLearning Symposium

You missed a full, fun day on the latest thinking on Workflow Learning from Gary Dickelman, Harvey Singh, Michael Littlejohn, Duane Degler, Ben Watson, and Jay Cross at the Innovations in eLearning Symposium at George Mason University in Fairfax, Virginia today. Transcripts are on the way. In the meantime, here's what Gloria & Bob Gery are up to in Romania.

Friday, June 03, 2005

A Workflow Learning Pattern Language

Here's a preview of my presentation for the Innovations in eLearning Symposium this coming Tuesday in Fairfax, Virginia. This is 34 minutes of a top-of-trees view of workflow learning, a new presentation from the ground up.

If you attend the conference (there are a few empty seats remaining), you'll hear half a dozen industry experts flesh out the pattern language with live examples.

We're devoting Wednesday to hands-on demonstrations to small groups.

Monday, May 30, 2005

Web 2.0

Tim O'Reilly has a great track record catching trends early on. He thinks Web 2.0 inevitable. Join this free Yi-Tan Consortium conference call for an update.

From Jerry Michalski:
Thanks to the work of many organizations, the Net is turning into a platform, not just a place through which we send email and request Web pages. Loosely coupled applications now know how to communicate over the Net, helping companies build flexible solutions from (relatively) interchangeable parts. It's a big architectural step forward, prompting us
to ask:

o What does SOA mean for the IT business? Who wins, who loses?
o What kinds of applications lend themselves to SOA now? Not?
o Where is this trend headed?

Wiki and chat-related things at, including a link to a real-time IRC chat during the call.
Well, you missed it. However, you can listen to the recording on the wiki site.

Saturday, May 28, 2005

Workflow Learning Pattern Language

Prepping a brand new presentation for the DAU conference June 7-8 in Northern Virginia, I began preparing a Pattern Language for workflow learning. It won't be out of the oven for a few days, but if you're curious about Christopher Alexander, the father of all pattern languages, take a look at my visit to his house.