Agile and Lean - Agile and Scrum Blog

Agile Prague Conference 2015

Agile Prague Conference 2015 – Sep 14-15, 2015 got awesome speakers for this year. We were able to get unique experts from all different areas of Agile and Scrum. We have talks on Agile Product Management, Scaling Scrum, DevOps, Test Driven Development – TDD, Behavior Driven Development – BDD, change and improvements.

This year we continue with 2 full conference days – every day we plan for 2 parallel tracks and one additional workshop/game/open space track in the afternoons at open area.

So far you can be looking forward to the following keynote speakers:

– Jurgen Appelo | selected by Inc. “100 Great Leadership Speakers for Your Next Conference”

– David Hussman | author of the Dude’s Law

– Joshua Kerievsky | protect people by engineering anzen (“safety” in Japanese) into workspaces and code bases

– Bas Vodde | creator of Large-Scale Scrum (LeSS), a framework for scaling agile development

and talks from:

– Vasco Duarte | improving estimates for software with #NoEstimates

– Andrea Provaglio | speaking at AgilePrague for the 5th time

– Jutta Eckstein | enabling Agile development on the organizational level

– Senta Jakobsen | enables distributed development

– Cliff Hazell | at Spotify we aim to build shared views and models to reduce unnecessary ambiguity

– Oded Tamir | DevOps is taking the Agile to the next level

– Pawel Brodzinski | the missing bit is almost never a tool or a method but sort of myth
and that’s indeed not all.

Those wonderful speakers are just beginning of our list. In addition you can be looking forward to games, workshops, case-studies, and last but not least an Open Jam session – believe it or not, for most of conference attendees the open space is the most valuable part of every conference. How it works? Bring your idea/question/theme to discuss and run a session yourselves. Or join any group where you are interested in the subject and share your experiences and hints. Or just listen. It’s an awesome opportunity to get insights from each other.

Looking forward? Have a look to Agile Prague Conference web site, full program and register now!

Sprint Planning in 30 minutes

How much time takes your team to finish Sprint Planning? To my experience it could be anything in between of above mentioned 30 minutes and full day. If you are closer to the second option and it feels scary, annoying, waste of time for you, let’s have a look at few recommendations how to cut it out into 30 minutes.

First, let’s see how to run the Sprint Planning itself. I recommend Product Owners to come to the Sprint Planning with physical cards for each User Story. They quickly introduce them, answer questions if needed and then let the team choose out of them. Don’t bring the exact ordered list; let them freely choose from the cards. There are two reasons for that. First, you maximize work done as they can organize themselves in a way they are most efficient, and at the same time there is higher commitment as well. Second, you build a trust between team and Product Owner. You trust them they will choose the right User Story which brings the highest value at the moment. Once the team select the User Stories witch they believe they are able to finish within the next Sprint and put them on Scrum board, Product Owner and Scrum Master can leave and let the team finish the Sprint Planning. During this second phase team will collaboratively split the selected User Stories into maximum one day tasks and revise the Sprint Backlog commitment. After 30 min they are done, have full board of cards and can start working.

If that still feel unbelievable, let’s have a look to the preparation. There are three key recommendations you should do in order to make your planning fast and meaningful. First is for Product Owner. 2-3 days before the Sprint planning let the team know what are your priorities for the next Sprint, so they can have a look and prepare themselves, ask questions, etc. Second is proper Backlog Grooming. The goal of Backlog Grooming is make sure the team understand Product/Release backlog (i.e. all User Stories, Super User Stories, Epics and vision). At this time team do the estimations and help Product Owner to split User Stories which are too big, or add Acceptance Criteria. Once understood, they are ready to be planned to the Sprint.

To summarize it, if you are not able to do such fast planning, improve your preparation (team time to prepare, grooming, pre-planning) so the planning is here not to investigate new functionality but to confirm how much we can make. Doing that, you gain motivated team who is not wasting time at never ending planning, better reliable Sprint plans and higher backlog quality as you are not pushed to do splits and changes at the last moment. Start step by step and continuously decrease your time needed. It’ll go much faster than you would imagine.

Online Scrum Board

I belong to the Agile crowd who believe the physical board is very much useful and can’t be easily substituted for any online board. Let me give you a few reasons.

5 reasons why not to use electronic board

We are *still* limited by the screen inches and no electronic board gives you good overall Sprint visibility.
No electronic card takes pride in your handwriting, so the ownership of the whole board is much closer to “someone else’s problem”.
You can’t touch it, move it, or throw it away. It’s annoying how many fields in the *average* system are required to add a task. It’s a tiresome to crumble tasks to one day activities.
It’s hard to draw on it. There is no creativity. It’s just reporting by definition.
Regardless of the ability to share, usually the ScrumMaster controls the board during Standups. Not any team members.

To make it simple, to organize yourself as a Scrum team you need very good visibility of what is already done, what is in progress and what still needs to be done and who is currently working on which part. As there are no assigned User Stories to any team member, every individual is responsible for finishing Sprint Backlog. To be able to organize your daily work yourself as a team, you might need a flexibility – depending on where you are you might decide to distinguish tasks by colors, next time by shapes, then you start tracking dots per day, and the next time tear the task if it get blocked or anything else. You can start right away, and stop any time it suits you and there is no need to win over your system.

To make it clear, I’m not suggesting now your overall backlog should be at the board even if there are companies who work only this way. However, for this time of being I’ve been focusing on Sprint commitment, and simple tool which helps team to synchronize themselves. So keep your *future* – the User Stories – in the system, keep Sprint tasks and team synchronization be driven by physical board not connected to any system, and then link back any commit or important note back to your User Story in the system so you have a history and traceability.

And yes, I understand that some teams might not be at the same location, and can be spread over the world. So if you have such situation, still you might prefer flexible tool which gives you a good visualization. There is no ideal tool like that, but there is one I learned from some of my distributed teams. You can see the picture of that board below. It’s easy to use, it gives quite good overview as well. And yes, I can come up with hundreds of improvements, but I still like the simplicity of this solution. You can try it here.

Scrum Master is Not a Secretary of a Team

I’ve been wondering why so many teams believe that Scrum Master is here to draw burndown charts, prepare reports and be the only point of contact for the team, whenever anyone wants anything from them. They maintain the board, write cards, and prepare all you can imagine. And the team works ok, but surprisingly they are not at all self-organized. Such Scrum Master role is quite boring. But that’s not what was intended by the role of Scrum Master. I guess the reasons for that are coming from two different motives.

Firstly, Scrum Masters are often missing the real experience with Scrum, teamwork and self-organization. They are in a new role and want to succeed in it. They biggest fear is they would not be useful to the team, and team would not appreciate their work. So they try to do their best to make their work visible for everyone. Be helpful. The biggest Scrum Master’s trap is to be locked in the position of caring mummy who is scared to let her grown up kids go their way. But in such case you will never get real Scrum team.

The other reason comes from one of the Scrum Master responsibilities – to remove impediments. It’s the only responsibility which seems to be easy to do for starting Scrum Masters. Seems to be. Unfortunately, that often comes with huge misunderstanding. The goal of the Scrum Master is to build self-organized team which in the ideal theoretical world means “do nothing”. In other words, Scrum Master is here to help a team to find solutions to their problems, not to solve problems oneself. Nonetheless, most of the beginning Scrum Masters are eager to help, happy to do any work needed if it helps their team. And they don’t see that by doing it they are destroying the team.

Is Product Owner part of the team?

When you ask this question in the companies, you find out that about 30% of teams believe that he or she is not. If you ask why not, you find out that they feel their Product Owners are far away from them, they don’t help them, and they don’t understand them. And I’m not talking about physical distance now. So where is the problem? In many companies, at the beginning of their Agile transformation, they simply move team to Scrum and the Product Managers to Product Owners. What happens? They don’t have a time to be Product Owners as they are responsible for several huge products. Luckily they understand the product, but they have no time to share their understanding at any higher granularity than general ideas or epics. And that’s indeed not enough. Such teams are having a Product Owner Proxy, or Tactical Product Owner who is in reality acting like real Product Owner and don’t miss their business Product Owner. Why is that usually not good? We are missing the “one PO voice” and we are losing the business driven approach in favor of technical point of view. In such environments we are as well missing the push to “maximize work not done”, which is one of the Agile Manifesto principles. That is indeed not good for either team or product.

Then we have about 50% of companies where they believe the Product Owner is part of the team, but he is not responsible for writing User Stories. Why not? Usually because he or she doesn’t understand the technical aspects, so how can he possibly do that? They usually don’t invite him or her to the retrospective either, because… well… he is a team, but retrospective is for development team only. So it’s kind of unclear.

The remaining 20% take their Product Owner as their member. They invite him to the retrospective, they trust each other. If that’s possible, they sit together. If not, they speak with each other often. Such Product Owner relationship is very helpful. Not only for your team, but the product as well.

Measurements are dead, let’s measure

During my career as both Director of Engineering and independent Agile Coach, I’ve been hearing still the same grumble from managers: “We can’t get rid of measurements and KPI’s. How else could we know if the person is performing well, how can we compare people?” and at the same time, grouching from the team members: “We don’t like the individual based KPI’s and measurements, how are we supposed to be a good team when our managers can misuse that against any team member?” It’s surprising but no one likes individual metrics, they all admit they are useless, but they are all afraid to try anything else.

So if you have a bit of courage, you may try this: It’s based on coaching relative scale and is team oriented: 1 stands for 🙁 and 9 stands for 🙂 and it’s great if you add a reason for rating lower than 4 and higher that 6. Firstly, let the Product Owner give a team his number how he is happy with the team.

As a second input, ask Scrum Master to give a number to every team member how much he is happy with this person as a team player. Let them discuss it, but make sure the discussion is not about “why I’ve got 5 instead of 7”, but is focused on future development of that person discussion “what should I do differently so that I’ll get 7 next time”.

And last number goes from the team members. The best you can do for this part is to ask everyone to divide 100$ to all team members except himself. You may worry that they can agree with each other and rotate all the money one by one, or distribute them equally, but that’s not common in real life. The great think on this evaluation is that the team members are giving a feedback to themselves. So every team member gets an answer to the question how do you value my contribution to the team? And if you find out the other team members don’t see any value in your work, you would most likely be very much concerned about that situation and asking how can I do differently so that you value my work more.

Combining those three inputs you will learn much more than from traditional metrics, regardless the company size and culture. It’s working just awesome, but you have to have courage to give it a try.

And when this is just normal for you, you can take it one step ahead. The fully Agile companies are using such tool as the only one appraisal tool across the company. No other bonuses than those distributed by employees to the other employees. So in such company, if you feel you would like to appreciate the receptionist, give her some part of your bonus sum. The other one can be for your colleague, another part for a developer from a different team who helped you with some issue. And when you are afraid it’s too crazy for you, I would like to remind you that we are only talking about bonus distribution, not the whole salary. When you do so, you will increase team cooperation over individual heroes work, and openness and transparency over politics and gossiping. And it would be fun. If you still don’t know, start with Appreciation cards. Make them available and encourage people to give them to each other. Even by that you will learn a lot about yourself, your team and the whole organization ecosystem.

The future of Agile and Scrum

A few weeks ago we’ve been hosting a board meeting of Agile Alliance here at Prague. And the last question at the local community event with board members was the future of Agile. You can have a look at what they said here:

I would say that I fully agree with what they said. In the future, we will not use any Agile or Scrum any more. It will be already overcome, but until that time, Agile and Scrum are the best methods we know and they are very useful. From nowadays perspective, Agile and Scrum is a Ferrari car or Lamborghini, TGV train, or an A380 airplane – depending on your preferences. Nevertheless, in the long run we would be looking at Agile as Scrum the same way as most of us feel about traveling on horse wagon. With some feeling of nostalgia, but pretty much happy it’s already gone. And the new method would have another cool name like “Queguer” or any other you can imagine and will be much better. But that’s a problem with evolution, you need many, many years to understand the backbone principles, do research, inspect, and adapt. It can’t be made faster.

I believe “Queguer” will be very much change responsive. It will be even more collaborative, going out from the specialization of an individual person to the team sharing knowhow. It will be focused on fast learning and very good at adaptation of whatever is around. But it will not be an ideal method. It will again create lot of pain while Agile teams would be passing through the “Queguer” transformation. It will not be any easy. And last but not least, sooner or later there will be another method which would overcome the “Queguer” and the evolution will continue. But until that time, let’s enjoy using Agile and Scrum.

Product Owner Development Model

What is the difference between requirements, use-cases and User Stories? I’ve been struggling with that question a lot. On one hand it is easy. It’s something completely different. On the other hand, that’s not anything which would help people to understand the difference on their way to implement Agile.
After some time working as Agile Coach, I created this Product Owner development model. It’s focused on product creation and Backlog item definition process.

Level one: User Story is just a special format of a sentence

At this basic level of understanding we are very close to the requirement-like specification. We keep the backlog in the Word document, as we anyway wrote very long sentences and extensive document chapters about the functionality. There is often huge mix of current functionality we want to keep, and new functions. The only change we do with that requirement document is to change/add User Story sentence instead of general name. So we get something like “As a MyCompany, I want new tariff, so that my customers are happier” followed by 2 pages long text description what the “tariff” exactly means. Such User Story may survive at team board for several Sprints without getting done. Surprising, isn’t it? We wrote User Story and it didn’t help!

So this stage is about documents. We create PowerPoint presentations to describe product goals and vision, we use complex roadmaps to define timeframe and we have written long specification documents to describe functionality. The more we write, the better product we have. The understanding of the role of Product Owner is very limited, decisions are often taken as a board of people without real product success responsibility.

Level two: We have ‘bigger fish to fry’, than write User Stories

At this stage we already understand that we have to describe our User Stories better. The team needs higher granularity and detail. But we don’t have time to write User Stories, so that we delegate that unimportant work to some administrative position called business analyst, business requirement specialist, business delivery manager, development team or whoever else is around. We don’t have a time to write such ‘technical’ details. It’s not important for us. Just make sure you will deliver it on time. We have bigger fish to fry. We have to talk to the customer. It’s more than enough to discuss our product ideas and high-level visions. We are responsible for Backlog, and yes, we prioritize it. However, the level of Epics is just about the right level of details.
So this stage is about big high-level decisions and quantity. We already have a Product Owner position, although that person is not often seen. Instead we have the army of people, who are willing to help official Product Owner with creating as many User Stories as you can imagine. What if we need that functionality in the future? Let’s describe all we can possibly do. And if we cover any potential functionality, it must be successful.

Level three: User Story is use-case

Here we finally got it. It’s about functionality slice, it should be INVEST. We have to make it concrete, understandable, and testable end to end functionality. Isn’t that easy? It’s like a use-case, isn’t it? Well, unfortunately, sorry to say that, no. There is a huge difference between use-case and User Story. So what’s the difference? Use-case is end to end functionality which defines what user does and how he is using the product, while the User Story defines only a new/changed functionality. We don’t repeat the current functions anymore and we focus on the changes only.
This stage is already user focused. We start describing different roles. We focus on functionality end to end. However, it’s still not simple and not clear enough. And it’s still not what we expect from the Product Owner.

Level four: We will design one big User Story and copy-paste the rest

This stage looks already pretty good. We have understood that every User Story has three parts – Who, What, and Why, and we think about all three of them. However, we haven’t still understood that every single User Story has its unique value, and it makes sense to invest an energy into individual detail User Story creation process. We are now spending energy describing Super User Stories (smaller and much more concrete pieces than Epics are, although nor small enough to be done in one Sprint yet.) We have great tools, which unfortunately offer a copy-paste feature. So we heavily use it to save our time.

This stage is about User Stories which already create some picture in your head once you read or hear them, but they are very similar to each other and hard to be recognized. We already have spent some time to investigate reasons ‘why’ for bigger chunks of functionality, and we are very happy about it, so we use it at every detail User Story which we create from it – just copy and paste.

Step five: Understand of business value and impact

Finally, we understand that it is worth of investing our time to every single User Story. And we are even looking at it more than once. We reprioritize individual User Stories and not only big Epics. Every User Story has a special role or persona. We have spent time and energy defining every one of them. We encourage ourselves to throw away or postpone User Stories already written, if they don’t match our product/release charter (vision, goals, success measures, timeframe).

We focus on business value and “maximizing work not done” which is one of the core Agile Manifesto twelve principles. We keep our product simple. We try to visualize business value for every User Story in the “Why” part of the formula, so that it helps us to decide on Backlog priorities.
Furthermore, we compare every new User Story with product/release charter and discuss how that User Story contributes to the defined goals and vision. Before we write the complete functionality, we try to measure impact, i.e. if the goal of Epic1 is to limit the traffic through the component A, than individual User Stories may propose different solutions how to filter that traffic out. In traditional management we finish most of them if not completely all. In this stage of this model, we try first to measure the impact by identifying of the percentage of possibly filtered traffic by each solution proposed. And then implement just the ones which have real impact with respect of our goal to limit the traffic. We may identify many great ideas, but we stop implementing as soon as the goal is achieved. At that time we don’t need any other functionality and we can move on to the next important area.

This final Product Owner Development model stage is about business value and impact. The less is more. Product Owner is feeling strong ownership and responsibility over the Product Backlog and individual User Stories. There might be people to help him as Product Owners rarely works alone, nevertheless he understands the importance of his role in defining even the small functional slices as User Stories are. Finally, in this stage the Product Owner is here to shrink possible functionality to the minimum which brings just enough business value. Product Owner must negotiate the functionality and focus more on understanding the customer real needs than all their wishes to come true.

Summary

To summarize it, Product Owner Development model is useful tool which helps you to understand where you are with your Agile Product management and product ownership. It also shows you the way where you shall continue and which areas you shall focus. Theoretically you don’t have to go through this model one by one, but it is very likely you will pass all next layers from the one where you are now even if you stay at that one just a very short time.

Agile Prague Conference 2014

It’s time to invite you to the 4th year of Agile Prague Conference. . It is in September and we have wonderful speakers this year. You can have a look at the full program, but if it’s too long to read it through, here is my personal recommendation:

Linda Rising is the person I like; no matter what the topic is it’s always one of the best talks. This time she starts a conference with a keynote The Power of the Agile Mindset.
Are you bored or struggling with estimates? Then Vasco Duarte will share with you his thoughts about #NoEstimates.

This year we have several speakers talking about scaling Agile and Scrum. One of the keynotes are from Dean Leffingwell and Scaled Agile Framework (SAFe).

We have speakers talking about Agile Architecture, Kevlin Henney is having a pre-conference tutorial on that topic. We have several speakers sharing their experiences about development, testing and practical case-studies.

And last but not least, we have several short workshops in the afternoons, for example my workshop – Coaching starter for Scrum Masters, a networking party where you can discuss and meet other people and an open jam session where you can share ideas.

See more details about speakers and talks.

Over self-organized teams

Self organization is one of the key agile artefacts. It’s all about self-organization we tend to say. The team should decide. The idea is great. It says that people who are doing the work can solve 80% of their problems themselves. It also implies that the team itself in it’s mature stage can freely decide on their internal processes, meaning how they organize themselves to atchieve the given goal. There is nothing said about changing the goal, not even about changing the external arrangements (i.e. roles and responsibilities of people outside of the team). The self organization is supposed to come with responsibility in hand. It cannot exists without it. And it’s not any easy task.

But every good idea can be misused and so more and more often we are observing over self organized teams who just don’t get it. Such team believes they can decide on anything you can imagine, the managers should not interfere, they are useless and not allowed to visit or even observe team meetings. They are redundant to their opinion. The same usually happen to the Product Owners who are surprisingly not anymore the ones who decide on priorities and functionalities. Sometimes the same happen to Scrum Mastes as well, who are for such team not any team members so they are not allowed to do anything either. Very strange situation indeed. Despite on how different such teams are, they have usually one thing in common. They believe they are doing great, but the rest of the organization should change. And becaus the “self-organization” they are trying to force them to do so. And where is the responsibility? No, no, no, it’s not ourfauls, theya re bad…

I have a couple examples. We have played our new Tulming Travel game a couple of times, and almost always someone from the team suggested they as a self organized team decide on priorities as they don’t like their Product Owner decision. And surprisingly to us, the rest of the team agree with the argument. Yes, that’s right, we are self organized so stop to tell us what shall we do. Second example is comming from real kind of startup company. We had a team of 7 developpers, one scrum master and one business & marketing team of 5 people including Product Owner doing research at markets at South America. They’ve been all located at one spot in Europe, the business team was quite matured, they know what they need and why, they’ve been willing to explain all that all over again to teh development team so they understand real customer needs. However, the team somewhere read there is a scrum and self organization and there was no force to stop them beleive they can decide on whatever. So they in the name of self organization changed basic Scrum pracsices, and pushed their business people and Product Owners away as they “don’t need” them to decide what to do. And in addition they did the same to their manager, as they are now Scrum so he is not allowed to tell them what to do anymore. There was no chance to stop them as they didn’t listen at all. They believed they are great despite they never delivered the Sprint Backlog.

The last example is from a corporation where they decide to bring fresh air into their agility. Their coaches got a generaly good idea I guess, that the managers should let the team to be self ogranized, but instead of starting at both team and management level, they pushed to the other side and forbid them to do anything with the team. They are not allowed to enter team meetings, coach, facilitate, give feedback, nothing. They are supposed to be “eavel managers” who are not capable enough to do their roles in agile world. For some it might be true, but there is huge number of others who are now struggling to do their job. And that’s not what we wanted to atchieve with self organization. We need mutual trust. Within a team, and otside i.e. to the manager as well. Otherwise there is no agile company and no self organization either. Just some group of small kids trying to shout to others without any purpose.