Sunday, September 18, 2022

Coach? Consultant? Trainer?

Language is a funny thing.

As philosopher Wittgenstein said "The limits of my language mean the limits of my world."1

Or, to take another angle, as Steve Young put it "Perception is reality" 2

Without wanting to re-iterate my whole earlier post, I would just like to shine a light to the fact that outside the agile realm the coach is much more prevalant in sports than in psychology (e.g. life-coaching).

And a sports coach acts quite differently from a life coach. Can you imagine a group of people who hire a coach because they want to become a soccer team and that coach would start by asking everyone how they think soccer should be played? If the distance of the goals is to their liking and whether a ball would be the best thing to play with?

If you can imagine this scenario, then I guess, it is either with a sarcastic glance at the way many agile coaches work today or you where reminded of some kind of comedy.

Life coaching, solution focused coaching, systemic coaching all have their places – even in soccer coaching – but usually not in the beginning when the players still are unaware of the rules of the game, not well versed in the moves and inexperienced.

And by the way: the oldest mention of a coach in what later came to be the agile realm was from eXtreme Programming (XP). To quote my aforementioned article and paraphrase from eXtreme Programming explained:

“... the [coache’s] job duties are as follows:

  • Be available as a development [programming] partner [...]
  • [make refactoring happen]
  • Help programmers with individual technical skills, like testing, formatting, and refactoring
  • Explain the process to upper-level managers.”

or – on a later page – “Sometimes, however, you must be direct, direct to the point of rudeness. [...] the only cure is plain speaking.” And also “[...]I am always in the position of teaching the skills [...] But once the skills are there my job is mostly reminding the team of the way they said they wanted to act in various situations. The role of the coach diminishes as the team matures.”(p 146)

So maybe – just maybe – it would helpful to be aware whether the team needs a sports coach or a therapeutic coach.

I find that both are appropriate at different points in time, but I have seen a lot of cases recently, where the client was looking for –and needed– a coach akin to the sports-coach metaphor, ended up with a coach conforming to the life-coaching metaphor and everyone just ended up really unhappy.

till next time
  Michael Mahlberg


  1. "Die Grenzen meiner Sprache bedeuten die Grenzen meiner Welt", – Ludwig Wittgenstein, Tractatus logico philosophicus, 5.6, 1922.↩︎

  2. "Perception is reality. If you are perceived to be something, you might as well be it because that's the truth in people's minds." - Steve Young↩︎

Sunday, May 08, 2022

«Creating Feedback Loops» is not about having meetings

In many modern approaches to work, like The Kanban Method, Lean Startup, Agile Software Development, or DevOps, Feedback is an essential part of the approach.

Sometimes the role of feedback is explicit, whereas in some cases it is more of an implicit assumption that is only visible upon deeper inspection.

The Kanban Method has it pointed out explicitly as (currently) the fifth practice (Establish feedback loops) while the DevOps movement has one of its "three ways of DevOps" dedicated to it (The second way: the principles of feedback) which in itself consists of five principles.

“Let’s have more meetings” – a common misconception

Unfortunately, some of the currently popular approaches have introduced the notion that implementing feedback loops implies having some special meetings for feedback.

Feedback like that, could be a daily meeting regarding the current status of the work – especially focusing on problems or things to solve, or an event-based meetings, like post deployment retrospection.

For example, if you look into The Kanban Method, you'll find a whole slew of other meetings to be held at different cadences to foster more feedback in your work.

While these meetings can be very helpful, they are not at all the best way to get real feedaback, really quick.

The problem with meetings as the primary source of feedback

The trouble with feedback that only comes periodically, and is dependent on human interaction, is that most of the time it comes too late.

Consider some feedback loops from outside the work organization world:

  • The speedometer of your car gives you feedback about your current speed – just waiting for the speeding tickets to come in would be way too slow as a feedback loop.
  • Or how about another thing in your car that you get information about: the oil in the motor via the oil warning lamp and the oil dipstick. For certain kinds of information the dipstick, at which we look from time to time gives us enough feedback. For the important short-term feedback that the oil pressure is too low, we need faster feedback. That's why your car comes with an oil pressure warning lamp.

How can we crate feedback loops inherent in the ways we work?

What we actually want when we talk about feedback, is usually a very prompt response from the system we are interacting with. This system can be anything from a technical system through a physical system or a mechanical system to a system consisting of people interacting with one another.

One of the best ways to get early feedback is to actually remove inventory.

You may have heard that removing inventory is a central tenant of all the lean approaches, but when thinking specifically about feedback, removing inventory has the added benefit of making sure that we get our feedback earlier.

So really, what we mean by “creating feedback loops” is finding ways to see the final impact of the things we just did as early as possible instead of waiting for the effects to happen somewhere very far down stream.

till next time
  Michael Mahlberg

Sunday, March 13, 2022

Three strategies to ease the meeting pain

“Since we started the new approach, I hardly ever get any work done, because we have so many meetings.” That is a sentiment, I here quite often when I’m visiting clients who have just started with some new approach. Surprisingly often that is the case if that new approrach is some flavor of “Agile.”

This seems more frequent if the client is a large corporation, but it certainly also happens at startups and SMEs.

And yet, on the other hand it seems to be increasingly hard to get any meetings scheduled. Let’s look at some approaches to make things a bit more manageable again

Once we start to differentiate between meetings that generate work and meetings that get work done it starts to get easier to handle the workload.

As described below, once we start making that distinction we can apply strategies like

  • planning the Work instead of the meetings (allocating time in my calendar for “getting stuff done” – especially helpful when applied –and negotiated– on a team or even multi-team level)
  • conscious capacity allocation (I will have 3,5 hours of working time and 3,5 hours of meeting time each day)
  • Actively keeping buffers open for unexpected, short term interactions (Putting blockers in my calendar that I remove only shortly before they are due)

Now let’s look at these strategies in detail:

Two types of meetings

Some people (maybe many) tend to view all meetings as “a waste of time” and “not real work” – I beg to differ.
I would say that we need to differentiate between meetings that leave us with more work than before and meetings that leave us with less work than before.

Work generating meetings (coordination time)

Some meetings leave us with more work than we had when before we attended the meeting.

  • Planning meetings, where the actual purpose of the meeting is to find or define work that needs to be done.
  • Status meetings, where the original intention is just to ”get in sync” but where it often happens that someone realized: ”oh, and we have to do X”
  • Knowledge sharing meetings, where not everyone affected is invited and thus we need to share the knowledge again.
  • Knowledge building and gathering meetings where the purpose is to better understand something, we didn’t fully understand before – be it a user interview in a product development company, a design session for something be build ourselves, some kind of process improvement meeting, or something else in the same vain.

This list is of course by no means conclusive, but it should give you an idea of the kind of meetings that could be put in that category.

Meetings that get work done (creation time)

On the other hand there are meetings that actually get work done. Especially for work that needs more than one person to complete it.

  • Design Sessions that end with decisions.
  • Pair-Writing an article or a piece of software
  • Co-creating an outline for an offer
  • Co-Creating the calculations for next years budget (if your company still does budgeting the old way)

Try not to mix the two types of meetings. At least not too much. Especially try to make the second kind of meeting really a meeting that gets work done. As in done-done. Make sure that there is no ”X will write this up, and we’ll review it it two days.”

If it’s good enough in the meeting, it’s probably good enough for work.

If we introduce some kind of follow-up work, especially follow up-work that has to be reviewed again, we actually prevent people from using the result of the work we just did in that meeting. Try to make it “good enough for now” and then let’s get on with creating value at other places.

And if it takes too long to create those documents in the Meeting with the tools you have available in the meeting, you probably have some great opportunity to re-think your choice of tools.

With this in mind, let’s look at the three strategies in a bit more detail.

And even though the strategies are persented in a specific order, there is no real ordering between them. Each of them works well on it’s own and you can combine them in any possible way.

Strategy one: Plan the work, not the meetings

Even if you apply only this one strategy it can be a real game changer.
Instead of keeping your agenda open for meetings and then work during the few times where no meeting is scheduled, no meeting needs preparation, and no meeting needs post-processing, switch it around.

Start by filling your schedule with “creation time” – time slots where you intend to do the part of your work that directly creates stuff. When you’re a knowledge worker in the times of a pandemic, this might also include meetings, but those should be only meetings that create tangible results. (This could be a design session with colleagues if you’re in manufacturing, it could be an editing session on a paper if you’re in academia, or maybe a pair- or mob- (ensemble) programming session if you’re in software development. Any meeting that outputs work.)

Only after you filled your schedule with a reasonable amount of time allocated to ”creation time” fit those other things, that I like to call “coordination time”, in some of the remaining spaces on your calendar.

This “coordination time” can include planning, status updates, learning and agreeing upon how you want to do things, understanding the challenge you’re currently working on, and so on. It is basically the coordination you need to efficiently get stuff done in the “creation time.”

Some people tend to call only “creation time” Work and the rest of the time Meetings. However, meetings that neither add value through creation nor through a better understanding of who is doing what when and how, should be eliminated altogether. And maybe replaced by an e-mail or

Especially when we work on process improvements or introduce new approaches we tend to start by planning when the related events (or ceremonies to use an older term ;-) ) should occur to include all the necessary participants.

I suggest to first try to agree upon the times out when all the participants can do their “creation work” and then fit the events and other necessary meetings around that.

Combining this approach with a conscious allocation on capacity makes it even more powerful.

Strategy two: Allocate capacity consciously

Don’t just look at the days of the week as a long stream of hours passing by. Make a conscious decision on how to invest the time beforehand.

If you’re involved with some kind of process framework you probably have some of the time allocation already done for you “daily standups”, “plannings”, “review” and “retrospectives” to name but a few.

But is the rest of the time really uniform? For most of us it isn’t. It consists of periods where I can just chop away at my work, of periods where I need information from other people and of periods where other people need information from me.

Creating even an informal and rough plan of how you intend to allocate your time helps a lot in reasoning about the number of meetings and makes the gut feeling a lot more tangible and negotiable.

Such a rough and informal plan might just look like this:

Allocation per Week (on average)
Process related       4h (8h in total every two weeks)
Creating stuff       20h (4h per day)
Helping others       10h (2h per day)
Slack for surprises   6h (a bit over an hour per day)

With this little list it is already much easier to argue for or against meetings. And if we start tracking how we actually use our time against this list, it usually gets even more helpful. You might want to give it a try.

Strategy three: Plan your slack ahead of time

Just put “Slack Spacers” in your agenda and remove shortly before their time comes up. This way if someone asks you whether you have time for them today you might well be able to say “yes” without having to move any other appointments.

To be able to react to things that are happening every systems needs some slack. If there is not enough slack in the system every little disruption or interference will wreak havoc on the system and might even result in a total system breakdown.

Back in the seventies it was “common knowledge” that in knowledge work one should never plan out more than 60% of one’s day. Simply because “things will happen.” How does that fit in with calendars that are filled up to the brim for the next two weeks?

If you allocate specific times for “creation work” and put them in your calendar you might already have one thing that absorbs some of the “things that happen”, but that’s not always quite what you intended to do with those allocated time slots.

A simple and effective strategy to deal with this is the usage of “Slack Spacers” – appointments with yourself, that are just in your agenda to make sure you don’t plan too much of time too far in advance.

Those could go from 30 minute slices which you remove on the evening of the day before they come up to 4 hour slots twice a week which you remove on Sunday evenings. Or any other sizing and timing that works for you.

Depending on your environment you might either declare them for what they are or hide them behind inconspicuous titles like “Preparation for the XYZ project.”

Wrap-up

So these are three strategies you could put into effect right now

  • Foster collaboration by planning the time you work together
  • Get control of the amount of work you can do by allocating capacity deliberately
  • Create maneuverability by explicitly blocking time for work that shows up unannounced.

till next time
  Michael Mahlberg

Sunday, February 27, 2022

Unplanned work is killing us – really?

One of the things I often hear teams complain about is the amount of unplanned work they have to handle.

Drowning in irrefutable small requests

This unplanned work also frequently seems to be “irrefutable.” But is it? What does it mean to take up an unlimited amount of irrefutable work that has to be done right away?

Starting a new task immediately when it arrives means that you either have been idle when it arrived or –just as plausible– you had to put the stuff you were working on to the side. As long as you only have one item of irrefutable work at a time that might work. However the problem begins as soon as the next piece of unplanned work arrives before you were able to complete the current one.

In this situation you’re most probably not idle (since you’re working on the previous irrefutable piece of work) and you can’t easily put away your current work (because, well, it is also irrefutable).

This dynamic usually leads to a cascade of interrupted work that has been labeled as “irrefutable” and that still gets tossed in the “waiting bin” at the back end.

Most of the time, deciding for stuff to hunker in some “waiting“ state late in the process makes the “client” unhappy – the very person who insisted on the the irrefutability of the work.

This problem gets worse because often there isn’t any time to inform the original client that their work has been paused. After all, the new piece of irrefutable work had to be started immediately!

Thus, even though people try to work on the requirements coming at them as fast as they can it seems to be an uphill battle without much chance of ever getting a grip on the work.

But is that really the only way?

Accept reality

Once we face the fact that in these situations things will take longer to be completed than the mere net working time, we can employ other approaches to get on top of the situation.

There is this seemingly little trick that enables us to transform unplanned work into planned work. It’s called Planning. And the cool thing is that it doesn’t have to be big.

Once you know how many irrefutable small request usually land in your lap each day you can re-structure your day to handle them way more effectively.

You can get that number either from your gut feeling, or from some simple kind of low tech metric like tally marks on a sticky-note near your keyboard. Or maybe just start with an arbitrary guess and iterate towards better numbers later.

Planning to plan

So if you come to the conclusion that if all that work came in structured you could do it in 2 hours a day on average, there are two structural elements you could introduce to your daily structure to handle this

  • Firstly block out those two hours from your schedule. You will lose 2 hours per day anyway in which you will not be working on standard work. This is part of the “accept reality” thinking.
  • Set aside a couple of minutes for planning when you will work on these items and for feedback every couple of hours. Assuming you work 8 hours a day, I would take 5 minutes every two hours for “planning” which leaves us with 2 planning events per day.

All you do in these 5 Minutes is a quick check whether the requests actually fall into the category of “small” request.

If they do, schedule them for later today or next day, based on a rough guesstimation of the amount of work you already scheduled for the respective window and the perceived importance of the task. After scheduling the request you might want to let the client know that you scheduled the item and for when.

If they are not of the category “small” you have a different problem at hand – here you might still want to reserve a small amount of time in the 2 hour window to draft a more detailed feedback on why this request has to be discussed on another level. Still, you do this answering as a planned activity.

With just accepting that the two hours you ‘lose’ per day are actually lost for standard work and subtracting 10 more standard-work-minutes from your working day, you can probably convert 90% of your unplanned work into planned work. Without adding to the actual customer lead time of the items that used to ruin your day in the form of unplanned work.

And as almost every situation is unique, you most probably will have to come up with different numbers, but the general principles statet here should be applicable to most situations.

till next time
  Michael Mahlberg

Sunday, February 13, 2022

Is the user story overrated? Some story patterns and formats to learn from

The term “User Story” or simply “Story” as a shorthand for a requirement has become quite widespread these days. But what does it actually mean and how can we benefit best from it?

We all know, what a story is, don’t we?

Let’s try this one on, for size:

“Once upon a time, there was… here goes the story … ever after”

That’s the kind of story that most people in the real world think about, when they hear the term “story.”

In the agile realm stories seem to be a different kind of beast

As I point out below, my personal recommendation is something quite different, but in the realm of Agile, stories seem to be something other than in the rest of the world. Within the realm of Agile, the majority of people seem to believe that the “requirements packaged in the form of a story” is the central element that everything revolves around.

That extends so far, that even the “speed” of development teams is (way too) often measured in something called story-points – even though at least one of the potential inventors of the story-point concept says “I […]may have invented story points, and if I did, I’m sorry now.

And almost everyone in that realm, as well as in its adjacent territories, have –at one time or the other– heard the stipulation that a well-crafted story

  • starts with “As a <role>…”,
  • has an important “…I want <System behavior>…” in the middle
  • and –in the better cases– ends with “…so that <desired business effect>.”

So – why is this incarnation of the concept “story” so prevalent in the realm of Agile? And is it really the best way to handle requirements in contemporary endeavors? To write better stories today, we need to have a look at how stories came to be such an important instrument in the realm of “Agile Software Development”1 in the first place.

How stories came to software development

Back in the day, before the “Manifesto for Agile Software Development” was written, there were several approaches whose champions called their movement “lightweight software development” and who would later come together and write down what unified their approaches under the moniker “Agile Software Development.” These approaches used all kinds of helpful ways to describe what the system should be able to do.

In Scrum they had the PBI (Product Backlog Item), in Crystal the use case was somewhat prominent, other approaches used comparable artifacts. Extreme Programming was the one that used something called a User Story.

This concept of the user story somehow had such an appeal, that many of the other approaches embraced the idea – more or less.

It was more about the telling, than about the story

A key component behind the idea to use “stories” has even made it into the Manifesto for Agile Software Development – To quote the sixth principle from that manifesto

“The most efficient and effective method of
conveying information to and within a development
team is face-to-face conversation.”

Before the recommendation that requirements should be talked about was written down in that form, it was embodied in ideas like CCC Card – Conversation – Confirmation or the nice quote from the Book XP-Installed from the year 2000 that a card is a promise to have a “series of conversations about the topic.”

Unfortunately, in today’s world the concept of On-site customers often has been reduced to a person who is called Product Owner but doesn’t have any real business authority and spends about two hours with the team every two weeks. Under these circumstances it seems questionable whether this approach to product development is still viable for all cases.

But I am convinced that understanding why it was okay to write only one sentence to represent a complex requirement back in the early days of lightweight methods helps a lot with writing good stories today.

The fact that the way of working that lead to the original user story is hardly feasible in today’s “corporate agile” with all its compromises, has a direct impact here. It implies that we need something more than just the concept of a “User Story” if we want to capture and process requirements in an efficient manner.

Don’t put the story in the center, focus on the value and the work item

What most approaches propose, is some container that represents “value for someone.” In the process framework Scrum this is called Product Backlog Item, in more general approaches –like the Kanban Method– it is often simply called Work Item.

Such a work item –to go with the broader term– can have many structures. A few common attributes of many such item types are:

Of course, one of the attributes needs to be the actual requirement. And that could be represented by a story. But does that have to be a user story? Actually, there are some pretty helpful alternatives out there.

If you use some kind of story, get to know several types of stories well

As it is often the case, the habitat of the original user story provided many things that were no longer present once the concept was mimicked elsewhere. And as time went by, some people re-discovered what a story could mean for them. Some other people –many, actually– got confused by the story concept since they never really saw it in action and only knew about it through very indirect word of mouth.

Stakeholder Story

After the “As a «role» I want…” format for user stories had been around for quite a while, Liz Keogh pointed out that many of the so called user stories out there are not actual user stories but instead Stakeholder Stories.

  • Format of the Stakeholder Story
    • Liz Keogh described her ideas and observations in the 2010 Article “They’re not User Stories.”

    • The generic form of this kind of story –the way I use it these days– is

      • In order to «the required business effect»

      • «some stakeholder or stakeholder persona»

      • «wants,need,requires,…» «some kind of system behavior or future state»

  • Context for the Stakeholder Story
    • This is an extremely useful perspective if you have to describe requirements that are not actually wanted by the end user of the system, or that don’t actually have a direct user interaction.
    • Most of the requirements I encounter in enterprise contexts are more stakeholder-driven than user driven. (Legal requirement for example. Something like “To avoid being sued for GDPR violations our CISO requires that we have some GDPR-compliant deletion mechanisms that could be executed at least manually if ever a user actually should file a complaint that conforms to article 17 of the GDPR.”)
  • Caveats for the Stakeholder Story
    • The stakeholder should be as tangible and concrete as possible. Unlike with the model of personas in user stories for stakeholders in user stories, it is extremely helpful to name a real person for stakeholders in stakeholder-stories.
  • What to avoid for the Stakeholder Story
    • The most common problem I see with stakeholder stories these days is that the required business effect gets confused mixed up with the system behavior or future state.

User Story

It was probably Mike Cohn who popularized the now so common form of user stories in his 2004 and 2005 books “User Stories Applied” and “Agile Estimation and Planning” but to my knowledge Rachel Davies came up with it around 2002 at Connextra (actually that’s also what Mike Cohn’s post about the three part user story tells us)

  • Format of the User Story
    • The now prevalent way to capture user stories is the well known

    • “As a «role or persona» I want «system behavior» so that «desired business outcome».”

    • This is described (amongst other sources) in the often quoted Article Why the Three-Part User Story Template Works So Well by Mike Cohn.

  • Context for the User Story in this sense
    • Helpful if you really have a product (sometimes a project and seldomly a service) that has actual interactions with actual users
  • Caveats for the User Story in this sense
    • It should describe an interaction between a user and a system that will be possible after the requirements has been implemented.
  • What to avoid for the User Story
    • A story like “As a team member, I want another team member to implement the database logic for the WhatNotField so that it will be available” is using the format alright, but misses almost all point of using User Stories.

Job Story

To my knowledge the whole “Jobs to be Done” way of approaching product challenges became popularized through Alex Osterwalder’s work with Strategyzer around the value proposition canvas. [Please let me know, if you know the whole back-story, I’d be really interested in learning about that] Soon after that the JTBD idea proved so powerful that it spawned it’s own community.

Thanks to my esteemed colleague Matthias I learned about the job story format and the whole idea of using job stories to work on product ideas

  • Format of the Job Story
    • The article Replacing The User Story With The Job Story describe the idea of the Job Story as separating situation, motivation and expected outcome by using the format
    • When ________, (the situation part)

    • I want to ________ (the motivation part)

    • so I can ________ (the expected outcome part)

  • Context for the Job Story
    • Good for very young stories, when you still try to figure out what you’re really talking about.
  • Caveats for the Job Story
    • Unlike Stakeholder Stories and User Stories, Job Stories don’t (yet) provide an easy way to fill out the ________ part, so you really need to dive into the ideas outlined in the above mentioned articles and there can be a lot of discussion about the “right” way to write such a story.
  • What to avoid for the Job Story
    • Don’t treat it like a piece of functionality that just needs to be executed. Job Stories make for good candidates or the narrative flow of Story Maps. There’s also an 2-page summary explanation of Story Maps if you want to know more about that concept.

Of course this only covers some aspects of the usage of stories in todays post-agile society, and I would strongly encourage anyone to look (deeply) into the stuff about INVEST and SMART and at User Story Mapping, to get event more background with regard to working effectively with stories to represent aspects of requirements, but I hope this article gives you some ideas on when and how to use some other kinds of stories to represent requirements that are really hard to fit in the “As a «Role» I want…” format.

till next time
  Michael Mahlberg


  1. (Remember: There is not really an Agile Manifesto)↩︎

Sunday, January 30, 2022

There is no Agile Manifesto

Just a little reminder: what many people nowadays think is a way of living or even a way of designing whole organisations was originally something quite different…

What most people call “The Agile Manifesto” actually has a title.

it is called Manifesto for Agile Software Development

And its authors propose the “Twelve Principles of Agile Software.

  • It does not specify a defined approach to continuous improvement – TPS (Toyota Production System) does that, for example
  • It does not elaborate on good ways to optimize lead times – The ToC (Theory of Constraints) does that, for example
  • It does not express any opinion on how a company should be structured in the post-Taylor era – Sociocracy and its derivates do that for example. So does New Work
  • It does not tell anyone how to handle finances without upfront budget plans – Beyond Budgeting does that, for example

And all of the approaches on the right hand side came into existence long before 2001, the year the “Manifesto for Agile Software Development“ was drafted.

If you look a bit further on the original web-page that launched the term “Agile” into the world, you’ll find that in the section “About the Manifesto” as well as in the headline above the twelve principles, it has been called “The Agile Manifesto” even by its authors. Maybe this helps explaining some of the confusion.

Personally, I find it very helpful to remember the context where the whole idea of “Agile” came from – maybe it’s helpful for you, too.

till next time
  Michael Mahlberg