Category Archives: Organisational Thinking

DevOps in a Nutshell. Treating Machines like machines and people like people.

Serviced by @buildcycleworks

A google search finds few posts called DevOps in a Nutshell. As far as I am aware this may be the only one that mentions nutshells and how metaphors can help us understand DevOps.

I’m going to use nutshells to start a conversation about how the language we use can affect how we think about how best to organise. Our hidden metaphorical understanding of how things work also informs how we think, limiting the ideas we have and bypassing our conscious thought and critical thinking.

 

The first way of DevOps is systems thinking, and the performance of the entire system. The entire system contains people and machines. The dominant organisation metaphor in use is machine based. I’ll show others, and how they help the DevOps First Way, and understanding the entire system.

Treating Machines like People and People like Machines

Before DevOps we treated machines like people – there was really no other way. They needed to be installed manually (pre Jumpstart and FAI, old people), it took about 9 months and when you’d finished you had an almost living thing that you gave a name. Installs were often path dependent. It mattered what order you installed things, and it was quite hard to get and keep multiple machines exactly in sync. The tools to work differently simply didn’t exist the way they do now.

There is also an old people management technique, called Scientific Management or Taylorism, after Frederick Taylor. Taylor invented and popularised scientific management at The Ford Motor Company. Taylorism took craftsmen & women (especially during WW2) and timed, measured, divided and controlled what they did until they were doing repetitive tasks on a production line, separated from the craft of what they were making.

Taylorism was popular with factory owners, much less popular with workers. Without it, we simply would not have the world of cars and technology we have today. Taylor also raised wages, so that Ford employees could afford Ford cars. It’s not all bad.

Scientific management treats people like machines.

Pre DevOps we had machines treated like people, by people treated like machines.

Now we have an abundance of approaches to treat people and computers differently.

We have automation, orchestration and monitoring tools for treating computers like machines, and the tools keep getting better.

Our approach to people is less mature. We still can often use the machine based metaphors and ideas. No one would have thought it unusual if I’d called the approaches to working with people ‘tools’. The machine metaphor goes deep.

If not machines then what?

It’s worth saying that it’s possible you may feel that people should be treated like machines. This is how Frederick Taylor saw the world and is dominant thinking in many areas.

Often I’ve found that those who feel people should be treated like machines usually mean other people. There is a difference between the thinkers and the doers.

If we think that other people should not be treated like machines, what are the alternatives? We could treat people like roads or books or gases. But I’ve not really thought these through…

Gareth Morgan wrote the seminal book Images of Organisation that contains a set of metaphors that cover most ways of treating people in our Organisations. Fortunately, he’s thought about the metaphors so I don’t have to.

Gareth Morgan’s Organisational Metaphors

Picture by Venkatesh Rao, ribbonfarm.com

Gareth Morgan’s Images of Organisation has eight metaphors; Machine, Organism, Brain, Culture, Political System, Psychic Prison, Change and Flux and Instrument of Domination.

All the metaphors can be true at the same time. Often people favour a dominant metaphor in their understanding of their organisation. This dominant metaphor may be colouring their thoughts and reaction and shaping the ideas they have. When I say their thoughts I mean our thoughts. Mine and yours.

The organisation that your DevOps team is part of will be working in ways they cannot see if they only think with one metaphor. Using a variety of lenses increases our understanding.

The metaphors shown in the diagram above are

  • Organisations as Machines
    • with a known input they should produce identical output
    • ‘well oiled’
    • People as machines are interchangeable cogs, once trained or certified
    • It’s about Input, process, output
    • We reduce variation (6 sigma)
    • A leads to B leads to C, always
  • Organisations as Brains
    • Who knows what
    • Information spreads through learning
    • Constant learning through feedback, and learning to learn
    • Management cybernetics
  • Organisations as Organisms
    • Adapting to the variety of a changing environment
    • Evolving from organisational DNA
    • Changing with a survival strategy
    • Looking for best fit with the environment
  • Organisations as Cultures
    • The way we do things here
    • What we value Systems
    • Behavioural Norms and Patterns of Behaviour
    • Dominant cultures and sub-cultures
  • Organisations as Political Systems
    • Influence
    • What’s in it for me?
    • Centres of Power
    • It’s not what you know, it’s who you know
  • Organisations as Psychic Prisons
    • Cognitive Biases
    • How else could I behave
    • We couldn’t do that here
  • Organisations as Flux and Change
    • Constant change
    • Less about boundaries
    • We’re connected and part of our environment
    • A system can’t change independently of its environment.
  • Organisations as Instruments of Domination
    • Prisons or Boarding School
    • Physical Punishment
    • Prison Industrial Complex

When using the DevOps first way to think about the whole system, the metaphors can help uncover and explain things that are happening. Viewing situation through the metaphorical lenses is revealing.

Conclusion

“In a Nutshell” not a great way to think about this topic. It suggests that “here is everything you need”, a complete explanation. When we’re working with people we need to be open to ideas and change.

I’ll use another post to discuss situations and responses through the lenses above.

IT From Common Resource to Strategic Partner

800px-cows_in_green_field_-_nullamunjie_olive_grove03

This blog post is about IT in large organisations, including public organisations like councils and businesses where new Tech competitors are changing the environment of business. The environment is changing, or more accurately, being changed by the strategies of competitors.

Organisations are looking for a digital strategy to combat this threat, moving digital to the heart of what they do. IT for it’s part is keen to become a Strategic Partner to the business. So what is stopping it?

Where did IT departments come from?

It’s worth looking at how IT departments may have been created. They traditionally exist as a cost on the balance sheet, providing common resource to other areas of the business, often underpinning other parts that are necessary, but may not directly exchange value with the environment, like accounting, marketing or HR. The may also run internal and external websites, but they are unlikely to be the core value propositions.

If not managed,  common resources can suffer from the “tragedy of the commons”. The popular example is common grazing land is so overburdened by people wanting to graze their animals that loses it’s initial value. People are assumed to want to maximise the number of animals on the common land.

Common IT Resources

Many people in IT departments will recognise this, with many unrelated customers in the business wanting their work to be prioritised by the limited IT resource. Like in the tragedy they want to get value from the resource. There is now no such thing as an IT project – they are business projects central to strategy. But with a common IT resource each may be another cow on the  metaphorical field.

The big issue is ‘Who would strategically partner with an unmanaged commons?”. It’s a very risky proposition.

From an Unmanaged Commons to Strategic Partner

Managed commons can and do exist. Elinor Ostrom studied working, managed commons, and found that there are 8 organising principals common to functional commons. Applying these to IT could provide the step towards being seen as a potential Strategic Partner.

Eleonor’s rules applied to IT may look like

  • Define clear group boundaries
    • This is perhaps the easy part, but it’s vital to understand where the boundaries are, so you can understand and manage the work and relationships across them
  • Match rules governing use of common goods to local needs and conditions.
    • Your rules will differ from other commons, best practices won’t work. You need to look at what is required by the people who use the resources.
  • Ensure that those affected by the rules can participate in modifying the rules.
    • For natural commons like an inland fishery the fish don’t get a say. For IT departments there is likely to be internal work that needs be done, for example upgrades, patching and audit requirements.  So the IT department itself, alongside Project Managers, Service Managers, Marketing and Finance should discuss the rules for using the common resoure.
    • Users of IT resources need a way of getting work done
    • They need a way of getting progress reports out
    • They need a way of getting ad-hoc questions answered by subject matter experts.
  • Make sure the rule-making rights of community members are respected by outside authorities.
    • We need to sell the idea that IT will work better for everyone as a managed common resource
    • We should have rule for getting urgent business requirements discussed and done appropriately – so that there is not a requirement for the use of higher authority to get work done
  • Develop a system, carried out by community members, for monitoring members’ behaviour.
    • Monitoring should be done by users of the commons. It is in their interests that the rules they helped create are followed.
  • Use graduated sanctions for rule violators
    • Starting small, and agree. What sort of sanctions would you like to see?
  • Provide accessible, low-cost means for dispute resolution.
    • Anticipate things may go wrong, and we know how issues will be resolved quickly and easily
  • Build responsibility for governing the common resource in nested tiers from the lowest level up to the entire interconnected system.
    • Create systems within systems, with each level being viable. The Viable Systems Model would be ideal for creating, or diagnosing this organisational structure.

There are many other areas of organisations, or entire organisations that are Common Pool Resources. Strategically some may aim to be well managed commons, others may need to use this framework to be seen as a potential strategic partner by other areas of the organisation or environment.

Why are people replacing robots?

Mercedes-Benz is replacing some of the robots in their factories with people.

http://www.theguardian.com/technology/2016/feb/26/mercedes-benz-robots-people-assembly-lines

3127953038There had to be some passionate conversations between factory managers, and executives at Mercedes-Benz with this one. Replacing some robots with people has caused all sort of problems. If the factories are quite new, and were built for automation they probably don’t have many toilets near the factory floor. Or a large car park, or canteen. Robots don’t drive to work, and need to eat. Factory managers will take personal and professional pride in running efficient operations and automation has made cars affordable, reliable and available.

What is going on when the factories start employing people to replace robots? Wearing an efficiency hat this doesn’t make sense. Robotic factories have been the only future imaginable for years.  What has changed? Does the factory rulebook need to be rewritten?

Increasing pace of change and complexity 

From the Guardian article:

The robots cannot handle the pace of change and the complexity of the key customisation options available for the company’s S-Class saloon at the 101-year-old Sindelfingen plant, which produces 400, 000 vehicles a year from 1,500 tons of steel a day.

We need to be flexible. The variety is too much to take on for the machines. They can’t work with all the different options and keep pace with changes.”

Robots can’t currently mange the complexity of the customisation options. People are currently able to outperform robots at tasks requiring variety, at least until the robot manufacturers catch up.

Managing Variety 1: Making something you can sell

To find where this variety has come from, we can start in sales. Mercedes is in business to sell cars to customers. To do the sales folk need to offer what the customer environment wants, a…

… dizzying number of options for the cars – from heated or cooled cup holders, various wheels, carbon-fibre trims and decals, and even four types of caps for tire valves –

There may be customer demand for these, or marketing could have created the demand. Either way, with a lean, efficient production line, the sales folk are selling something the robotic factories can’t make.

Variety 2: Selling something you can make

Mercedes need to reduce the variety their customers demand to a level their factories can cope with. Balancing this equation is essential. Of course the ultimate offer would be a custom Mercedes for each customer, but this is not possible for the cost of a Mercedes.

To do this marketing and production have to work together to design and market cars that they can make in their factories. Mercedes-Benz are a luxury brand, so cost efficiency is not the sole purpose of the factory.

Marketing has to create and manage demand for the sort of customisation that their factories, restructured with people and robots, can produce. People can cope with the operational variety that robots, or people behaving like robots can’t.

Using robots, machines or computers increases efficiency, but reduces the ability of the system, in this case a factory, to cope with variety in an fast changing environment.

At every level we must ensure that the variety equations balance. If a car dealer can’t supply what the customer is asking for, they will buy elsewhere. If the factory can’t make what the car dealer is selling then the business won’t last long.

 

 

 

Why we need Models, and why it’s hard to change them.

  •  It’s 460BC. Your job is a map maker, and your maps show the world to be flat. You’ve a lockup garage of flat earth maps to sell. But you also like astronomy, and understanding the planets.
    • Is a model of a flat earth of any use? Is it good?  It was good enough for me to get to work, and to drive a cart to London.
    • But it’s not good enough for astronomy, you need another model.
  • You hear of the model of the earth as a sphere. Hmm, this fits simple astronomy, but does it make your lockup full of flat earth maps worthless? Which model do you believe? How hard is it to change your mind to a new more complicated model?
    • Is the model good enough? It’s great when thinking on a global scale – like where is Australia relative to where you are.
    • But maybe it’s a bit complicated for driving to London. A flat earth map will be fine for that.
  • From the international space station, is the model of the earth as a sphere good enough?
    • Maybe not. Gravity may be affected by the shape of the earth, and the movement of planets may need more complicated models. But perhaps you don’t need a model of the earth that shows the Himalayas.
  • Is that enough models?
  • What if you are cycling to London? A flat earth map won’t show you the hills, but a spherical model with enough detail is far too much information. You like to avoid hills, so you need another model.

Using the examples above, I think we can learn:

  • We need models. A model is a synonym for an understanding
  • Multiple models of the same thing exist at the same time
  • New models should compliment existing ones
  • We should use the simplest model we can, but no simpler
  • We need awareness of other models
  • Believing in one true model is an Anti-Pattern
  • If you have an interest in a model being true (like a business selling flat earth maps) it could be hard to learn a new model. The greatest resistance against a new, different model may be those who currently benefit from an existing model.
  • All models are wrong, but some are useful. Is the only up-to date model of the earth the earth itself?

This cartoon shows Calvin explaining his simple model to his toy tiger.

Calvin-Toast

This model of how to make toast is sufficient unless:

  • Calvin starts to sell toast in his yard and
    • He may be asked to contribute towards the electricity bill
      • “There is electricity and you have to pay for it?!”
    • He may have to buy his own bread
      • “Can’t I reuse the bread I just put in somehow?!”
    • There is a drought and the price of bread rises
      • “So I’m losing money on everything I sell?!”

Systems Thinkers love models. It’s how we understand the world, and different perspectives and contexts.

We can also see that if you insist on using a simple model, for example one that will fit on a napkin, or can be explained to a 6 year old, then you can only use it in simple situations. More complicated systems need bigger models.