How We Got To Agile

Frederick Winslow Taylor

Frederick Winslow Taylor

1900. Allentown, Pennsylvania. The dawn of a new century. Manufacturing looked nothing like today.

For one thing, there were no managers. There were owners, There were overseers. Foremen. But not pocket-protectored MBAs wandering the shop floor, clipboard in hand. Back then owners were upper class, landowners, holders of property. They didn’t bother hanging out with riff-raff. Sure, if necessary, you could find the owner or one of his lackeys in the main office. An overseer, a shop foreman or two were on the floor keeping the men sober and busy. But aside from production goals and tradition, the workers were left to their work.

And why not? For skilled work, after completing a lengthy apprenticeship, the job was handled by craftsmen, many times from a guild. They made their own decisions about how their work was performed. Simple work? Workers were hired off the street and paid by how much output was desired. No matter the type of work, each job was a special creation, a snowflake, even if it was just something that fitted into a larger whole.

For unskilled work, workers didn’t do anything. Or rather, they did everything, but weren’t particularly good at any one thing or another. Need more output? Threaten more. Promise more. Workers were miserable.

Craftsmen were the paid same amount no matter how much was produced, and centuries of the apprenticeship system taught workers the wisdom that the more they did, the worse things got. The more they did, the more mistakes they would make. The more they did, the more they would stand out. The more they did, the more they would work themselves out of a job.

Let’s say you got a wild hair up your ass and came in one week and killed it. What do you think would happen next week, kid? I’ll tell you what. Next week the bosses would expect you to work at the same rate, that’s what. And they’d expect everybody else to work at that rate too. Do you have any idea what kind of living hell they’d make for us if we were all constantly working at top speed?

Nope, twice the people working at half efficiency was much better for all concerned than all of us working as hard as we can. It was even better for the company, because then there were extra people around to take up the slack when people got sick. The main job here, if you were smart, was convincing management that you were working as hard as humanly possible — while slacking off as much as you could without getting caught. More people, more money, more leisure time, more security, more everything.

It was, after all, us against them. We have to create our own slack. We can’t trust the business to do it for us.

The early 1900s were the sunset of the Victorian Age, and owners really didn’t much care. They were owners. They were obviously better educated and more hard-charging than the average knuckle-dragger on the shop floor. Foremen provided the raw material, said some encouraging words from time-to-time, and mostly just left workers to sort things out the best they could. Fire the slow workers. Make sure nobody was goofing off. The beatings will continue until morale improves. This is the way it is. This is the way it always was.

It was, after all, us against them. We have to use a stern hand on the workers or nothing gets done.

And so what if things ran 10% better this month than last month? The company made money through political connections, through inside contacts, through an old-boy network. Not by making a better widget.

It sounds odd today, but back then nobody actually organized their work in the abstract. Sure, work was orderly, but nobody sat down and looked at how it all fitted together. There was no master plan or detailed architecture at the task level. There was no optimized design. For management, it was beneath them. How would they know how to do things better than the workers? For workers, using rough guesses and rules of thumb was the way of organizing the work handed down from person to person. Your mentor guessed his material needs a certain way. This was the way you guessed your material needs. Your mentor paced his prep work a certain way. This is the way you paced your prep work — if either of you thought about pacing at all. Or prep.

One mentor taught it one way, another a different way. The standard was that there was no standard.

It was a true collection of artisans, and because everybody was an artisan, nobody was. True skills in such an environment were social ones. Get along with each other. Be respected by your peers. To owners, in terms of measured technical expertise, all the workers looked mostly the same. Other than that, it was all reputation.

Taylor fretted. How much activity was being wasted? How much material over-ordered and never used? How many workers were doing jobs they had no ability in? How many men were living lives of pointlessness, either working hard and not smart, or not working at all while nobody was the wiser?

No. No matter what kind of happy face you wanted to put on it, whether you were owners or workers, this arrangement was killing morale, efficiency, and most importantly, led to distrust, resentment, and a bitter us-versus-them attitude for all involved.

And what of the larger picture? What kind of impact did this work structure have on factories? On the nation as a whole? On the families of the people caught up in it?

Just because we had always done it this way didn’t mean it always had to be done this way.

It was unacceptable. Something had to change.

A hundred years later and it had changed, massively, but somehow we ended up just trading one misery for another.

2001. Snowbird, Utah. A small gaggle of software developers, consultants, and process wonks — nerds, really — met to discuss the state of technology development.

It wasn’t that good. Over the past couple of decades, as more and more people entered the workforce as programmers, dozens of different methods for improving productivity had been tried. Books written. Whitepapers prepared. Doctoral theses defended. People were certified. People set up huge organizational systems to manage technology creation. People honed and tweaked. They lived lives of quiet desperation inside huge corporate and governmental machines dedicated to making technology happen.

For skilled work, nobody did anything, or rather, they were busy, but little got accomplished. Workers were miserable.

Early on, people developing technology started with a simple premise: things happen in a certain order. It’s sequential. People need something. They tell us. We write some code. We debug. We give it to them. They are happy. One thing naturally leads to another. Work naturally happens in a clear sequence.

So it looked like developing technology was just a matter of listing out each step and doing each step correctly, right? Pieces of work come in at a high level, then cascade down to product coming out the other end. It was a big waterfall. You got on at the top, you got off at the bottom. This idea that sequencing was the most important part of the structure of technology development and that we should order our work as some version of a large linear sequence was called “waterfall methodologies” or just “waterfall”.

In theory, waterfall was natural and intuitive. Only problem, in practice it didn’t work so well. In practice, there wasn’t always a single path production could take. There were branches, loops. Once product was released, it might enter the system again as people needed to fix problems or add new features. Instead of a simple waterfall with a few stages from “beginning” to “end”, technology development started looking more and more like a flowchart. Then it started looking like a really complicated flowchart.

The idea that technology development is a flowchart with cycles, not a straight line, over it’s entire lifetime is called a Software Development Life Cycle, or SDLC. The idea is that whatever you’re doing in a modern technology development shop, it’s just work inside a little node on a flowchart somewhere called the SDLC. Your node has inputs and outputs. Rules of operation. Quality criteria. You worked the node. Or the node worked you. Depended on how you looked at it.

Lightweight SDLC by deliverable -- documents to be delivered. Each deliverable would have it's own complex flowchart of processes to create it.

Lightweight SDLC by deliverable — documents to be delivered. Each deliverable would have it’s own complex flowchart of processes to create it. Projects started in the upper left and ended in the lower right.

Even at this level of complexity things weren’t complete, though. Or rather, no matter how complicated we made our flowchart, there was always something new that needed to be added, and it was important to add it. After all, companies in this world made money by being as precise and efficient as possible, by making a better widget. Chasing the correct SDLC, tweaking it and making it better, looked a lot like something that over time created diagrams of infinite complexity.

So practitioners doubled-back. Instead of looking at things over a lifespan, how about if we changed focus? Looked at things in small increments? You’d do the same things over and over again, releasing a new little bit of stuff when it was ready. Instead of focusing on the tiny details of what you are doing, couldn’t you just tell me what you can do by next Friday? Maybe technology development wasn’t a flowchart, maybe it was just a string of little nodes all the same, each one representing a unit of time.

Focusing on delivering pieces of a whole in small units of time was called incremental and iterative development, and it’s mostly stuck with us. It helped. But not as much as we hoped.

First, we still had the question of process. What did we do inside of each release, or increment? Nobody knew. How long should an increment be? Should there be a standard increment length? Should we just work until we think we’re ready? Obviously if we wanted any chance of improvement, we had to break down, define, and optimize each piece of our work, right?

People added in details to answer all of these questions, and, just like before, a simple idea of doing small increments over and over again in an iterative fashion became many super-huge systems of how to program, how to report your time, how to create one of the standard project management plans, what sorts of status reports were necessary and when, and which of the 83 standard artifacts you needed to complete for the upcoming release meeting next Thursday afternoon. Yes, the nodes were all the same. They were just little boxes of time. Inside each node, though, we crammed a miniature SDLC. And that SDLC could get just as complicated and unwieldy as the old ones did.

It was the dawn of the Age of Dilbert, and workers led days that were full of meetings, mostly management meetings, mostly asking why things weren’t completed yet. Each and every bit of their time was tracked. Everything had a check and a cross-check. There was a standard structure and system for everything they did and everything was meticulously measured and tracked. We define. We measure. We optimize. Productivity was king.

But productivity did not increase. Instead, it plummeted. There was plenty of activity, mind you, at least on paper, but nothing of any value got released in any reasonable time frame. Developers were miserable. Managers were miserable. Customers were miserable.

Trust was lost between the business and the workers. Promises were never kept. Deadlines always slipped. No matter what kind of happy face you wanted to put on it, whether you were owners or workers, this arrangement was killing morale, efficiency, and most importantly, led to distrust, resentment, and a bitter us-versus-them attitude for all involved. Eventually each side recognized the game. It was the same old one. It was, after all, still us against them.

Looking back, some extremely smart people across the industry had worked at this problem of how to organize technology development work over a period of many decades. All sorts of various systems had been erected, diagrams created, but no matter where we started, no matter the philosophy and rules used, technology development always ended up a complex mess of rules, contracts, meetings, flowcharts, late hours, unhappy spouses, and kids with missing-in-action parents.

It was unacceptable. Something had to change.

Early "modern" production line. It was found that when workers were required to take "tests" during their work, productivity magically improved.

Early “modern” production line. It was found that when workers were required to take “tests” during their work, productivity magically improved.

Back in 1900, Taylor agreed. Something had to change. He looked at the shop floor. The company was spending millions of dollars on labor, and yet nobody was using any kind of system to keep track of how things were done. It beggared belief. Surely we could learn something from technology and science and apply it to the most important thing we do during our day — our work. Surely work had to be something more valuable than simply a force of wills between owners and workers.

What we needed was a scientific method for managing our work.

1. Replace the old rule-of-thumb work methods with methods based on a scientific study of the tasks.

2. Scientifically select, train, and develop each worker rather than passively leaving them to train themselves.

3. Cooperate with the workers to ensure that scientifically developed methods are being followed.

4. Divide work nearly equally between managers and workers, so that the managers apply scientific management principles to planning the work and the workers actually perform the tasks.

In short, we needed to emphasize science, reason, and hands-on management over tradition, old wives’ tales, and class separation. Science was the way forward for all of us. Science was our hope. It was insane not to use science to help us improve the way we created things. In Taylor’s mind, it wasn’t that class distinctions would go away, lord no. You’d never take a lowly worker bee and make much of a thinker out of him. That wasn’t even in the cards. We couldn’t change reality. We couldn’t change attitudes and values, but we could, and should, change our processes and tools; in fact we should break down, define and tweak each and every thing we do during our work day to make sure it makes sense to us and is providing the maximum amount of value.

Looking out over Allentown, Taylor knew our emphasis was correct, we just needed to plan, measure, and change every little bit of our action we took during the day, use the scientific method to meet our goals better.

Which is exactly the opposite conclusion the Agile Consultants reached a century later after living in the world Taylor created.

Looking out over the state of technology industry from Snowbird, no matter where people started, no matter what philosophy and rules they used, technology development always ended up a complex mess of rules, contracts, meetings, flowcharts, and late hours. The business always lost trust in the workers. The workers always lost trust in the business. The customers lost trust in both. Everybody ended up unhappy.

It wasn’t that any of the specific things we did or tools we used were bad, really. In many cases there were really good reasons for folks using them. It just always seemed that somehow things got out of hand, and processes and tools that were supposed to be supporting the work and helping things run smoother instead became prisons that people lived and worked inside of. It wasn’t enjoyable. It didn’t even accomplish the things it was supposed to.

We couldn’t always change our processes and tools. Many times they were the best we had, and we had to use something. But we could, and should, change our attitudes and values.

We needed a change in emphasis. Thus the Agile Manifesto.

The Agile Manifesto is one of those things that started off as a few bullet points in a conference somewhere in the backwoods of industry and somehow gained a life of its own. Sure, there were a lot of other things that came out of the conference — they decided on the buzzword “Agile”, they came up with a set of principles, there were various structured systems of work that spun off. They drank beer. There was a lot of writing and marketing. But nothing with as much staying power as the Manifesto itself.

The reason the Agile Manifesto resonated was because the Manifesto was not just another system of telling folks how to run technology teams. Instead, it was a reaction, a reaction to the hundreds of methods that had already been proposed and tried (and billions wasted) developing technology. It wasn’t a set of instructions. It was a set of values. A polite primordial scream. People who read the manifesto could choose whether to agree with the values or not. If they agreed, they were then free to apply them in any way they desired. They were asked to agree on where our values had gone astray, not about the perfect way to run our next technology development program. There was always plenty of that.

The Agile Manifesto went something like this:

“We are uncovering better ways of developing
software by doing it and helping others do it.
Through this work we have come to value:

Individuals and interactions over processes and tools
Working software over comprehensive documentation
Customer collaboration over contract negotiation
Responding to change over following a plan

That is, while there is value in the items on
the right, we value the items on the left more.”

Might want to read that list again, mentally comparing the items on the left to the items on the right.

Pretty simple stuff, and it might have gone without much notice, but the timing was good. There was a website. People saw it and liked what it said. A movement was born.

Then came the hard part: proving that the principles had practical value and weren’t just more happy-talk or managment-speak, sophistry, which is exactly where Frederick Winslow Taylor found himself back in 1900 when he started spreading the gospel of Scientific Management.

The very first "light paintings" were created by time and motion pioneers interested in defining and optimizing how workers worked

The very first “light paintings” were created by time and motion pioneers interested in defining and optimizing how workers worked

At first they hated his ideas, and some of the ones that hated it the most were the ones you’d least suspect: factory owners.

That’s right, because although we associate management today with close oversight and control over the workers, early 20th century owners already had all the control and oversight they wanted. They could hire and fire at will. They were comfortable. They had tradition.

In many cases they even enjoyed very close, paternalistic relationships with folks on the floor. There was rarely outright hatred. A good worker was loved by his boss and he loved his boss too — all while he slacked off as much as possible. All while the boss planned draconian incentives to keep the workers honest. No hard feelings. It’s just the way things have always been. Upper class cracked the whip. Lower class pulled the oxcart.

Owners hated Taylor’s Scientific Management ideas. Here come these guys, these Management Consultants, carrying this radical departure from tradition. Required a lot of oversight and work that looked silly, and at the end of the day, what do you have? Same old lower class, uneducated riff-raff, god bless ’em, that you had to begin with. You couldn’t make a silk purse out of a sow’s ear. These ideas amounted to nothing much more than an an unnatural mingling. Of ideas. Of classes. Pointless bean counting.

He was the first person to ever use a stopwatch on the factory floor. Who would use a stopwatch for menial tasks? Those were for sporting events. Train schedules. But Taylor would. He would sit and time people as they did their jobs, breaking down each task and cataloging it, watching closely how long it took to accomplish, what the averages looked like, and how much variance there was. Which methods worked better than which others.

Scientific Management: the idea that careful definition, measurement, planning, and architecting could make a huge difference overall.

Taylor watched some men moving pig iron. On average, they moved 12.5 tons per day. If necessary, the company could offer huge sums of money to get them to move more, say 47.5 tons, but experience showed that the only thing that would happen would be that the men would try as hard as they could for a while and then give up, exhausted.

If, on the other hand, a specially trained overseer, a real manager, conducted experiments to determine how much rest was required and at what intervals, then the manager could synchronize the lifting and resting among all the men such that every man could move 47.5 tons per day, every day, all without becoming exhausted. Manager does the thinking. Worker does the working.

Don’t just whip the horse harder and pull on his reins. Take apart everything the horse does, climb inside his brain, and optimize each motion. Think for the horse.

Of course, not every worker would be capable of moving that much pig iron. Perhaps only 1/8 of the general population could. Not all of our skills and weaknesses are the same. Some folks have attributes that make them very good at moving pig iron. Most do not. The company would be doing both the men and society a favor by selecting only those people for which moving pig iron came naturally. This could be accomplished with physical fitness tests. Science helped the managers define the work, plan the work, measure the work, and select the workers for the work.

No detail was too small for Scientific Management to analyze. Taylor studied shovel design. How long does it take to operate a shovel? How does that interval change depending on how much weight the shovel is carrying? What is the optimum weight for an operator to lift and move using a shovel in order to maximize throughput?

Based on time and motion studies, he determined that the optimum weight a worker could lift using a shovel was 21 pounds. Since various materials being shoveled had various densities, obviously shovels should be sized according to the type of material being moved so that they would, at their fullest, weigh 21 pounds. Companies started providing workers with appropriate shovels. Prior to that, workers brought their own shovels to work. Substantial productivity gains were realized — and more importantly, concretely proven — by recording productivity metrics before and after the change. No more running things by anecdote or tradition.

Science rocks.

As this caught on, others joined in, including a husband and wife team that used new and highly-experimental motion picture apparatus to study motions of bricklayers. More dramatic improvements were proven.

As it turns out, Scientific Management was something that any educated person with discipline and a scientific mindset could get in on. Just required a new way of looking at things. Which is exactly the way the Agile Manifesto consultants wanted things a hundred years later. A new way of looking at things. It was just common sense: we should share values. Empathize those. No-brainer.

But it wasn’t a no-brainer. People hated the Agile Manifesto in 2000 just like they hated Scientific Management in 1900. And surprisingly, many of those who hated it were also the ones you’d least suspect: developers living in old oppressive systems.

Why should they like it? They had a complex, unwieldy system, sure, but it was their system. It was something they had spent many years putting together and tweaking to optimum performance. Sure, some of the new guys thought the system sucked, was overly complex, and ate people’s souls, but that’s just because they were new guys. Hippies. Slackers. They didn’t understand the reasons why things had to be the way they were because they weren’t looking at all of the things that could go wrong.

And here come these new schmucks, these…Agile Consultants? What do they offer? A bunch of hot air and loose talk to encourage youngsters and all those already unhappy with the way things are going. These Agile Consultant guys would be better off spending their time and energy explaining why things had to be the way they were instead of building fantasy castles in the air. Perhaps doing some real research on processes and procedures we might have missed. Give us more detail. Provide practical advice instead of regurgitated marketing speak.

And yes, things were taking a lot longer than they used to, but that’s not important. What’s important is that we’re finally beginning to operate this place like an engineering shop instead of a mob. Optimization comes later, after standardization and baselining. Everybody knows that. We’re becoming professionals. We have a system. We make complex things, therefore our system should be naturally complex. We are just now getting around to adopting some intricate yet practical and important processes, and these guys are suggesting things that sound like they were invented last week in a beer hall and could be explained on a page or two in the back of a comic book.

But the really crazy thing? The thing that drove us nuts? Most of the time, the practices developed from the Agile Values? They worked.

The daily stand-up is practiced by hundreds of thousands of technology teams around the world

The daily stand-up is practiced by hundreds of thousands of technology teams around the world

For example, they had this accomplished, committed, needs-help cycle. Once a day, each person would go through a brief communication cycle that covered — and only covered — their accomplished, committed, and needed-help items. The entire team would stand in a circle and toss around a Nerf ball. When you caught the ball, you communicated to the rest of the team — not a manager! — what you accomplished over the last day. You communicated commitments about what you thought you could accomplish over the coming day. Finally, and this was bizarre, you communicated what you were weak or worried about and asked for help from the others. And the team provided it! Without being told! On their own!

The same thing happened at the next level up, the team. The team repeated this cycle every so often, demonstrating what they had accomplished, making a commitment to do more, admitting their weaknesses, then talking about where they needed help. Then management would be directed, by the team, no less, where they should help. By the team!

It was something you could train people on in five minutes. Heck, just reading about it and you’re already trained. Yet every team that honestly gave it a shot found productivity increasing, happiness increasing, participation increasing. Some even called this communication cycle the “five minute miracle.”

Then some of the other Agile guys had this idea of the team being the unit of work delivery. Instead of assigning work to a person, as has been done since the dawn of time, the idea was to let the team as a unit decide, commit, to what it could accomplish. Then the team, as a unit, was responsible for delivering it, not a specific person. All the problems and variances among different people still existed, of course, but it was up to the team to handle them. From the outside there was the team, and the team was all there was. Work went into the team, the team did work, and work came out of the team. What people did inside the team was of no concern to anybody but those in the team. Insanity.

But genius, too. This freed each team to dynamically optimize its work patterns depending on the project, the team members, the type of work, perhaps even the individual work item or task. The company had a trade to make: time and money to the team for delivered product back to the company. The team had the freedom to make the value creation happen as productively as possible inside its commitment time box, which somebody decided to start calling a sprint.

Teams working in this fashion, with no hangups or baggage carried over from previous systems, customizing their optimizations as they went along, were reporting some astounding successes. So much so that most outsiders found it hard to believe. Two, three, ten times faster? Happier workers? Customers singing praises? Technology development, Management Consulting, had never seen that kind of performance improvement. That kind of excitement. Ever.

Then there was the whole Product Owner and Backlogs thing. Since the team was the unit of production, what was going to be the unit of value for the team to produce? There had to be some way of enumerating and measuring the value of the things the team was working on. There had to be some kind of interface between the team and the outside world. After all, how would the team know what to work on? And so the Product Owner was born. The Product Owner made a list of things that had business value. The Product Owner ordered this list from most valuable to least valuable. The Product Owner gave the list to the team, and the Product Owner was always there to answer questions about the list and the items in it.

This list, the list of things that had business value, was called the Backlog. The Backlog was the Product Owner’s, The Product Owner and his list represented things the business was willing to pay for, things that had real value. Not a lot of explanation was forthcoming at first about what went in the backlog, or what exactly the Product Owner did all day, or how the business knew these things were valuable, or what made a good or bad item on the backlog, but teams had a guy responsible for telling them what to do, and a list of stuff to do, and that seemed good enough for most folks.

A Product Owner and a Backlog eliminated long sessions with users who didn’t know what they wanted. Instead, the Product Owner could handle that. It eliminated huge hunks of time where people dickered about which thing was more important than which other thing. The Product Owner could handle that too. Why should the team care? Give them a list, a person they can talk to, and they can make things happen. We solved the problem by assigning it to somebody.

Big swaths of time that used to be spent coordinating and meeting were returned to the developers to do productive work. Tough problems were made simple. Developers made happy.

Now if it only would take off.

Frederick Taylor had the same struggle with Scientific Management Theory, but he didn’t give up. After working at a few places validating his ideas, he retired at age 50 and spent the rest of his life writing books and proving his ideas correct to anybody that would listen. And because people could see the results, he quickly gained a following. More and more managers saw the natural connection between work, management, and science.


Scientific Management went everywhere. Taylor single-handedly invented the job of Management Consultant. Most all of modern management theory is based in one way or another on Scientific Management Theory. Name a famous management consultant; Deming, Drucker, McKinsey; talk to them or read their work, and pretty soon they’re talking Frederick Winslow Taylor. Name a famous management theory: Six Sigma, TQM, MBO, Balanced Scorecard. It all goes back to Scientific Management Theory.

That’s right. Frederick Winslow Taylor is the guy responsible for just about every management fad you’ve ever had to suffer through.

His ideas of time and motion measurement and optimization, along with his use of the scientific method for process improvement, are used everywhere. In fact, it’s impossible to find anywhere they are not used. The entirety of modern life, with all the great increases in technology and efficiency, is built on Scientific Management Theory.

Even facing stiff initial resistance, the logic of Scientific Management quickly outstripped the emotional appeal of not rocking the boat.

Agile was slow going at first, too, but eventually took off for completely the opposite reason. While Taylor supporters had to argue science and logic, Agile adherents could just jump up on a soapbox and hold forth on the cruelty of the existing system. You couldn’t support Taylorism without supporting precise measurement and studied action, but you could support Agile without actually taking a stand about anything at all. You could just take a stand against stuff. It was enough to know that the system was broken. The system was broken! Tear it all down!

Early Agile adopters faced a ton of friction from the Scientific Management crowd, mostly in the upper echelons, but the emotional appeal of Agile Values quickly outstripped the logical appeal of Taylor’s measurement, planning, and optimization, at least for the people doing the work.

And as much as it sounded like another feel-good management fad, applied Agile had the distinct benefit of actually working. But it worked in a probabilistic, not scientific manner. People could take the values that the manifesto described and create various concrete processes to try in their teams. Then they could compare notes. Over time, some processes were mostly found to work well most of the time. Some only rarely. But practitioners could collect these process tricks, or games, like postage stamps, trading with other practitioners and selecting the appropriate ones depending on what kind of work was going on. This wasn’t about the right or perfectly optimized way to do things. This was about the way to do things that worked better than others most of the time. This difference turned out to be critical. Agile, at least the way the community applied it, wasn’t a standard; it was a heuristic.

Hundreds of books sprang up around Agile. Pick an area of technology development: coding, databases, hardware, the Internet. Successful, proven Agile techniques are there and continue to grow there as people organize new processes and structures of work and then place those processes, games, katas, or tricks, inside the larger controlling Agile value system.

Two opposite philosophies. Each deeply involved in all progress and organized activity on the planet. Each looking at the world in a completely different way. Scientific Management, which makes sense to anybody creating and running the machine of business. Agile, which makes sense to anybody making technology solutions happen.

Scientific Management tells us that life is rational, causal, and works according to logic and reason. If we define what we are doing, break it down into smaller pieces, we can optimize those pieces, thereby optimizing the whole. One set of people plans the work and another set of people implements the work. In fact, that this separation between planning and execution, combined with scrupulous measurement and adaptation, is the only sane way to organize our work.

Agile tells us that no matter how we break down and optimize our work, without our values being in the right place we end up creating a living hell for ourselves and others. That progress involving ideas is many times surreptitious, not directly causal. That you can only break down things so far. That trust is more important than measurements.

It’s as if there’s one world full of business people with suits, stopwatches and blueprints, measuring each little thing and determining how a little tweak here or there will change things by a percentage point or two. Another world is full of technologists with flip-flops, games, and Nerf balls, not talking about measurements or machines yet creating new products that change reality for the rest of us. And doing it on a regular basis.

And backlogs? Right at the boundary, in the hand of the Product Owner.

Follow the author on Twitter

February 11, 2014

2 responses to How We Got To Agile

  1. paul said:

    Why no mention of Lean and its history/impact along with TQM and Excellence and alike. Even though it was not widely used in the USA, i would imagine it had impact.
    Agile is very Lean thinking and Scrum has a lot in common with the origins of Kanban.
    Just wondering?

    • admin said:

      Great question, Paul.

      This is part of a book on backlogs. As such I was trying to bring the average reader up to speed in regards to the two main forces at work in technology development today, the production-line mentality and the social-development theory.

      What I found was that the more I tried to explain the background, the bigger the chapter got. So I decided to dump it here, then cut even more for the book.

      Something like this, you know? Could be a book on its own.

      So — you are correct, along with several other commenters. There should have been a lot more mentioned. Sorry about that. Hope the explanation makes sense.

Leave a Reply

Your email address will not be published. Required fields are marked *