Scaling Agile with Frameworks – a SAFe path to LeSS work?

Ok, I admit right away that am just fooling around with the names of the two known large scale agile frameworks in the title – SAFe is the Scaled Agile Framework by Dean Leffingwell and company, and LeSS is Large Scale Scrum by Bas Vodde and Craig Larman.

The first book I read about scaling agile was from Dean Leffingwell, in 2007. I met Bas Vodde at my first Scrum Gathering in Stockholm 2008, and after I heard him talk about feature teams, I bought the book about scaling agile by him and Craig Larman as well, and of course we tried to apply much of that during our agile transition at Siemens Healthcare SYNGO. The book came very helpful since it clarified many of the questions behind the simple mechanics of agile, e.g. which effects queues have on you development organization, or how should the organization structure adapt to agile, and which other topics or areas beyond R&D are concerned.

In the last couple of months various people kept asking me about “Should we apply a large-scale framework to our agile transition?” – and asking back I get the underlying questions, of course, will we be faster, safer, cheaper, avoid start-up problems, will we have to read less stuff and discuss less if there are solutions available online and even people rolling out the one or the other framework licensed by the respective framework’s prophet(s)?

OK, let’s see if they themselves come with a warning message, like the drugs do in Germany, a little paper full of text that is in each box.

 

LeSS

„LeSS is not “new and improved Scrum.” Rather, it is regular Scrum, an empirical process framework in which you can inspect and adapt to any method and work in a group of any size. Large-scale Scrum is a set of additional rules and the set of tips that we have seen work in large multi-team, multisite, and offshore agile development initiatives. These tips are experiments to try in the context of the classic Scrum framework.“ (http://less.works/less/principles/large_scale_scrum_is_scrum.html)

 

SAFe

The Scaled Agile Framework (SAFe) is a proven knowledge base for implementing agile practices at enterprise scale.” “We, the contributors, offer this web version to the market in the hope that it will help all software development practitioners and team – as well as their employers – enjoz the satisfaction that comes from delivering ever-higher quality software at ever-faster speed” (http://www.scaledagileframework.com/about/)

OK, sounds good, but how can people start using it?

Now I switch to the “Implementation” page of SAFe: (http://www.scaledagileframework.com/implementing/)

Implementation is done in three steps, which come with lots of training and certification with trained and certified SAFe trainers: step 1 Change Agents, step 2 Management & Executives, step 3 teams and agile release trains – challenging, in 5 days per train, 8-12 Scrum teams per release train. Maybe it is intended for different kinds of enterprises than those I know, which have a lot of legacy to reflect about: structures, processes, products, codebases… and include properly into their new setup, so that they can go on delivering something to their customers.

 

So what do the LeSS guys say about implementation?

“These principles are crucial to an organizational LeSS adoption:

  • deep and narrow over broad and shallow
  • top-down and bottom-up
  • use volunteering”

“Prefer applying LeSS in one product really well over applying LeSS in many groups poorly.

Focus LeSS adoption effort on one product group, give them all the support they need, and ensure they work really well. This minimizes risk and if you fail it triggers a focused learning opportunity. And when you succeed it creates a positive “word on the floor” that’s vital nourishment for further adoption.”

And so on. I could not have phrased it better.

It reminds me of what a manager from another big company reported last year: lean and agile transitions worked great in their organizations which had self-chosen to transition with a mixed bottom-up/top-down approach, and where there were convinced local coaches. It did not work at all in organizations that were made “lean” through a big top-down rollout.

Frankly, I cannot imagine that any internal coaches who have neither seen nor done any agile work themselves, in product development teams on the ground, can be of much use in a large-scale rollout. So you need to start small, experiment, learn your own lessons.

By the way: yes, the LeSS guys are also offering trainings and even certifications. I would recommend to send people there who are already such local agile practitioners and have either worked in an agile team themselves for at least one year, or have been active participants in a transition team at smaller scale for at least the same time, and take the course in order to know more and improve, or roll out to a larger part of their organizations.

Don’t get me wrong: I think the SAFe framework can also a great source of inspiration and hints for solutions for organizations going agile in small or large-scale. I remember that we had a problem with making the product owner team feel responsible for prioritizing technical debt and technical improvements high enough, and we found a good example in SAFe how this can be solved: by agreeing on a percentage of the organization’s velocity dedicated to technical improvements by mutual agreement between architects and product owners, and then prioritization of actual technical backlog items by the architects.

Then only thing that I am very convinced about is that you cannot buy a large-scale agile transition, you do have to do it yourselves.

Self-organizing companies start to shape the world differently

The book „Reinventing Organizations“ by Frederic Laloux is a very special kind of economy book: it talks about self-organization in modern companies. This is a very hip topic today and one of the mega trends of the current decade.

reinventingOrganizationsInterestingly the author gives first the sociological and historical context of mankind since Stone Age. He relates to each context the types of organizations that have been possible and that have mainly existed in this context: their basic assumptions about the world and about human beings, their ability to plan, to scale, to deal with complexity and to react to unexpected events. During all these ages, it was possible that organizations with totally different value systems and organization level co-exist and compete, and today they do as well. Now in the last couple of decades organizations emerge that are only possible because of the complex experience in the modern world, and that are able to self-organize on team level and organization level and evolve their own organization in very short cycles. He gives all these organizations roughly spectral colors to distinguish them.

In the context of companies, the following colors apply:

  • The amber organization is ordered hierarchically, and the hierarchy level of a person is already clear from the beginning. This is traditional school, army, church, and the start of industrialization.
  • In the orange organization, there is as well hierarchy, but a person can grow if he/she meets the goals that the company has set for him, and can earn more. These companies see mainly selfishness as driver for doing better and receiving a bonus. Both orange and amber companies have a “machine” model of the company.
  • The green organizations give everybody the same, rely on consensus, and want to do good things for the world. These are mainly social organizations, they can also be companies like the famous Berlin newspaper “taz”.Only good people can work in such organizations.

Now the new evolutionary-teal organizations have a fundamentally different view of the world and the human being: they trust in the common sense and the social sensitiveness of the individual, without postulating that everybody needs to be similar or even have similar needs and wishes. They trust people to collaborate across different goals and needs, and to bring all this into a common context that makes sense for all of them. They trust people to make different contributions and they trust the team to evaluate these. Human beings in a teal organization are trusted to be able to learn nearly without limits, going beyond job titles and role names in self-organizing cross-functional teams, learn about their customers and their needs, and improve the ability of their company to fulfill these needs. Seeing the company as a growing, living organism helps them to take much more good decisions than in the machine context.

This is no speculation – the author gives lots of examples from the work life of more than a dozen very different companies, here are four examples:

  • Buurtzong – a social company for elder care in the Netherlands with 7000 people
  • FAVI – a producer of special metal parts in France with 500 people
  • Morning Star – a food processing company from the US with 400-2400 people depending on the season
  • Sun Hydraulics – a global company for hydraulic components with 900 people

All these organizations, however different in location, purpose and size, have solved a couple of fundamental problems in similar ways.

  • Their basic structure is team oriented, with teams that are self-organizing and self-administering.
  • The teams care about hiring, pay and personal feedback
  • There are no middle managers, mainly there are only coaches, or they elect project managers for a certain project for a limited time
  • There is a high transparency of all information for everybody
  • Decision taking is not based on consensus, rather decisions can be driven by an individual after consulting, listening to and considering input from all relevant colleagues
  • A basic principle is trust instead of control – any adult can be trusted to take reasonable decisions
  • Most of the companies have an explicit value system that they have created with the members, and which new members will learn in trainings
  • Important training topics are communication, collaboration and handling conflict

In an earlier blog post I wrote about the Brazilian company Semco SA which started already beginning of the 1980 to transform into a self-organizing company. It does also a lot of things similarly. All of these companies are also very successful compared to competitors in their area, in growth, margin, and customer satisfaction. It also seems that they have a much higher ability to manage crisis and adapt to changing markets and situations than traditional hierarchical companies. So it can happen that these companies will be the most important actors in the 21st century’s markets. It would at least not be a surprise to me, because in the same way as the modern democracies are more able to sense things happening and react to them than it was the Soviet Union with their top-down 5-years-plans, a self-organized company relies on the senses, minds and hearts of many people, not of a few.

HenrikKniberg-CultureOverProcess-OrganicStructureStrangely, no software companies are mentioned in Frederic Laloux’ book. – Looking to the area of software development, the combination of self-organizing company culture with agile methods and lean startup principles seems to be the most successful in the last couple of years. In principle, lean startup and agile companies have the self-organizing company culture in their DNA. However, it can happen that it gets lost when the company grows, or the CEOs do not have a real good value system. An interesting case of a rapidly growing company with a genuine agile culture is Spotify – at least according to their agile coach Henrik Kniberg. In this presentation (video here, and the slides) he is talking about their agile company culture. There are many others said to fall into this category, like Google, Twitter, salesforce.com, Dropbox, Netflix, Streetspotr – at least all of them are using agile and lean startup methods from the beginning, yet it is not guaranteed that all these companies have also a self-organizing company culture in the organization above and around the teams. But there are much more of them than we know, and I am sure that soon there will be much more of these companies around, looking forward to it!

Large Scale Agile at XP2013 Vienna – exchanging knowledge at a great conference!

From my perspective, the 14th International Conference on Agile Software Development XP2013 in Vienna was a great success. It took me to another level of confidence about what is needed to create and sustain a large scale agile organization.

The XP conferences are traditionally about programming and testing in an agile – XP – way, and organizing the team so that it supports XP practices. But they have grown into a conference that also covers  product ownership and design, leading agile teams and organizations, and even extending agile to the rest of the organization – this is agile real life in the industry.  What I like very much at XP conferences in general is the good mixture of experienced agile people from industry, some very well known consultants, and a lot of academic researchers who also are working close to industry about agile topics.

image

(Photo: Hubert Baumeister)

On the first day I got absorbed by the Executives & Managers Track with firsthand experience from four different companies: ABB, Ericsson, Johnson Controls and Nokia. Most valuable! The managers were speaking openly about what it takes to get agile, how their company transformation programs took one step after the other to establish agile on all levels. And this is still an unfinished journey, but it has a clear north. An important point from the talk of Hendrik Esser, who is Head of Portfolio and Technology Management at Ericsson, is

To embrace change, you have to change 3 things together
· Culture
· Practices and process
· Structure

You should never see agile as a process only, this will inevitably lead to failure. The most important cultural foundation for the agile transformation is building trust, and on top of the trust you can build transparency.

These three important transformations were mentioned by Per Branger from ABB in Sweden, but are basically identical to what Ericsson is doing:
· Continuous Portfolio Management
· Continuous Release Management
· Continuous Development

When they noticed at ABB that they are really a big software developing company, coming from a background of electrical engineering, they launched a massive knowledge and skills improvement program. The remarkable thing is that they measure progress by self-assessment of every developer against description of expected skills, and that the training comes in small portions and by self-assignment as well. Wiki based knowledge bases and Q&A tools that remind a bit of the famous stackoverflow website support the learning as well.  “Carrots, no sticks” opens also the path to using common tools.

Gregory Yon is Agile Coach at Johnson Controls and talked about how he is extending agile into the rest of the organization, to the non-development teams as well as convincing different levels of management from the agile values and needs. From him as well as from the other managers I learned that we can never communicate too much to higher management about the advantages of agile, and that we need  to measure things and compare with previous projects to show how we have advanced since the old times of waterfall.

Jorgen Hesselberg, Senior Manager, Enterprise Agile Transformation, Nokia (Chicago US) explained how they are using on all levels Agile Working Groups, mixed from management and project roles, to start and sustain agile transitions at each Business Unit, and keep them up and to extend agile to the whole company. The positive results of agile on the results as well as on the employee motivation are impressive.

My own presentation on our Distributed Product Owner Team for an Agile Medical Development was on the second day, and I loved the discussion with a couple of other speakers and participants about the needed knowledge and skills for this role, and the needs for communication in the PO team and with teams and customers. I have also learned that there is actually an open group of experts from industry and research with the goal to foster software product management excellence across industries, the International Software Product Management Association (ISPMA) , who are interested in collecting such practical experience, and are creating resources for professional software product management training.

At the academic track, I found a couple of presentations on the last day very interesting: A research paper by Jeanette Heidenberg, Max Weijola, Kirsi Mikkonen, and Ivan Porres – A Metrics Model to Measure the Impact of an Agile Transformation in Large Software Development Organizations asks whether an agile transformation was worth the effort. For this, they were looking for metrics that support agile values, focus on the whole organization, not individual or teams, and are applicable to both waterfall and agile projects. Also they should be feasible to collect for past and ongoing projects, in any size of project, and be objective and clear. They did an iterative approach with first formulating the goal, then ask practitioners for metrics used, compare them against their values and goals, and finally select a collection of metrics. They have some really helpful metrics that we can apply to learn how much we have already improved through agile, at least in some aspects.

The paper Continuous Release Planning in a Large-Scale Scrum Development Organization at Ericsson by Ville Heikkilä, Maria Paasivaara, Casper Lassenius, and Christian Engblom was complementing very well with the talk of Hendrik Esser, as they are exactly describing how the release planning for individual features works, and which experiences people at Ericsson had with this method.

Of course there were also great keynote speakers at XP2013 in Vienna, a helpful Open Space, a wonderful conference reception, and great conversations in the breaks.

image

As always, thank you very much for the photos, Hubert Baumeister.

The next XP conference will take place in May 2014 in Rome, which is also a nice place to go to.  I will convince some of our colleagues and managers of submitting a talk and participating – we have a lot of experience we can share!

Best Regards
Andrea

Slides for ScanDev 2013 DONE: distributed product owner team – strategies

One week to go for this year’s ScanDev conference in Gothenburg, and slides are done.

image

Topics covered are mainly:
Strategies for growing a distributed product owner team, when the problem space is very different from the normal experience world of a software developer. Strategies for communication and customer collaboration in the distributed setup.

For preparation, I made a lot of personal interviews with product owners on different hierarchy levels and multiple sites. I learned a lot about the reality behind our official agile framework, and how much the real success depends on people and their skills and goals.

To convert my knowledge into an interesting presentation I used a marvellous book, a classic: Presenting to Win: The Art of Telling Your Story by Jerry Weissman. What is very helpful in this book is that he tells you a lot of stories to make sure you know what you want to tell, to whom, and how you will bring this across to your audience. So I started to create my presentation on a lonely day at home, using hundreds of sticky notes on my personal whiteboard, instead of bothering with Outlook templates or -beware!- assistants. Then I selected a structure for the slides that would allow me to tell my story, and then I created each single slide on a sheet of paper, put them into the structure, re-ordered and finally bright them to the office.

image

There, of course, I had to take a corporate template, and also picked a lot of photos, that would this time nicely fit with the topic. As structure I had selected a pyramid. I created it with corporate colors, and let it build up throughout the slides.

At the beginning of this week, it was finished, and I submitted it for review and release. And -wow! -I got it released after considering a few remarks, in less than two days. I can say, the release process got much more agile since I first used it for an external presentation four years ago, when it still took me four weeks, but obviously as well my own professionalism in creating publications has increased a lot since then.

Speaking at ScanDev 2013 about our Distributed Product Owner Team @ Syngo.via

I am speaking at SCANDEV 2013

This week I received the news that my new talk has been accepted for SCANDEV 2013 on March 4th to 5th in Göteborg, Sweden. The title is „Distributed Product Owner Team @ Syngo.via“

What is it all about?

We are developing medical imaging and workflow software in an agile way with development teams distributed to several countries. One of the major challenges is how to set up and communicate within the Product Owner team.

  • There we have to deal with the distribution, e.g., have the Product Owner either onsite with her peers or with her Scrum team, travelling, or with proxy.
  • We need people who are good in two different fields of knowledge: medical and software development.
  • As a third issues, the environment of the customers may be different in different countries.
We have ramped up local Product Owners in different countries, have found local collaboration customers, and have developed  a set of communication channels and workshops how to synchronize Product Owners in the team, share a common vision and backlog with their Scrum teams, and collaborate with customers locally and globally.
While I am putting my presentation together, I would be interested very much in what my customers want to know about. You can ask me questions, and whatever is feasible I will include into my talk. So Product Ownership is being directly applied.

Agile is about the people!

Sometimes I see blog posts or LinkedIn discussions where some agile specialist or special agilist reasons about the exact amount of minutes to spend on each Scrum meeting according to sprint length, or he asks how he can measure and compare the productivity of teams.
Often I have the impression if I asked him now “How is the motivation of your teams? How do they feel, are they happy or frustrated? Do they have all the tools to get their job done? When did you last time clear some organizational obstacles out of the way for them? Have you hired the right people, and given them lots of opportunities to learn?” or anything alike, this kind of specialist would stare at me without getting the issue.
Actually, this is the centre of the universe – is the people, not the final twisting and tuning of an ideal Scrum process. Scrum is about the people!
Even the managers I know, some managers who have done so incredibly many things right, – even they, at rare times, of course, come up with something like “but now if I would like to compare the productivity of two teams, how would I do it?” and they think very hard about it. 
 
Then I ask them: Have you invested already enough time into how they can learn something new about the product domain, stay up to date with their architecture and programming knowledge, know how to use the latest tools you have bought them? Have you spent a similar amount of time on how to sharpen the product vision with the Product Owner, so that he can really make developers happy working on this marvellous project? If I wake up a Scrum team member in the middle of the night, would he or she be able to tell me the product vision? Have you thought about organizing a FedEx day so they can try out new features of their favourite programming language just for fun, and get fresh motivation for a couple of sprints?

This is what I ask my managers in such cases. Metrics and measurements are fine, but you always need to seed before you harvest.
Hey, it’s about the people!

An incredibly intensive 1st ALE2011 Unconference – Agile and Lean Europe in Berlin

 (Olaf Lewitz, one of the central organizers, on a photo by Marcin Floryan)
In February, 2011, the Agile and Lean Europe network was founded in LinkedIn by Jurgen Appelo, Author of Management 3.0. He asked the European lean and agile practitioners and communicators to join. We were about 1000 members within a month, and there were lots of active interesting discussions, “bathtub conferences” and many ideas how to collaborate more closely. The first real-life meeting happened at XP2011 in May in Madrid. Since then, 47 people with a vision created the best and most intense (un)conference I have ever attended. My own role was to be part of the “industry sofa” – we had “sofas” instead of “chairs.” I spent some of my free time reviewing abstracts and finding out whether these people were good speakers, if I had not heard them speak before. The final result we composed is amazing – http://ale2011.eu/speakers/ .
audience!

    The audience was amazing

All further work was organized via real-time collaboration tools: Skype, Basecamp, GoogleDocs, Mindmeister, Twitter, and Conftool. Real-time mostly meant evenings, sometimes even weekends. The final event structure included one keynote every day – with Rachel Davies, Bjarte Bogsnes and David Snowden, we had three highly interesting speakers, two of which are not from the software world, but are teachers of lean concepts for management.  Each day started with a funny coding dojo warm up, followed by 30-minute talks in the morning, lightning talks after lunch, and Open Space all afternoon. Virtually everybody participated actively in something: more than 220 people from at least 27 European countries. Talk topics ranged from “Software Craftmanship” and “Metrics in a complex world” to “How to change the world.”

Bjarte Bogsnes from Statoil, Norway
For me, Bjarte Bogsnes with his “Beyond Budgeting” talk was most inspiring (http://ale2011.eu/2011/09/17/ale2011-keynote-bjarte-bogsnes-on-beyond-budgeting/ ), but… yes, but… the 7 levels of hierarchy between me and the CEO of our company make me think that this is not the easiest thing to put into practice by myself. Fortunately, many other talks also had inspiring contents!
From Henri Kivioja from Ericsson, Finland, I learned how we can guide managers to practice go and see with the Scrum teams: they just got rid of all kind of upward reporting from project to line managers. They also reduced their full test cycle dramatically, from about 1 year for the whole system (100%) to about 1 week for 90%.
dojo2Eva Kisonova and Sabine Canditt presented a funny game of cultural differences they have practiced with our Scrum teams in Slovakia. It showed the stereotypes that may exist on both sides, which can make collaboration difficult if the teams have not reflected on them. Putting it into practice in a small example among the participants was really fun.
Rob van Lanen explained why and how they had realized FedExDays with his company’s developers in the Netherlands. This is a 24-hour slot, in which the developers can develop whatever they want – the only condition is that they must present it after that time. The department provides food and drinks, and the CEO is present at the demo at the end. The participants created 4 products, a traffic light tool for the software build, and a gaming application. They self-organized to do Scrum in one-hour slots and even pair programming. It was a great motivational boost for the teams.
Claudio Perrone gave an excellent introduction to A3 and Kaizen, which can actually be understood when you look at this outstanding presentation: http://www.slideshare.net/cperrone/a3-kaizen-heres-how. In this way, continuous improvement can be introduced on all levels: in the project team and on organizational level with the managers. This is something we should put more emphasis on soon.
Torsten Kalnin explained how the Wikispeed team builds modular speed cars using very little fuel with lean and agile virtual collaboration of volunteers around the world – an amazing example for agile hardware development – see also at http://www.techflash.com/seattle/2011/02/the-future-of-cars-and-everything.html.
I visited a few more talks related to big agile transitions, offshore and distributed experiences, which we later followed up with discussions in the Open Space.
Open Space facilitated by Mike Sutton
The most intense part of the unconference was surely the Open Space sessions: everybody posted his/her topics at a common marketplace, and there were a lot of different spaces in the venue where we could start discussing around a flipchart. On the first day, I proposed a talk about organizational impediments, to get stories of what happened and how people actually resolved them.  Later, I was in another big agile transition discussion, and there I met a couple of people who also used communities of practice in their companies. So I had my topic for the next day: how to get CoPs going, and how to keep them alive in their original sense, as a means for knowledge acquisition, best practice exchange, and as a catalyst for improvements.
You can find almost all references from the conference in two places: http://ale2011.eu/resources/ and with #ALE2011 on Twitter. A lot of lean and agile conferences in the next future will be powered by the spirit of the ALE-Network, I am sure! 
DSC_6380 ivanKostial JurgenAppelo respect4people
Andrea Heck