The failureship and turkeys voting for Christmas.

In some organisations, the individual succeeds if the organisation as a whole is a success. This is often where the organisation is at risk, either a start up, an organisation in trouble, or continuity requires achieving success. In these organisations, success is only possible if all the individuals align their goals with the goals of the organisation. It is not possible for an individual to succeed if the organisation fails. A “risk managed” culture will normally emerge in this context.

In some organisation’s where continuity is guaranteed, the success of the individual is best achieved by out performing their colleagues. The individual succeeds by placing their own goals above those of the organisation. A “failure” culture will normally emerge in this context if the leadership do not act to prevent it.

Puzzled by their colleague’s strange behaviour in December.

In a “risk managed” culture, adoption of new values and practices will be natural and easy. In a “failure” culture, adoption of new values and practices will only occur it aligns with the goals of the individual, something that leadership has a huge impact upon.

The failureship in a failure culture is dominated by self interest. Even if the CEO is trying to improve the share price, they do not necessarily care about the long term future and health of the organisation. The goal is short term value extraction rather than long term growth. I worked at an organisation that was deeply damaged by the financial crisis and specifically the mismanagement of government. Prior to the crisis the CEO had been leveraging the strong national brand to build a trusted global brand. This was a slow and complex process of investment that would take years to evolve this centuries old institution to it next incarnation. The Crisis eventually lead to a new CEO. The new CEO retrenched and shifted to a short term national strategy and sold off the international assets to make the short term profit look good. This strategy distracted from his poor revenue figures. The CEO got their bonus at the expense of the future of the organisation.

Why does this self interest impact the quality of the developers that an organisation hires?

Consider two organisations. The first organisation hires the cheapest staff that they can find. They normally do this through a consultancy which is also driven by self interest. More later on that. The second organisation hires staff who cost twice as much but they need half the amount. The first organisation has two thousand (2,000) developers costing one (1) unit, and the second organisation has one thousand developers costing two (2) units each. Both organisations spend the same amount two thousand units (2,000) on developers. The organisations look like the following.

Although the cost of the workers is the same, the management headcount and the the cost of management have more than halved. Furthermore, if the workers are significantly more experienced, they will require much less support from management which means that those managers in Organisation 2 that have an outline could also potentially be removed. So the final result is a reduction in management headcount in the region of 50% to 75%. So if you want to understand why managers in a failure culture do not want to increase the quality of the workers, you only have to ask “Why don’t Turkeys vote for Christmas”. In a risk managed culture, management is focused on value rather than cost. The desire to move fast means that they want smaller teams of higher quality individuals.

In a failure culture, the failureship does not want to be held accountable for the quality of the workers. As a result, they do not hire directly but instead make extensive use of consultancies with a “global brand”. This also has the added advantage of reducing the quality of the workers even further which increases the demand for poor quality managers. Any decent worker can normally find their own work, either as a permanent or temporary member of staff. People join consultancies because they want to gain experience either in a skill set or in an industry. This is particularly true of graduates (that includes me) who want to get some experience in a variety of industries before deciding where they want to work. Whereas a permanent or contract individual receives 100% of the money paid by the organisation, a consultant from a “global brand” consultancy receives only 30% of the money paid for them. In effect the individual consultant is paying 70% of their fees to the “global brand” consultancy. They are paying that 70% fee for the learning experience that will eventually enable them to receive 100% of the fee. So why would the failureship pay such a huge premium for such poor quality staff? We have to remember that the motivation is self interest rather than the goals of the organisation. “Global Brand” consultancies look after their customers. In the event that a member of the failureship is fired, the “Global Brand” consultancy is incentivised to find that person another job, ideally with more authority as they are more likely to use the services again at their new organisation. Some consultancies even act as car parks for members of the failureship. Once again, asking a member of the failureship to move away from “Global Brand” consultancies to hiring competent individuals is like asking someone who is falling from an airplane to hand over their parachute.

My favourite experience of this nature was a meeting with a senior executive about Safe.

  • Me: It is my strongest recommendation that we do not use SAFE because it is rubbish.
  • Exec: No! We are going to use SAFE!
  • Me: Why?
  • Exec: Because I say so.
  • Me: In that case let me introduce you to Martin (I still miss you Major Tom) who knows all the people with lots of experience with SAFE.
  • Exec: No! It is more than SAFE so we are going to use the “Global Brand” Consultancy called *********
  • Me: Why?
  • Exec: My best friend is a partner with *********. He finishes his paternity leave next month so I want him to do it.
  • Me: Is that the door? Must be going.

In this case it wasn’t even pure self interest, the executive placed the needs of his friend above the needs of the organisation… Though it always helps to keep the “Global Brand” happy in case you need them. 😉

Leaders in a “risk managed” culture would ensure the quality of each individual. They would invest in continual pipeline of high quality candidates preferring to find talent over filling roles.

A large organisation of poor quality individuals in a failure culture gives more opportunity to “leaders” in its failureship. Using “Global Brand” consultancies provides further job security.


Failureship and EXPENSIVE, low cost people

One of the key differences between organisations with “risk managed” and “failure” cultures is the attitude towards talent. “Failure” cultures prefer large teams of “cheap” individuals who have little experience, knowledge and skill. “Failure” cultures prefer individuals with potential. “Risk managed” cultures prefer small teams of “expensive” individuals who have significant experience, knowledge and skill. “Risk managed” cultures prefer individuals with a track record.

When you discover the driver is inexperienced, its not an accidenture, its negligence.

One problem with “failure” cultures is that the failureship do not understand the difference between cheap and low cost. Consider two potential team members. One member costs one unit and delivers one piece of work. The second costs two units and delivers ten pieces of work. Which of the two individuals is cheaper? The second costs more than the first, however we need to consider the value delivered as well. The second delivers five pieces of work per unit of cost whereas the first delivers one piece of work per unit of cost. The second is cheaper than the first. One explanation is that the work or value delivered is subjective whereas the cost is fact based and defend-able if the investment fails to deliver the desired return.

“Failure” cultures gorge themselves on expensive low cost individuals whereas “Risk Managed” cultures carefully select cheap high cost individuals. I worked on a project where we had twenty developers based in London working closely with the users of the software. An executive blindly following the edict of another member of the failureship, and insisted we move the software development to an off-shore location with lower day rates, which was about one fifth of those in London (five units of cost in London, and one unit of cost for offshore developers). The London developers became product owners and fifty off-shore developers took over the development. The value delivered dropped like a stone to a quarter of previous levels ( The value delivered went from 600 to 150 ). The cost went from 100 units to 150 units. The cost of a unit of value delivered went from six ( 600 / 100 ) to one ( 150 / 150 ), an 83% reduction. The exercise was considered a success by the failureship because the average cost per developer went from five to one and a half. It is only possible to think this way if you completely ignore value.

BeforeAfterChange
Head count2070+250%
Value delivered600150-75%
Cost100150+50%
Cost per head count51.5-70%
Cost per unit of value delivered61-83%
Agile project lead (Me)10-100%

The video below provides a wonderful illustration of this difference in the cultures.

Even though there are 100 children playing football, only a few are ever close to the ball. What are the rest doing… nothing of value to the game. This reminds me of a transformation project I worked on. A well known consultancy provided a team of twenty individuals to work on the project. We told them that we wanted to focus on delivering one or two epics (i.e. Limit Work in Progress). Very quickly sixteen epics were work in progress and we did an intervention asking why so many epics were being worked on. It turned out that the team members did not have the skills to work on the epics we wanted and so they started a new epic when someone in the team could not work on any of the other epics. They bought in an agile expert. It turned out the agile expert’s most significant experience was sitting next to a team that was using Scrum.

Each “Risk Managed” organisation I have worked for has had a ridiculously high bar for entry. At ThoughtWorks and many other high quality software organisations there is a coding test where the candidate has to write some code and then explain their design decisions in a interview with experienced developers. Thoughtworks hired one out of every one hundred applicants. A tribe lead at Spotify said that they hire one in three hundred applicants. Any developer who does not want to do a coding test because they are a well known “rock star” should not be considered a good cultural fit. As soon as one person is considered “special”, where do you draw the line. A “rock star” developer should want to share the interview experience of their colleagues rather than be important enough to be exempted. A “rock star” should be able to complete the task in an hour whereas another candidate might take a few days to complete the task. The coding tests do not just test for capability, they also test for developer culture fit.

A “Failure” culture struggles to recruit high quality talent because it does not know what good looks like. It takes several high quality individuals to recruit someone of similar quality, because we all have blind spots and it takes several of us to make sense of someone’s experience. This is particularly the case as many people now know the “words of power” used by high quality developers such as “refactoring”, “clean code” and “patterns”. If you have never done refactoring, you might be impressed by the person who “refactored” an entire system without needing any automated tests because they are just so awesome. Once a “failure” culture develops a reputation for rewarding bad behaviour and tolerating mediocrity, high quality individuals will avoid it and no amount of recruitment effort will help. One of the easiest ways to identify a tolerance for mediocrity is to look for organisations that make extensive use of low cost and expensive consultancies like Wipro, Tata Consultancy Services and Accenture rather than apply their own recruitment processes to hire individuals as either permanent or contract staff.

“Failure” cultures find it even harder to retain high quality individuals than they do to recruit them in the first place. High quality individuals want to deliver value to the customer. They quickly become frustrated in cultures that are only interested in personal survival and promotion, where helping others solve problems is not considered sensible. Things that are easy to achieve in a “risk managed” become almost impossible to get done in a “failure” culture where everyone is too busy doing their day job to have any spare capacity to help someone else fix an organisational problem. Eventually high quality talent will leave, shifting the culture even further into a “failure” culture.

The flip side is that organisations that truly embrace hiring high quality talent get a reputation and grow rapidly through word of mouth. The London Agile Developer community descended on Thoughtworks, then BNP Paribas, UBS, Sky, Springer, rapidly growing high quality teams, simply because they knew they would be able to work with similar high quality developers and they would be valued by the organisation. A cautionary tale though, at one of the organisations the leadership drifted into a failureship and all the high quality developers left. The developers who remained “knew where the bodies were buried” and quickly achieved senior positions where they prevented the adoption of practices that undermined their status, thus effectively inoculating the organisation against “risk managed” culture.

As with all leadership, deeds need to match words, or leaders need to “Walk the Talk” as some books call it. Any transformation of an organisation needs to start with building a team of cheap, high cost, high quality individuals and developing a recruitment pipeline focused on excellence. Any organisation that tries to improve by hiring a large quantity of expensive, low cost individuals from a poor quality consultancy is choosing to fail.

If its so obvious that smaller teams of high quality individuals are better, why do the failureship continue to gorge themselves on EXPENSIVE, low cost consultants? I’ll answer that question in the next blog…


Step away from the office, and join the team!

Years ago I shared a table at a wedding with a primary school (Kindergarten) teacher. We were not surprised to discover that my job in technology was very similar to their job. Understanding behaviour was a big thing we had in common. One common situation they encountered was children behaving badly and the parent’s not believing them, “They are an angel at home” was the common response. The teacher explained that they would arrange for the parents to observe their child’s disruptive behaviour secretly through a window. I’m sure that often the parents would rather not know because then they would have to do something about it. And so it is in organisations. I’m sure we have all seen people behave in an aggressive and domineering manner when they perceive themselves to be the HIPPO in the room, only to assume the demeanor of a timid mouse in the presence of a bigger cat.

Photo by yang miao on Unsplash

Failure cultures prevent access to the big boss, the hierarchy acts to prevent uncontrolled access. This benefits the person potentially behaving badly, and it also benefits their boss because they do not become aware of the bad behaviour. People who tells tales out of school are often told “to escalate through the appropriate channels”. The “appropriate channels” being those that can be used to filter the message.

The leadership in “Risk Managed” cultures spend a huge amount of effort to make sure they are accessible to everyone in their organisation. As well as making themselves accessible, they will also make the effort to reach out to people. They understand that “going to the Gemba” is not a planned royal visit with courtiers and presentations, often it means simply working near to a team for a while. As any ethnographic researcher will tell you, it takes a few days for people to drop the formality and act normally. Spending a week sitting in an area, listening to the buzz of a floor will tell you much more about the mood of the people than any survey or presentation. Leaders are better able to detect issues earlier when they sitting with the team at the Gemba.

The failureship in a “Failure Culture” are like an unruly child in the kindergarten along with their irresponsible parents who do not even want to know about the behaviour of their star child. Uncontrolled access to the big boss is to be avoided at all costs. Meetings with the big boss will be arranged by respecting the hierarchy at all times. Your superior will attend the meeting, acting as a guide to ensure you do not reveal any information that might taint the ears of the big boss. Door stepping a senior member of the failureship with a problem / issue / idea will be met with sanctions of a career limiting nature. Individuals will be given controlled access to senior members of the failureship at intimate breakfast meetings and coffee corners with a dozen of your fellow colleagues, normally those with the same rank.

This leads to a behaviour I observed in the original “Seeing Culture” article. The failureship have offices, and the leadership sit with the team. This is perhaps one of the most visible differences between a “Risk Managed” and a “Failure” culture.

“Turn the ship around” by David Marquet tells the tale of how his submarine went from a leader – follower culture to a leader – leader culture. I prefer the language of Eric Berne’s Transaction Analysis of Parent – Child to Adult – Adult relationships. In a “failure” culture, a leader – follower (parent – child) relationship means you treat your subordinates like a child and control access to the big boss. Going around your superior and violating the hierarchy is normally a career limiting move. In a “risk managed” culture, an adult – adult relationship means you treat your subordinate as an adult and respect them to make the proper judgement as to whether they need to speak to the big boss. The big boss will determine whether the contact is appropriate and will advise your subordinate accordingly, perhaps directing the person to someone else or even back to you. If the contact is inappropriate, its a learning opportunity helping them understand how to address an issue, rather than a violation of the hierarchy.

I previously wrote about how Al-Noor Ramji and JP Rangaswami reduced the power distance index at Dresdner. At Skype, Mark Gillett was even more explicit at breaking down the power distance index. I first met Mark shortly after joining Skype in the coffee area next to his office, he introduced himself and when he knew I was an Agile Coach we discussed User Stories for a few minutes. A year or so later we moved office into a state of the art office as part of the Microsoft family. A couple of years ago Mark told me about his design for the new office. Microsoft had wanted him to have a corner office but he wanted to maximise the possibility for chance encounters. The Skype office was based on three floors. Mark installed a sweeping spiral staircase to link the floors. The staircase was the easiest and quickest route between the floors and as intended by Mark, it was a great way to bump into people. Furthermore Mark installed a coffee area on the middle floor next to the staircase, complete with a professional Gaggia coffee machine with associated Barista Training. Mark’s office was situated right next to the coffee area. If you really needed to chat to Mark about something, all you needed to do was go for a coffee and hang around for a short while.

At Tesco, Jim Banister was entitled to an office. Instead of sitting in his office unable to hear what was going on, he sat out with his teams. His office was made available to his teams for meetings that were either private or would be loud and disruptive to everyone else.

By comparison I worked in a company where senior managers were entitled to an office… and they always used one. They even used an office when they visited other locations. It was common practice for them to use the building where they could have an office rather than sit in the building where their teams sat. Team members would have to travel to a different building to meet the manager visiting from an overseas location. The same company further reduced the chance of unintended access by locating senior management in separate buildings to the workers. Even in open offices with hot desking arrangements, the failureship will soon establish themselves with dedicated seating for senior staff members and reserved seating for important individuals. Nothing signals a flattening of the hierarchy than seeing a senior executive hunting around for a desk because they arrived late, and them sitting with a new group of people, or perhaps in the coffee area.

At one company I worked the executive had an office in London. They only spent one or two weeks a year in London. However, they had an executive assistant sitting at a desk outside the office with instructions that no one was to use the office. Furthermore, they had an additional window added to improve the light to the office that no one was allowed to use. As one colleague commented “I wish they would fix the toilets so that they did not flood instead of improving an office they never use.”

When I did some work at Bank of America, a number of people excitedly told me the same story. A developer was working on some code when a new joiner asked them if they could join them and do some pairing. After an hour of coding the new joiner excused themselves. The developer asked the new joiner what his job was “I’m the new Group CIO”. That story spread through the organisation like wild fire, that the head of a department of tens of thousands of people was approachable and could code!

Along with “Servant Leadership”, “Go to the Gemba” is possibly the most widely held expectation of Agile Leadership. The first rule in “Six Simple Rules” by Yves Morieux of Boston Consulting Group is “Understand What Your People Do” or “Go to the Gemba”. “Go to the Gemba” and “Sit in an office” are the easiest behaviours to spot to identify whether you have leadership or failureship, which is why they are so powerful. Nothing signals change to an organisation better than a senior member reaching out and connecting to real people in the organisation, rather than simply communicating through the hierarchy (or through McKinsey if they do not trust their hierarchy).

If you are genuinely interested in moving from failureship to leadership, move out of your office and turn it back into a team meeting room. Sit with the team but give priority to them, offering to give up your seat if someone else would benefit from it more. Move around your organisation, spending a few days with different sets of teams. If you do need to spend time with your colleagues in the leadership team, split your day between them and the real workers. You will discover so much and you will enjoy your job more.


aacennprrsTy and the failureship

I ran an exercise where I asked a group of executives “What is transparency?”. Take a moment, think what it means to you and say it out loud before you read the answer I gave.

Transparency is when you give someone the information they need, in the format they need so that they can make the decisions they need to make to fulfill their responsibilities.

Transparency is a collaboration. Sometimes the person asks for the information they need in the format they need. Sometimes the person with the information decides that another person needs it and works out the best format to present it. Formatting the data is often about constructing the context so that the information has meaning.

Transparency is not open kimono, the sharing of all information. It is not drinking from the fire hose, receiving huge quantities of data to be sifted and sorted. It is not the pretense of sharing everything, expecting the receiver to find Transparency in the aacennprrsTy. Transparency is the sharing of information with context to allow everyone in the organisation to achieve their goals. Sometimes the information is focused on an individual or a small group, and sometimes it is shared with the entire organisation. Sometimes it takes a great deal of effort to help someone understand that they need the information you are providing.

In “risk managed” cultures, the leadership create an environment where everyone, including themselves, share information so that others can decide and act. In “failure” cultures, the failureship create an environment where information is carefully controlled, and sharing of information can be seen as a crime against the organisation. The failureship do not want information to be freely shared as they want plausible deniability in the case where they fail to fulfill their responsibilities. Speaking truth in a “failure” culture can be career limiting or career destroying.

In “failure” cultures, information is carefully curated and filtered, often translated through several excel sheets, powerpoint decks and summaries until it presents the story that subordinates want the failureship to read… which is the story that the failureship want to read. Drilling into the information requires the failureship to navigate the reverse trail through the organisation, offering the opportunity to correct or delay access to the truth. In a “risk managed” culture, information is automatically assembled to be always available for all concerned, with alerts when appropriate. The information acts as an “invitation to the Gemba” for the leaders in the organisation. “Invitation to the Gemba” is a phrase that Jabe Bloom introduced to indicate that the leadership should not act on the information they receive but rather use it to identify the places where they should “Go to the gemba”, go to where the work is done, and see for themselves what is actually happening.

In a risk managed organisation, the leadership create the culture and environment where everyone is encouraged to provide information to those that need it. They actively work to encourage people to share information up, down and across the organisation. This requires a great deal of work on the part of leaders as they will constantly be employing new people who are used to cultures where sharing of information is forbidden, discouraged, or where information is “power”. Leaders not only send out memos or give organisation wide speeches on the subject, they also encourage it in every interaction or meeting that they attend.

All leaders ask for transparency, they all want to know what is going on. It is their behaviour when they receive information that determines how willing people are to share it. Getting people to share information requires the leadership to build trust. A second question I ask executives is “What do the RAG statuses, Red, Amber and Green” mean to you?”. The traditional meanings are:

  • Green – Everything is OK
  • Amber – The project might be late
  • Red – The project will be late – (actually, it means someone has revealed the truth about what is going on.)

Anyone who has worked in a failure culture knows what happens when someone reports a status with “Amber” and “Red”. It is a very unpleasant process for all involved, the failureship swoop in, possibly deploying someone from outside of the project to find out what is going on, even though the failureship already know. A lot of weight is thrown around with an associated dollop of Kabuki (Japanese theatre with the players cutting all sorts of shapes on the stage) until some sort of reset is performed and the status can return to “Green”. The reset normally involves moving out the target date if it can be negotiated, or reducing the scope until it no longer delivers any real value for the customer. Normality resumes until the next RAG crisis. Everyone knows that nothing good comes from reporting status of “Amber” or “Red”. The failureship lament that most projects are “water melons”. Green on the outside, and red on the inside. They lament that they normally only find out that a project is “Red” when it is too late and they can do nothing about it. They already know what is going on. They find out it is too late because they want to find out too late and make it unpleasant for those that tell them. They do not find out too late as they already know, they are forced to acknowledge the problem publicly when the options to address it have expired.

In a “Risk Managed” culture, the RAG status have a different meaning:

  • Green – We are good, and are unaware of any problems that will prevent success.
  • Amber – We might need help. There are problems that we might need help addressing. We are signalling that the leadership may need to step in and help.
  • Red – We need help. There are problems that we are unable to address and require the support of leadership to address them.

In a “Risk Managed” Culture, the responsibility for delivery remains with the team, including the responsibility for requesting help. No team should be green until they have completed their first end to end production delivery of value. In a “Risk Managed” culture, the leadership act as servant leaders to the team. The leadership work hard to build trust with the team so that there is no barrier to revealing the real status of the delivery. The leadership see it as their responsibility to help the teams. The crime in a “Risk Managed” culture is to sit on information that you know others need.

Transparency is not about systems and reports, it is about how the leadership behave when they encounter transparency or the lack of it.


Failure Cultures Reward failure.

For over twenty years I have been reciting the mantra:

“In agile, we fail fast to win early. Traditional teams are afraid to fail!”

For many years I have been observing agile teams and non agile teams, and their attitude towards failure. Once I started to observe them using the lens of “risk averse” and “risk managed” culture I realised that the agile mantra about failure was wrong.

Photo by Caleb Woods on Unsplash

In agile teams, practitioners do not fail fast to win early, or fail fast to learn. In agile teams, they act with self discipline as a single team to deliver value in small increments. At regular increments, the team and/or team of teams come together in a retrospective to identify what is going well and what should be improved. Agile teams do not focus on failure, they focus on value delivery and continual improvement. When failure occurs, agile teams acknowledge the failure so that they can learn from it and improve.

It is not the goal or focus of an agile team to fail or learn, it is the focus of an agile team to deliver value.

In addition, traditional team are NOT afraid to fail!

Teams in “Failure Cultures” know that their managers in the failureship are even more scared of failure than they are. As a result, their managers will never acknowledge failure. The result of this is that the failureship will REWARD failure to prevent anyone knowing that a failure has occurred. The bigger the failure, the bigger the reward. This inability to acknowledge failure and the subsequent rewarding of failure is why I call this “Failure Culture”.

Career progression in a failure culture is based on avoiding failure rather than being successful. If your portfolio of interest contains a large strategic project, it has to be hailed as a success regardless of whether it is a failure or a success. If you have a large strategic project in your portfolio and the key individual(s) are not rewarded for their contribution, people may question why they have not been rewarded. So if you are in a failure culture, the easiest way to get promoted is to join a large strategic project in a key role. Once you are established in the role, your promotion is guaranteed regardless of whether you are a success or failure. In fact, the bigger the failure, the bigger the chance of promotion.

I worked on a project that was meant to be an Agile Flagship. As an Agile Coach I gave it my full attention. The business users and product team adopted the agile approach with gusto, even reorganizing to sit together in the same space within earshot of each other. The technology team was a different matter. The technology lead refused to engage and turned a blind eye to his lead developer’s misogynistic and bullying behaviour. Eventually the team refused to even listen to suggestions about how they could improve the way they worked. “The developers are working sixteen hours a day, seven days a week. They do not have time to work in new ways”. The senior executives were informed of the situation and responded by saying “That’s not Agile, Its not even Waterfall, Its a death march project!”. The first production delivery on the Agile Flagship project took over a year. It failed to be useful because the the business users had built an interim shadow tech solution which better met their needs. The technology leadership blamed the business users for not being engaged enough. The same executives who declared the project a “death march” promoted the technology lead who was the main cause of the failure.

Another project started with a decent plan. The introduction of a new pervasive technology would begin with a three month pilot period with teams scattered across the organisation. Even though the technology was well established in the market, the internal implementation at scale needed to be debugged. After the three month pilot, the plan was to roll out the new tool to the entire organisation. I am a huge fan of the work of Chris McDermott and Marc Burgauer with Social Practice Theory. Introducing a new tool is normally a way for getting the organisation to adopt a new practice. Building organisational muscle memory in a new practice takes time and should ideally occur before the tooling. Those engaged in the pilot were the best of the best in the new practice and easily adopted the tool. Hundreds of teams adopted the new tooling. At the end of the three month pilot not one single team had successfully migrated. Just before go live, we summarised the status as a “shit show!”. The executive explained that it was a game of chicken. “Who will blink first, the central team or the teams implementing the new tooling. No one wants to call it a failure.” The mass migration of the entire organisation went ahead over Christmas. Some highlights from the project:

  • A lot of teams discovered that the migration had failed for them and were unable to work.
  • The central team had not considered that thousands of developers in India do not take vacation over the Christmas period.
  • The entire central coordination team went on vacation the day before the week long migration started. No one knew who to contact for help.
  • The central team did not even think about creating a help desk or support team until weeks after the migration by which time teams were in chaos.
  • The definition of done for the migration was cut back, and cut back, and cut back as the end date for the migration moved out quarter by quarter. As a result, cost savings were not realised as licenses needed to be extended.
  • There was a lot of disruption for teams, with many millions of dollars worth of lost productivity across the organisation.

I think you know what happened next. The migration was declared a huge success by the Group CIO, and a few weeks later the Group CIO was rewarded with a promotion to the Board of the Group.

In a “Risk Managed Culture”, stage gates based upon success criteria rather than arbitrary dates would have determined progression, and retrospectives would be held on a regular basis to learn and improve. There would be proper transparency rather than a game of “Chicken”. The retrospectives would seek to ensure that similar failure would not occur in the future for the same reasons.

In a “failure culture”, the failureship do not want genuine transparency. They do not want to know that there are problems. They just want to make sure that they can contain and hide it. They promote the people who cause failure so that they do not need to acknowledge it.

Failureship doesn’t avoid failure… Failureship REWARDS failure.


Introducing Failureship – The dark twin of Leadership

The culture of an organisation is vital to its ability to change and acquire new capabilities so that it can adapt to new contexts. Whilst academics and thought leaders will tell you that culture cannot be changed and is different for each organisation, careful observation will reveal that there a couple of key cultural archetypes when it comes to change in organisations. Like blind men scrambling over an elephant, there are many ways to describe these polarities. Currently I name them a “risk managed culture” and a “failure culture”. Associated with each culture is a set of behaviours. In a “risk managed” culture, we have a set of behaviours that lead to change, a set of behaviours we call leadership. In a “failure culture”, we have a set of behaviours that prevent change and perpetuate the status quo, a set of behaviours we shall call failureship.

Image by KRISTEN FOSTER from Pixabay

“All that change needs to fail is that good leaders do nothing”

Failureship is the set of behaviours that prevent change and perpetuate the status quo. It is a mixture of conscious AND unconscious behaviours that undermine and block change. Failureship behaviours are not the fault of an individual “managers” but a complex interaction with the culture itself. They are a behaviour-plex where one behaviour leads to another which leads to another which feeds back and stabilises the original behaviour. Sometimes they are held in place by an enabling constraint rather than a governing constraint.

An understanding of behaviours in a “risk managed” and “failure culture” and the associated “leadership” and “failureship” behaviour sets can help us better predict the success and failure of a change within an organisation. As most people consider leadership synonymous with leaders, I will refer to “leadership” as “changeship”, the act of leading change rather than leadership which most consider to be the rulers of an organisation.

It is worth noting that failureship often leads to huge success for the individuals but ultimately leads to long term damage of the organisation they work for. The “failure culture” rewards individuals engaged in failureship and although it does not punish those engaged in leadership, they are likely to be undervalued and will seek organisations where they are valued. The result of failureship is the squeezing out of leadership and the eventual dominance of a failure culture and the establishment of a ruler group who are experts in failureship… the failureship. As the goal of the failureship is value extraction, the failureship will ensure that the organisation will appear to be in rude health even though it is unable to change to meet changing contexts. The Failureship will trade long term health for short term gains.

An understanding of failureship will help the following groups of individuals:

  1. Leaders of organisations that are genuinely interested in change. An understanding of failureship enable them to identify colleagues that are consciously or sub-consciously preventing change. It will also help them to understand the behaviour-plexs that need to be disrupted in order for change to occur. Most importantly it will help them to understand how their own behaviour contributes to change and prevents it.
  2. Change agents can use it to identify whether the leadership in their organisation is commited to change or is consciously or unconsciously opposed to it. They can identify whether their efforts to help an organisation improve will result in success, or frustration and failure.

It is worth noting that any organisation consists of many sub organisations. It is possible for different sub organisations to have a different culture, especially if the leadership of the sub-organisation is strong enough to resist the culture of the wider organisation.


Wrong-order-o-meter (An experience report)

About a year after implementing Capacity Planning ( aka Delivery Mapping ) at Skype, the product leadership team asked for reporting to support the process. Tony Grout, Lisa Long and myself implemented the “Strawberry-jam-o-meter” and the “Wrong-order-o-meter” with the product leadership team ( report names inspired by Dr Seuss books which were a favourite of one of the product executives ). The product organisation consisted of about two hundred product owners working with about three hundred scrum teams. Capacity Planning meant that for a year the product organisation had come together once a quarter to prioritise an organisation wide product backlog based on the available capacity of teams.

Skype consisted of several front end products (Windows, Windows Tablet, XBox, iOS, Android and Windows Phone) that implemented back-end products (Audio/Video Codecs, Transport, Login). The architecture was constructed using Reverse Conway principles with the aim of allowing each scrum team in a release vehicle (a set of scrum teams) to be able to deliver independently.

The organisation wide backlog (known as the meta backlog) was a set of business goals whose success was measured using metrics. For example, reduce the day one churn (number of users that only use the product once and never return), conform to regulatory rules to prevent significant fine or enable asynchronous chat by moving it chat from peer to peer to server based.

Anyone who has worked in a large organisation knows that it is pretty much impossible to set up your organisation such that every business goal can be delivered by a single scrum team. Several teams are often required to achieve a business goal. Large organisations need to create a single ordered backlog to provide focus and create alignment between the teams.

Once the leadership team had established the process to create an ordered backlog, they were able to get better insight into what was actually happening in the delivery. To that end, they reduced work in progress by over 50%. They also created two reports:

The Strawberry-Jam-O-Meter was used to identify teams that had the most backlog items in progress. Inspired by Jerry Weinberg’s “Law of Strawberry Jam”.

The Wrong-Order-O-Meter was used to identify teams that were working on items in the wrong order according to the organisation wide backlog that the product owner organisation had specified.

The reports were to help the leadership understand which teams needed their help and support. The reports provided bad smells that helped them identify issues to resolve. To quote Jabe Bloom, the reports Were an “invitation for the leadership to go to the gemba”. There was an understanding in the product leadership team that the “behaviours” were often a result of constraints in the system. For example, the team is blocked waiting for another team and moved onto the next item. In some cases the product owner was not aware of the impact of too much work in progress which was a “learning opportunity”. In one case, the product owner on the 16th priority was very aggressive at getting their own delivery done such that they put the 1st item at risk.

The two reports provided leadership with a system wide view so that they could identify constraints. The leadership were more interesting in solving system wide issues rather than issues with an individual team.

Layout and Calculations

The Wrong-Order-O-Meter was very simple. It showed the team name and the effective position in the backlog that they were working. It was possible to drill down on each team to show the items they were working on.

TeamEffective Backlog Position
Griffindor3.5
Slytherin3
Hufflepuff1

Drilling down into Slytherin’s backlog:

Organisation Backlog ItemPriorityEffortPriority Weighted Effort
Improve App Store ratings in Asia18%0.08
Reduce churn in Asia235%0.7
Improve on-boarding conversion rate314%0.42
Increase new customers into Funnel435%1.4
Items not on the Organisation Backlog58%0.4
Totals100%3.0
Effective Backlog Position3.0

Fifty percent of the effort in the organisation will be for “Items not on the Organisation Backlog” because the teams will not be able to work on the backlog item as there will not be enough capacity in the teams acting as the constraint in the system. For this reason, the leadership needed the strawberry-jam-o-meter as well to identify teams that are spread too thin both for items on the backlog and for items not on the backlog.


Stress testing Skills Liquidity

Corona virus has raised the importance of Skills Liquidity or “key person dependency”. As well as managing the normal skill liquidity risk, organisations now need to consider the “summer holiday” effect. During the summer holidays the productivity of many teams plummets for an extended period because of skills liquidity. The absence of one or more team members means that the team experiences a significant reduction in throughput and a massive increase in lead time. To address this, organisations need to move quickly address risks that are normally annoying but are now significant.

To address skills liquidity it may be necessary to redeploy people from non-core applications to core applications. So the first step is for organisations to classify applications as core and non-core (possibly with more refined classifications).

Then create a skills matrix for each application ( I like the way Cat Swetel describes it here ). We end up with a skills matrix like this one:

Screenshot 2020-03-29 at 10.30.18

Then we create a scenario where team members are unavailable for any number of reasons such as:

  • Being ill for a few weeks
  • Looking after sick friends or relatives
  • Unable to login due to lost internet capability
  • Called up by the government to assist in army or medical capability

We assess the scenario impact on the ability of the team to fix problems or make critical emergency changes:

Screenshot 2020-03-29 at 10.30.33

Where we do not have enough skills liquidity (options), we need to create them. Bring new people onto the team, possibly former team members if they are not supporting other core applications. Creating new features is probably higher risk and can reduce the stability of the application. The new team members can refactor the application, create automated tests and pay down technical debt. They can automate the build chain and generally make the application more resilient.

Core versus Non-Core

A follow up on core / non-core and the risk profile of each skill / component to follow.

 


There are no stakeholders in (Scaled) Agile

A month ago a colleague of mine asked me about stakeholders in Agile. After a moment of thought I told them that there are NO STAKEHOLDERS in Agile.

street-3169981_1920

What is a stakeholder?

In traditional development, someone puts together a business case, some business value that will be delivered, often with a high level description of the solution. A sponsor then secures funding for the business case and it is handed over to a project manager to deliver it. Funding is corporate gold, quite literally, and it attracts attention from anyone who has an idea that might have a claim to it. Some of those claims are more legitimate than others. Those people who have a claim on the budget are referred to as “stakeholders”.

The project manager takes direction from the stakeholder (often indirectly). The waterfall practice of “Stakeholder management” or “management of stakeholder expectations” involves minimising the impact of the stakeholder claims on the Sponsor’s deliverables. Where the stakeholder has a legitimate claim, it should be minimised. Where the claim is not legitimate, it should be ignored… politely in a way that does not offend the stakeholder.

In finance, examples of legimate stakeholders might be:

  • Someone from finance who needs a feed to general ledger (so that the CEO does not go to prison for a Sarbanes Oxley violation).
  • Someone from the regulatory department who needs a feed to the regulatory reporting system (so that the CEO does not go to prison for a MIFID violation)

Examples of stakeholders that are not legitimate might be:

  • Someone from marketing who requests a feed to the customer targeting system.

The project manager will ensure that the finance and regulatory requirements are met with the least possible effort.

Stakeholders in Agile

There are no stakeholders in Agile.

Stakeholders are a phenomena that only occurs in larger organisations. Stakeholders do not occur in small organisations.

In Agile, Stakeholders are either customer segments with a need, or more normally, they are product owners who represent customer segments with a need.

In Scaled Agile, there are no stakeholders because the customer and their needs are represented by the product owner. In the simple case, the product owner order the customer segment’s needs relative to the needs of the other customer segments and the associated value to the business. In complex situations, a group of product owners comes together to create a shared organisational backlog. To facilitate that process, the organisation can use “Capacity Planning”. In really big organisations, the “Capacity Planning” session would involve Product Executives who own parts of the business organisation.

Stakeholder management and big up-front business cases are an hangover from waterfall development. The organisation decides up-front what “Value” is going to delivered, and scope is managed (reduced and increase avoided if possible) to ensure the value is delivered. In Scaled Agile for Practitioners, executives decide which capabilities (teams etc.) they want to invest in, and then the product organisation optimises the business value delivered given the constraints inherent in the organisation. “Capacity Planning” provides feedback to the executives to indicate where investment in capacity needs to be increased and reduced.

If we consider the stakeholders above, it is possible that the product owners may consider that the need identified by marketing delivers the most business value to the organisation if it is satisfied.

In Agile, there are no stakeholders, only customer segments with needs, and the product owners who identify and prioritise those needs.

 

 

 


The London Agile Software Camerata

Gitte recently recommended this fantastic must-read article on Cameratas by Jessica Kerr. My only criticism is this statement:

One camerata emerged inside ThoughtWorks (consultancy) in London around 2003–2006. This group gave us Continuous Integration, Continuous Delivery, DevOps. They weren’t all on the same project, but they talked to each other, and they solved the problem of: Does deployment have to be so hard?

Jez and Dan and Chris Read produced The Deployment Production Line. Later, Dan went to invent BDD, Sam Newman became a prophet of microservices, and more. I keep meeting conference speakers who were part of this group.

The Camerata in London was not centred on ThoughtWorks, it was centered on The Extreme Tuesday Club. In fact, ThoughtWorks was one of several companies that played host to a chunk of the Extreme Tuesday Club, others being Connextra, BNP Paribas, Google, UBS, Sky, Springer and Zuhlke.

The_Old_Bank_of_England_(geograph_5585326)

The Old Bank of England was the second or third regular meeting place for XTC. It was the first place where I attended.

Why Extreme Tuesday Club?

The Extreme Tuesday Club was one of many Local Agile Communities that existed in the early naughties. Other notable communities were The Salt Lake City Round Table (Todd Little, Alistair Cockburn, Jeff Patton), The Silicon Valley Patterns Group (Rus Rufer, Tracey Bailik, Joshua Kerievsky, Mike Hill), The Portland Patterns group (Ward Cunningham, Diana Larsen) and many others. Members of these local communities were networked thanks to the OOPSLA conference.

In 2003, the Salt Lake City Round Table organised the first Agile Development Conference which went on to merge with XP/Agile Universe and become the Agile20xx conference. In 2004, the second Agile Development Conference was organised by members of the Extreme Tuesday Club, namely Rachel Davies, Andy Pols, John Daniels and Steve Freeman, of whom only Steve ever worked for ThoughtWorks.

I was the only employee of ThoughWorks London to attend in 2003. Partly as a result of the write up of my experience, in 2004, about seventeen employees of ThoughtWorks London attended, and we had a ThoughtWorks stand. This launched ThoughtWorks London on the Global Agile Stage as the group made a major impact on the conference.

Like the other communities, the combined impact on Agile of members of the Extreme Tuesday Club has been quite significant. I will list a few below. Many of those members worked for ThoughtWorks for a period of time. However, it is the Extreme Tuesday Club that provided the community that generated the ideas.

The Extreme Tuesday Club had weekly feedback loops. People would suggest an idea one Tuesday, a week later they would get feedback. XTC was a hub. Visitors from outside London would head to XTC and the global network was strengthened. Famous visitors would also attract new members. XTC organised XP Day which was the only place in London to learn about XP/Agile for many years. Mainly, XTC was a friendly face and a empathetic ear for those trying to make XP/Agile work in their work lives.

Why ThoughtWorks?

ThoughtWorks was hiring people with experience of eXtreme Programming at a time that no one else was. Paul Hammant actively recruited people from the Extreme Tuesday Club into ThoughtWorks. Hence there was a very large overlap between the Extreme Tuesday Club and ThoughtWorks. I am eternally grateful to Paul for being one of the people he recruited from XTC.

ThoughtWorks was great. It was on the bench that Dan North explained BDD to me ( c/Assert/Should ) which made it much easier to coach people to adopt TDD. “Should” provided me a link between Business Analysis and Agile Development. I then introduced “GIVEN” to express mocks. Dan and I then collaborated to come up with the “WHEN” and replace “SHOULD” with “THEN” to complete the trio. This was only possible because of ThoughtWorks positive attitude to supporting innovation in the Agile Space. ThoughtWorks left us to work on things we considered valuable rather than find busy work for us.

I left ThoughtWorks shortly afterwards in 2004, continuing to develop Real Options, Feature Injection (which includes GWT), Kanban, Capacity Planning, Skill Liquidity, Tech Debt in collaboration with members of XTC and the wider network of communities. When people left TW, many remained together as a community as part of XTC. Sometimes, XTC members would help people find jobs when they left TW unexpectedly.

Why ThoughtWorks rather than Extreme Tuesday Club

XTC is only something that you will hear of if you are connected to an existing member or visitor.

Why do so many people think that the London Software Camerata centres around ThoughtWorks rather than Extreme Tuesday Club? The answer is simple… Martin Fowler, Dan North and Jez Humble are more famous than the Extreme Tuesday Club.

Within ThoughtWorks, like all consultancies, they promote their own and play down the role of others. Within ThoughtWorks, you are more likely to hear about the contributions of ThoughtWorkers that the shoulders of giants that they stand upon. Martin Fowler has 250K+ followers on twitter. He is brilliant. His job is to promote ThoughtWorks and he does it well.

So people see Dan and Jez and other famous ThoughtWorkers rather than the XTC.

The London Software Camerata

XTC is an important part of Agile History. It would be great if we tried to document that history, especially showing the collaborations between members.

Years ago I tried to list the contributions of XTC members to Agile. I gave up after about listing fifty people. I just want to give you a taste (Please add a comment with people I have left out):

  • Rachel Davies
  • Tim MacKinnon*
  • Ivan Moore*
  • Andy Pols
  • John Nolan
  • Nat Pryce*
  • Ade Oshineye*
  • Joe Walnes*
  • Liz Keogh*
  • Gabrielle Benefield
  • Robert Benefield
  • Benjamin Mitchell
  • Portia Tung
  • Keith Braithwaite
  • Mike Hill*
  • Richard Watts*
  • Paul Simmons
  • Tom Ayerst
  • Antony Marcano
  • Andy Palmer*
  • Giovanni Gasproni*
  • Douglas Squirrel
  • Marc Johnson*
  • Gojko Adzic
  • David Evans

* – Indicates they worked for ThoughtWorks at some point.