Thoughts for improving your business

What is agile? Get in touch

Creating an ATDD Ready Sprint Backlog

The arcticle about ‘Creating an ATDD Ready Sprint Backlog’ can be found here

Ralph Jocham is the founder of effective agile. GmbH 

effective agile. believes that agile is more than just working in iterations, doing a daily standup or writing a unit test after the fact.
effective agile. believes that people are the most important part of a successful project and that we should focus on ways to enable them.

effective agile. believes that the world for knowledge workers is changing faster than ever. Only by harnessing the skills of each individual and putting trust into them, technology companies are able to survive.

effective agile. believes that we have to address our current believe system and that we need to challenge everything.

The Separation of Power in Scrum

There is one element in Scrum which I really appreciate. It is the separation of power in Scrum. 
What exactly do I mean with this? Democracies are based on the separation of powers they require.
  • Legislative
  • Executive
  • Judicative
Each one has their rights and responsibilities. The other two watch out and make sure, that third doesn’t abuse it’s power.
In totalitarian governments this is not given. One entity reigns over all three. The usual result is that a few benefit and many suffer – from individuals to whole economies.

What does this have to do with Scrum. Well, nothing – at least at a first glance. But if we apply this concept to classical management, the project manager has the possibility to act as a dictator. Screen Shot 2012-04-12 at 13.59.42
He or she can decide about all three elements: scope, schedule, people.

The image above is the incomplete ‘iron triangle of quality’. It tells us, you can choose two out of the three, the third has to give. For example, if we have a certain amount of scope to implement by a given date, we need to adjust the number of people working on it.
If all three are set then the quality of the product under development will be sacrified when things get tough. Quality is the forth hidden element. Often the manager tries to convince us about the attainability of the goal with sentences like ‘I know it is aggressive but …. ‘, ‘You are not a team player ….’ and more. Screen Shot 2012-04-12 at 14.00.11
Quality dies first on every software project. It is not transparent per se, it won’t show until very late in the project or even until the product has been released. This manifests in very high TCO (total cost ownership). Eventually the developers require to rewrite the piece of sh.. garbage software. (see my blog Resetting the Shitty Counter)

How is this handled with Scrum? In Scrum the Definition of Done (DoD) states certain attributes or activities which have to be present in order to guarantee a high quality, potentially releasable product. The compliance of the DoD is paramount to a high quality product with happy customers and low TOC. Now Scrum is not immune to crunch times, times when the Product Owner (PO) is tempted to push the Development Team a little further. In those situations it would be just to easy to abandon the DoD and reduce quality to keep the date and make the Product Owner happy.
This is when the Scrum Master comes in. She will make sure that the DoD stays enforced and keeps the PO at bay. Essentially she protects the people (Development Team) so that they can work in the agreed way and thereby create a high quality product. 

In Scrum the Product Owner has the right to decide which feature gets developed in which order. His tool for that is the Product Backlog.
Screen Shot 2012-04-12 at 14.00.26
and the Scrum Master ensures that the Delelopment Team has the right to estimate the work according to the Definition of Done and to implement it according to the Definition of Done.
Screen Shot 2012-04-12 at 14.00.39
The separation of power protects the Development Team and allows it to deliver high quality product increments throughout the project. This sustainable approach guarantees high quality software with high ROI, low TCO — easy to maintain, easy to support, easy to enhance — for a long time. Best, you should see happy, engaged developers.

In the end everybody wins!

The 1000 students challenge (2)

In May of last year I was able to run my first Pay It Forward Scrum Training at the University of Bern in Switzerland. I had blogged about this event there:  The 1000 Students Challenge
Last week-end I had the chance to do the second run at the University of Applied Sciences (FHNW) in Brugg Switzerland. 20 students volunteered their week-ends to learn about Scrum by attending the Professional Scrum Master Training or short PSM. It was great fun for them and myself and it is good to know that 20 future IT specialists will join the workforce pre-equipped with Agile and Scrum.
967 more to go …
PS If you are interested in hosting such an event at your institution please contact me at

Cross-Functionality in Scrum

Scrum recommends that a team should feature all the skills required in order to deliver the releasable product increment by the end of the Sprint. 

Why is it a good thing to have all the skills needed? It is all about dependencies. We try to design our software systems to be loosely coupled and highly cohesive. The same principles apply to team composition. We want our team to be like a special unit, self sufficient and able to deliver on the chosen assignment.
So, what would happen if one significant skill is missing from the team? The team will have a strong external dependency and will need to ask for support from external sources. This is a huge risk. Will the team get access to the person at the right time for the right duration? This creates an additional drag factor and affects the delivery date or reduces the scope of the product to be delivered.

Scrum does not require cross-functional teams, it only recommends them. In practice they have shown to be a significant boost for productivity. Especially in combination with self-organization.

Often it is misunderstood that cross-functional means that any person in the team needs to be able to do all the upcoming work. This is wrong. The unit that needs to be cross-functional is the development team. The dev team has the responsibility to self-organize in order to maximize the usage of its skills.  Hence, the team might be composed of specialists only, however the sum of the individuals needs to possess all required skills.

Nevertheless, as described above, if you only have specialists then you will have rather large teams, as for every skill you will need at least one developer. This will cause larger than needed teams and probably all other kinds of problems: part time team members based on FTE mathematics, work organized by activity not by feature, bad ‘unskilled’ estimates, …

Therefore agile teams favor generalists, developers with a rounded and versatile skill set. I like to use the term specialized generalists, strong all-rounders with one top notch special area.

The Weight of the Christmas Cake

At my daughters school they organize a charity event every year. There are all kinds of different foods, games and lotteries were you can spent money for the good cause. Since it is a british school the mandatory guessing of the Christmas cake weight must not be forgotten. Christmas-cake-kit-007
In Agile and Scrum we praise ourselves about our estimation skills. It is not that we claim to be infallible. No, we also claim that laws in mathematics help us to be precise. One in particular is of interest: The law of large numbers.
The cake weighing proved to be change to put this practice under test.
This year we had 45 guesses and the weight from all the guesses is 3’742 kg. The real weight was 3’520 kg. The guessed weight is 94.1% accurate. Pretty impressive!

Agile A3 Sprint Report

I really enjoy working with A3 Reports. They have the habit of making you to distill out the crucial information, to really drill down.
There are different kinds of A3 reports. The book ‘Understanding A3 Thinking: A Critical Component of Toyota’s PDCA Management System’ describes three different report types. Problem-, Proposal-, and Status Report. The Problem Solving is the one I use most often.

So what is an A3 report. A3 Reports are size limited and have to fit on an A3 size of paper (for US folks, it is 11.69 × 16.54 inches). The size limitation requires to only present important informations. Furthermore, the use of charts, graphics and other none textual descriptions is highly encouraged. The less words the better. Since the Problem A3 Report describes the current and the target situation they serve as excellent PDCA (plan do check act) tools.

In my profession as Scrum coach, I grew tired of all the different and bureaucratic status reports. Usually the decision makers require lots of text and a traffic light. The text does not get read and all decisions are based on a single color. So, my idea was to create an A3 Report which is hardly text and more differentiated then only one color. Providing a quick and easy way for more information width. 

This is how it looks like:
Top Left – Sprint Burndown
Current Sprint burndown. Blue remaining work in hours, Green remaining Story Points

Top Right – Release Burndown (Burnup in this case)
Up to date, including last Sprint, release burnup

Bottom Left – Risk list 
Up to date, risk list a la Frederic Brooks. New risks are continuously discovered and either handled by outside help (i.e. by the Scrum Master) or they get put into the Product Backlog and mitigated when the Product Owner puts them into a Sprint. The risk list is dynamic and subject to change from Sprint to Sprint. (*) Mitigated risks are kept on the list.

Bottom Right – Open Bugs
Up to date number of open bugs. If I can have my way on a project, the 10-Finger-Rule applies. The moment you run out of fingers to count bugs, you need to fix bugs first. This guarantees clean, maintainable and sustainable code. If a bug is rather tricky to fix, it can be put into the Product Backlog analog to a risk. Again, the Product Owner then schedules the fix, if she sees it to be appropriate. Also, this avoids the growing number of bug triage meetings towards the end of an project (**). If you work on an legacy project with X bugs, the rule changes to X+10. Done right, X will decrease over time and product quality will go up.

(*) On many projects the institutionalized company process requires a risk list. So, at the early beginning of the project a risk list is compiled, checked off on the check list and then put into a drawer. No transparency and accountability.

(**) On many projects you you have severity levels from P1 to P4. With P1 being worst and P4 minor. Usually, you have a company mandate, that no software can be shipped with open P1 and P2 bugs. Guess what … in the growing number of meetings after each bug fixing cycle, panic tends to creep up. The human reaction is to ‘mis’-label P2 as P3. Problem fixed, process followed.

An Agile Transition done Right

A little over a year, I kicked off an agile change program at a leading swiss devices company. Two weeks ago, I had a chance to visit them during an open house on their new premisses.

This company was fully committed to walk the talk, they decided to do what needed doing.

In August 2010, we started of with a five day PSD (Java) training. I trained 12 very strong developers of various expericene levels. This training was intended to ascertain whether they really want to work in an agile manner. Brave move from management, but to no surprise the developers fell in love with Agile and Scrum and were hooked on day three of the five day training. This training is excellent!
After the developers gave their thumbs up, it took about a month to set everything in motion. This is is a mid size company with over 300 employees at their head quarters. Once we got the ball rolling the company was ready. They had torn down walls to create right sized team rooms, organized great Story Boards or shall I say walls. Again, they were serious. In the first two weeks, apart from coaching the teams, I trained ten people as Scrum Master (PSM) and ten more as Product Owner (PSPO). We didn’t really know who is going to fulfill which role, we just knew that the future person would be coming from those groups. Also, the intention was to really create an agile culture and what could be better then to reach out and train and convince as many people as possible. After that I gave about ten introductions to Scrum, each lasting about two hours; the target audience was marketing, sales, support and other departments on the projects periphery.
For the frist two sprints — 2 weeks in duration — I was coaching a 100%. Essentially being a Scrum Master. After two Sprints I was only around at the beginning and at the end of the Sprint. During my absence the teams had the chance to walk on their own without training wheels, getting first-hand experience into when they started to struggle. After 5 Sprints the teams decided on their own who should replace me as the Scrum Master. In the remaing four Sprints, I worked closely together with my replacements. My presence became less and less. At the end it was about 20% of the time. By the end of last year, they didn’t need me anymore and I moved on.
This whole engagment lasted about four month in calendar time and about two month of coaching from my side.

Now, during the open hours, I had the chance to meetup with a couple of guys from the old teams and we chatted a little. It was a great to see, that all of them really had drunken the agile CoolAid. They did not roll back one inch – instead they kept on pushing hard. Now, there are three teams and more coming soon.
For me and my company effective agile. this coaching approach has become my favorite modus operandi for change engagements.

By the way, the last two releases were two and six weeks early! It humbles me to know, that I was the initation.

The right Sprint Length

In order to keep the blog short here is the answer:

     Let the Development Team decide

If you are curious why this is the answer then keep on reading.

Again and again, I hear from outsiders that we should extend our Sprint length. Usually, I like to work with two weeks. When I ask for a reason, the answer is typically the following: ‘With all the meetings you don’t have time to do real work, so I think you should add another week or maybe even two!’.

Interesting statement, but completely wrong and obviously from someone who does not understand Scrum. 

Here is Why:
1)  Apart from the 15 min Daily Scrum all meeting durations are proportional to the Sprint length. The Product Backlog grooming which is paramount to an executable Product Backlog and often neglected in literature is about 5% of the Sprint length in average by my experience 
Here a chart showing how the meeting time per day actually slightly increases when the Sprint is longer
2)  All of the above meetings are real work and important. They are the empirical process controls you need when you work in complex enviScreen shot 2011-10-20 at 17.29.15ronments and domains. Especially the Daily Scrum is the catalyst for an effective and hence productive working day. The Daily Scrum is not disruptive it creates focus and team spirit. I just read the slides of a recent talk of Tobias Mayer. He stresses that Scrum is essentially five things.
  • self-organization
  • collaboration
  • focus
  • alignment
  • rhythm
Each of these points is addressed by one or even a couple of those meetings. 

3)  The rhythm is to be discovered by the team. In my experience two weeks is a great fit for most environments but if the team cannot find it’s own rhythm and cadence within a given Sprint length, then let the team decide. Often they adjust their defined working rules, and sometimes they decide a different Sprint length is what they need. If so, let them choose whether they want to go shorter or longer. The rhythm is a function of the domain, the planning horizon, team configuration, dependencies, other Scrum Teams, and more. It cannot be decided by an outsider.

4)  Finally, the underlying problem is, that people with that recommendation are usually bottlenecks themselves. It is a latency issue by those very individuals. When problems do pop up it takes too long for them to turn around. Since a loss of a couple of hours or even days is more visible or shall we say dramatic in a two weeks Sprint then a four weeks Sprint. Those ‘well’ intended recommendations try to mask the problem and thereby slowly turn the project into a ScrumBut.

Scrum is Open for Extension and Modification

Today, and Scrum Inc. are announcing a formal model for modifying and extending Scrum. Scrum’s creators, Jeff Sutherland and Ken Schwaber, are inviting practitioners from around the world to contribute to Scrum’s future.

Scrum was first developed 15+ years ago, and it has been evolving and adapting ever since. Informed by their experiences and those of the Scrum community, Jeff and Ken have carefully codified the framework in the Scrum Guide, which documents the basic rules, artifacts, and events of Scrum.

Today’s announcement marks a new era in Scrum’s evolution by making available a public mechanism for providing feedback on the Scrum Guide and a model for proposing extensions to the basic framework.

The formal process for proposing and integrating changes into Scrum is available online at To learn more about Scrum, or proposing your own contribution to Scrum, you can use the following links:

Read It –
Change It –
Extend It –

We look forward to hearing from you.

Capacity: Help with Excel

For every Sprint Planning the capacity, the hours available for the team is an important ingredient. Usually, this is rather straight forward and easy to do. At my current project we have two teams. Each team is cross-functional, i.e. being able to deliver a piece of done software. However, in this case, the skills are very different, so that the overall number of available hours can be deceptive as there might be 120 h of development tasks and only 80 h available. We actually run into this problem when we got additional business analysts. The overall capacity went up so that we could commit to more user stories. As a result, we had far too much programming work and not enough analysis work.

Now, we break the capacity down per skills we need in our teams. Once Sprint Planning Two is over, we add up the hours per skill and see if we are still in the range. If yes – great! Otherwise the Dev Team and Product Owner self-organize and figure out a way to handle it.

Long story short – here is the link to the Excel sheet. It is not fancy but does a great job for the current project.