Tuesday 17 June 2014

CDE and Big Brother

Dear Junior

What can a sleazy soap opera teach us about coaching agile teams and their managers?

It was somewhat a surprise to me when I realise that the reality TV show Big Brother makes a good metaphor for explaining one of my favourite models for understanding self-organised teams: the CDE-model.

The Big Brother TV-show is a non-scripted soap opera where a bunch of people are locked up in a house together and followed by cameras day and night. The "thrill" of the show is how the people interact and react upon each other.

The CDE-model is a system-theoretic model to reason about how teams self-organise as a reaction on their surrounding. Of course these reactions are very non-linear and hard to predict.

What on earth have these two things in common?

Let us put ourselves in the shoes of the Big Brother producers. Say that the events in the locked house have become a little bit dull lately. The people locked up have settled for a pace of life where they all sustain without unnerving the others. Nice for them, but not thrilling TV. We need something to happen.

The problem for us is that Big Brother is non-scripted. We cannot direct Susan to "go and snug up Jonathan", even if we think it would be an interesting turn of events. We need to find other ways to change the behaviour inside the house without giving explicit directions.

So we decide to shake them up a bit by changing the conditions under they live. A rough partition can separate three types of conditions: containers, differences, and exchanges.


One type of conditions we can change are the containers.

We can lock the yard so that they are locked indoors. That will unsettle Bob who is used to take his morning strolls there. He might have to spend his mornings in the kitchen together with Alice, who has a terrible temper before getting her coffee. Or, we can open up an extra room, one that is a little bit hidden away, with very little insight - apart from the camera of course. Or, in the middle of the night we could push in an extra wall dividing the house into two parts; let us ensure that Tom and Lisa are locked up in one part, and Toms rival Jerry in the other part. Or, we could simply remove all doors. That could be fun.

All these are examples of changing the containers that contain the contestants - making the containers bigger, or smaller, or less connected, or more connected, or even dissected. 

Containers need not to be physical, there are other ways to divide a group. We can create a competition between two teams, then the teams become sociological containers. The teams can follow some obvious partitioning like city of birth or gender, or by some arbitrary dissection. 

The important part is that containers effect the patterns of interactions, so changing containers will change the behaviour.

Of course we could interpret "containers" literarily and put them all in a freight container. That is an idea.


Another type of conditions we can change are differences and how they are resolved.

We can stir up events by introducing differences. For example, if we have an group of contestants with the same ethnicity we can send in a new contestant of a different ethnicity, for example sending in a white guy when there are only hispanics in the house. Or send in a professor of philosophy in a house with high-school drop-outs. You get the idea.

Sending in someone fair-haired when all are brown-haired will probably not make a lot of fuzz. In Big Brother, hair colour does not make a difference to how people treat each other - at least not in a way that make a significant change. We say that we only consider "significant differences". 

It differs from situation to situation what is a "significant difference" but in general the nuance of hair-colour is not one, whereas gender is - men and women are (sadly enough) treated differently. Same goes for ethnicity, sexual orientation and lots of other traits. All those are significant differences.

Introducing or enhancing differences are sure thing to induce change, but sometimes reducing differences can also unsettle the state of things.

Let us say we want to send in one person when there are four men and one women left in the house. Sending in a man would enhance differences from 4-1 to 5-1. Sending in a women would reduce gender difference to 4-2, but would probably be more interesting.

Apart from the differences as such, we have the issue about how these differences are resolved. 

One contestant might enjoy a particular kind of music, and preferably at high volume. Another contestant might not be so fond of that particular kind of music. However, they might be able to stand each other on a day-to-day basis. The difference is manageable, handling it is not difficult.

Now we can amplify the difficulty to manage differences, for example by introducing a large amount of liquor. Should either of them get drunk, or both, it will be harder to resolve the difference in taste of music, and we will probably see some interesting conflicts. 

Another difference is food preferences, one way to amplify the effect of this difference is to insist that everybody in the house agree on what should be served for dinner. Can be fun to see how Mark "must have meat" tackles Vegan-Lisa, even more interesting when he gets hungry. 

As differences and resolving differences are a major driver for interaction, obviously changing those differences will change behaviour.


Third and last of the conditions are the exchanges with the outside.

If we let the contestants interact with the outside things will happen. We might let each of them have a (filmed) phone-call with a friend. Or we can have a small room where one at a time is allowed to speak to the audience. Or we might take away that room. We could put up a big TV showing news from the outside. We could fake the news we show. Surely things will happen.

Exchange need not to be communication. We can change the way food is delivered to the house; instead of small deliveries every day we make one big delivery once a week. That will cause some interesting effects at the end of the week when someone has eaten all the goodies on day one. If we are diabolic we can give them slightly too little to eat. Surely things will happen.

As exchanges are the connection between the very limited system in the house, and the very large system on the outside, it is not surprising that how the inside and outside are connected will effect the behaviour on the inside.


Of course there are a multitude of ways to unsettle the status quo in the house. However considering containers, differences, and exchanges (CDE) gives a good start to think about what leverages we can pull to cause the people in the house to behave differently. Or phrased in system lingo: to make the system reconfigure itself in another configuration.

System theory teaches us that lots of systems are dynamic and non-linear. This means that it is very hard to predict the exact outcome of a change.

As producers of Big Brother we know this. When we nudge the system in the house, we know that something will happen, but we do not know exactly what. We can have a guess, but we do not know. So, we need to be at our toes to watch out if things go another direction, and take compensating actions - or actions we hope are compensating.

Well, obviously neither you nor I are producers of Big Brother. And most probably we will never be. But same ideas can be applied to think about agile software teams that have self-organised and when coaching them or their managers. However, that is a subject which is a letter of its own.



PS To be honest, my description does not perfectly fit how CDE was described by Glenda Eoyang in her Ph D thesis. But I think I am truthful to the main idea.

PPS Glenda Eoyang’s thesis can be read in full at http://www.hsdinstitute.org/about-hsd/dr-glenda/glendaeoyang-dissertation.pdf. A briefer introduction can be found at http://wiki.hsdinstitute.org/cde.

PPPS I think first time I came across CDE was when Mike Cohn introduced it to me during is tour when he released Succeeding with Agile, a good book for a lot of reasons and which also includes an introduction to CDE.

Thursday 5 June 2014

The Three Elements of Agile

Dear Junior

Not that I necessarily claim to be an old hip-hopper, but an interesting part of the hip hop culture is the concept of "the four elements". 
  • Breaking - more commonly known as breakdance
  • MC - more or less what rap is about
  • DJ - you know disk-jockey
  •  Graffiti - public painting
An interesting thing is that within the hip hop community it is commonly agreed that all four elements are needed. If you should take away any of them, the hip hop culture would not be whole, it needs all four of them to be complete.

Also, all the elements are on the same level. That MCing should be "higher" than graffiti, or vice verse, is an idea that would be refuted as ridiculous. All the elements are of equal importance and each indispensable. 

Obviously, the individual hip-hopper might not practice all of the elements. He or she will most probably immense in one of them. A few might practice two or three. Some might practice all four but are certainly not expected to master all of them.

Still there are a lot of respect between practitioners of the different elements: a talented MC easily gives public cred to a vicious graffiti painter. They are both needed and they would both feel poorer without the other. 

The four elements must be in balance for the hip-hop culture to thrive.

Well, that sounds all nice and cosy. But what does that have to do with agile?

I think the Agile community have a similar situation. Cutting the cake with a blunt knife I see three elements within the "agile culture": tech, process and organisation. OK, the knife is blunt so the cuts might not be perfectly clear, but I think it suffice to clarify my point.

Agile Tech consists of the technologies we have developed and use for the direct building of software. In this category we find frameworks as Hystrix for resilience; we find products as nosql-databases; we find tools as Chaos Monkey. Aside from the code we also have the close-to-code practices like Test-Driven Development, Domain-Driven Design, Continuous Integration, Build Pipelines etc. All these things enable us directly to write awesome software.

With "process" in Agile Process I simply mean "the way we work". We work in sprints á la Scrum, or WIP-limited continuous flow á la Kanban; we demo at regular intervals; we use information radiators and stand-ups to synchronise work; we make forecasts to synchronise our work with other departments; we run retrospectives after sprints and projects to have a structured learning etc; we set targets and measure progress visavi effect goals, e g using Impact Mapping.

Finally Agile organisation is about how we structure ourselves. Within teams we self-organise, sure, but this is what we do on a larger level as well. We must ensure that information propagate across the organisation in a sound way; we want the architecture to stay reasonably consistent, preferably without a command-and-control chief architect. We must also synchronise and prioritise initiatives across this organisation so we also find portfolio management and finance/funding practices like Beyond Budgeting. Last but not least we find Agile HR/Agile Talent Management as part of this element.

Each agile practitioner might not do all the elements. On the contrary, most will stay mostly within one element or have their foot-hold in one and cross over to the other. 

Comparing to the hip-hop culture I must claim that the agile community is not complete without all three elements. Doing only tech does not help us; doing only process does not help us; doing only org does not help us. Skipping any of them would hurt us.

Nevertheless I have during the history of agile I have seen competition between the element, sometimes even hostility between practitioners where one side have claimed to be the "true agile".

I think it is unnecessary and harmful.

Furthermore, I think it is unworthy of us, of a community praising "humans and interactions" over all things, a community that thinks "cross functional teams" are the superior model - should we not strive for a "cross functional community"?

In the same way as the hip-hoppers give cred across element borders I would love to see the kick-ass developer praise the agile-minded HR Director for "getting it"; I want the beyond-budgeting-inspired CFO to thank the engaged and engaging scrum master for the fantastic project retrospective; I want to see the impact-mapping business analyst to thank the devops for coming up with insightful real-time metrics about what customers actually do.

OK, I might be guilty of hippiesm here, but I want the three elements of agile to share love.



PS One of the reasons I care so much is that deep in my heart I feel that Agile is different.
PPS There is a video in Swedish from the conference Agila Sverige 2014 where I do a blitz presentation on this subject and some related stuff.