Hierarchical and network organisation structures
This page is published under the terms of the licence summarized in the footnote.
All free-to-read materials at http://avancier.website are paid for out of income from Avancier’s training courses and methods licences.
If you find the web site helpful, please spread the word and link to avancier.website in whichever social media you use.
This paper explores the organisation of actors into hierarchical and network structures.
The concept of organisation is not limited to bureaucratic structures, and begs definition.
Organisation structures are commonly divided into hierarchies and networks.
A hierarchy arranges elements (from the top-down) in a cascade of 1 to N relations.
A network arranges elements in N to N relations.
In a system description, a relation between elements A and B may be one-way or two-way.
It may imply only that A depends on B in some undefined way.
But it often implies the conveyance of command, control, direction or decision messages.
Hierarchies and networks in biology
When the biologist Bertalanffy used the word “organisation”, what did he mean?
He meant how persistent components (e.g. brain, heart, lungs and legs) cooperate in repeatable transient processes (e.g. breathing, running).
Some animals are hierarchical in the limited sense that they have a central nervous system, which directs much bodily activity.
But bodily organs cooperate in a hugely complex network that does not all require direction from the brain.
And other animals (molluscs) have a distributed nervous system; obviously, an oyster is a perfectly viable system.
Hierarchies and networks in social systems
When the systems thinker Russell Ackoff said “All organizations are social systems”, what did he mean?
He was thinking of systems in which the elements are human actors - a narrow view of system element.
And the organisation is a bureaucracy – which administratively organises the many actors who need to work together.
Public and private sector businesses do rely on bureaucracies to function.
Organisation structures are commonly divided into hierarchies and networks.
A core theme of sociological discussion has been the contrast between totalitarianism (hierarchical control) and individualism.
Where individualism might be anarchy, or liberalism or participatory democracy.
The actors in a bureaucracy are connected (directly or indirectly) by communication relations.
A message may be sent to convey:
· information needed to perform a regular activity, or a reply to a request
· a decision about which of several defined activities to perform
· a direction to do something not previously defined.
Some social groups have a dominance hierarchy; consider the pecking order in a chicken coop.
But the dominance hierarchy rarely prevents communication between any two members of the group.
You might describe a bureaucracy as a hierarchy, while in reality it operates as a network, the manager’s role is primarily to report results
Or else describe a bureaucracy as a network, while in reality it operates as a hierarchy, the decision-making power structure is hidden from you.
This paper goes on to discuss an unusual business (Morning Star) which has a network organisation.
Some social systems are (or appear to be) hierarchically organised, but this is not a generally-required property.
In naïve management methodology, executive directors set goals, which are then cascaded down the organisation structure to all employees.
By contrast, agile system development methods favour peer-to-peer collaboration over top-down direction.
And some propose that managers are the servants of knowledge workers.
(Read Social groups and social systems for a distinction between social groups and social systems.)
Imposing a hierarchy on things
Humans often organise actors and activities into hierarchies to make them manageable and/or understandable.
They often organise actors and activities into a client-server hierarchy.
They often organise communication actors and activities into a hierarchical communication stack (see Information Theory for more).
Hierarchies can be defined from the top-down, the bottom-up, or both, iteratively.
Restructuring the top-level sometimes implies restructuring lower levels also.
If it doesn’t, that suggests the higher level actors are disconnected from what actors are doing at lower levels.
Hierarchies in enterprise architecture
hierarchical structures on aims (cf. a balanced score card), management
structures, business function/capability structures, geographical locations
These different hierarchies can sometimes be aligned
general-purpose methodology has to regard aims, managers, functions, locations etc. as distinct hierarchies.
You can map items in one hierarchy to items in another using matrices or diagrams.
In a typical business, every employee can trace a reporting line from their position to the chief at the top.
And some or most, business operations are performed according to direction from above.
The Roman army was broken down into different groups to have a clear chain of command during battle.
1 legion of about 6,000 men was divided into 10 Cohorts
1 cohort was divided into 6 centuries (only eighty men on average).
1 century was divided into ten conturbenium, a group of about 8 soldiers who marched and roomed together.
Classical organisation theory discusses different ways for goals, decisions and plans to be determined in a hierarchy.
For example, you can analyse a hierarchical management structure in terms of its position on five spectrums below.
Most if not all real organisations sit somewhere between the two ends of each spectrum.
Top-down command control (Napoleonic?)
Goals, decisions and plans are cascaded from the top downwards.
Perhaps supported by the use of a “Balanced Score Card”.
Managers are empowered to determine local goals, decisions and plans
Internal planning (Napoleonic?)
Goals, decisions and plans come from a separate/remote strategic planning body.
Managers are empowered to change plans in the light of events.
Managers determine their own goals, decisions and plans, and may do it alone.
Managers must consult their peer group and others
(Beware avoiding responsibility and “risky shift” phenomenon.)
Flat or shallow structure
Managers manage small number of actors
Managers manage a large number of actors
Managers make near-random decisions or choices between options
Managers are fully informed of options and their likely outcomes.
It is usually difficult to position a business organisation firmly in any of the five spectrums above.
Different business divisions sit in different places in each spectrum, and they move back and forth along the spectrum.
However, different businesses may be compared, and a rough analysis may suggest where changes might be made to one or other.
And this paper goes on to discuss an unusual business (Morning Star) which might be characterised to the right on each spectrum.
Some evidence suggests senior management decisions are closer to random than you might think
“We evaluated 287 allegedly high-performing companies in 13 major success studies.
We found that only about one in four of those firms was likely to be remarkable; the rest were indistinguishable from mediocre firms catching lucky breaks.
By our method, even in the study with the best hit rate, only slightly more than half the high performers had profiles that were credibly attributable to something special about the firms.
In short, what qualifies as remarkable performance is anything but self-evident.”
The full study used to be found here http://www.deloitte.com/persistence
A management hierarchy is the norm large business organisations; it may be necessary for reporting and other practical reasons.
Hierarchologists often mistakenly attribute successful business outcomes to the decisions of top-level managers.
And over-estimate the extent to which middle managers direct business operations.
In hierarchical organisations, managers don’t make all decisions, don’t always make informed decisions and don’t always direct what happens.
Some managers work mostly as officials or administrators, allowing decisions to be made by people who report to them
Some managers are content to make near random decisions (to keep things moving), rather than seek out, collect and analyse information.
Much ground-level work has an impetus of its own, dependent on skilled employees and/or directed by customers and other external forces.
In this Harvard business Review piece <https://hbr.org/2011/12/first-lets-fire-all-the-managers> Gary Hamel talks about Morning Star.
Morning Star is a business that works as a network of actors (business units and employees) that define their own roles.
Read the paper for details. In short, it is claimed that:
• No one has a boss.
• Employees negotiate responsibilities with their peers.
• Everyone can spend the company’s money.
• Each individual is responsible for acquiring the tools needed to do his or her work.
• There are no titles and no promotions.
• Compensation decisions are peer-based.
Before I say any more, note that as many as 50% of seasoned hires leave Morning Star within two years because they have a hard time adapting to the system.
Vision, goal and personal mission statements
Morning Star’s vision that all employee will be “self-managing professionals, initiating communications and the coordination of their activities with fellow colleagues, customers, suppliers, and fellow industry participants, absent directives from others.”
The company’s overall mission/goal is to “produce tomato products and services which consistently achieve the quality and service expectations of our customers.”
Every employee draws up a personal mission statement to support that mission/goal: “e.g. turn tomatoes into juice in a way that is highly efficient and environmentally responsible.”
The core functions in the value stream
Morning Star is a supply chain business that moves and transforms physical materials.
It has three major functions or capabilities: tomato harvesting, tomato trucking and tomato processing.
The core value stream is: harvest tomatoes > truck tomatoes to 3 processing plants > transform tomatoes into 100s of product varieties > truck products to customers.
The management hierarchy
The management hierarchy of this privately-owned business is shallow.
There is 1 president and 23 business units, which each employ less than 20 people each on average.
Each business unit has its own profit and loss account.
It is not clear how business units are opened or closed or whether each business unit has a manager or vice-president.
It is not clear whether business units are divided into teams, some might be.
The service-oriented activity network
Every year, the 23 business units negotiate customer-supplier agreements with one another
And each employee negotiates a Colleague Letter of Understanding (CLOU) with associates most affected by his or her work.
“As a colleague, I agree to provide this report to you, or load these containers into a truck, or operate a piece of equipment in a certain fashion.”
A CLOU is an agreement that covers as many as 30 activity areas and spells out all the relevant performance metrics.
The outcome is a network structure in which there are roughly 3,000 formal relationships between Morning Star’s full-time employees.
One actor can be both client and server to another actor.
So Morning Star is a collection of hierarchies that are dynamic.
The CEO knows his people will think about the business holistically only if everyone has access to the same system-wide data.
That’s why there are no information silos and why no one questions anyone else’s need to know.
Each business unit publishes detailed financial accounts twice a month, which are available to every actor.
Actors are encouraged to hold one another accountable for results, so an unexpected uptick in expenses is bound to get noticed.
With this sort of transparency, folly and sloth are quickly exposed.
EA as a metasystem
Actors in business operations monitor and direct the activities of actors (customers, suppliers, employees and machines) that the business cares about.
Actors in EA monitor and direct the fitness for purpose of roles played by actors in business operations.
EA plans generational changes; it replaces business system Av1 by system Av2, or perhaps by system B.
EA as action research
Some schools focus on how a social/business “organisation” can learn from experience,
There is “Organizational learning” after Argyris and Schön.
And “Organizational self-awareness’ after José Tribolet (http://algos.inesc-id.pt/~jpa/InscI/poisson/varwwwhtml/portal/ficheiros/publicacoes/3982.pdf)
And “Action research”, which is iterative refinement of business operations by people involved in the processes.
In our terms, people switching from working in the operational system to the meta system that designs the operational system - and back again.
EA in Morning Star
Traditionally, the EA team is composed of actors who sit aside from the employees in business operations.
Suppose the EA team is composed instead of the employees?
Suppose employees and business units, by annually re-negotiating SLAs with each other, make generational changes to business operations?
Morning Star’s EA is not the flat management hierarchy.
Rather, it is in the self-defining service-oriented network of activities.
In effect, the CEO is the EA team manager, and every actor is an EA.
EA does not tell us how to design an organisation’s management structure, recruitment or remuneration policies.
These things may be done from the top down, or bottom up – as in Morning Star.
EA does encourage us to encapsulate business actors behind documented service-level agreements.
In Morning Star, the SLAs are pervasive and self-documenting rather than imposed.
The outcome is a business that is a service-oriented activity network.
Morning Star’s EA is not in the design of harvesting, trucking and tomato product manufacturing processes.
Rather, it is in the information created by those processes and how that information is used to steer the business.
EA is about information-intensive activities in information-intensive businesses
It is about formalising, optimising, standardising and integrating information-intensive activities.
And making best use of information captured from business activities.
Morning Star encourages coordination and sharing of business information.
I’m in favour of it in principle
I assume, like agile software development, it suits the way small social groups work.
It probably implies small teams that are well motivated, and implies they are competent as “knowledge workers”.
Remember that many as 50% of seasoned hires leave Morning Star within two years because they have a hard time adapting to the system.
I’m not sure how well this approach scales up to (say) how millions of customers make claims for compensation against their insurance policies.
The tension between micro-scale agile development and macro-scale enterprise architecture is addressed in other papers.
The fact is, almost all EA teams today work in more traditional organisations.
Even in businesses with a formal management hierarchy, much of the work is done according to an informal activity network.
This can be formalised along the lines discussed above.
Be wary of:
· assuming the structure of the management hierarchy is the most important thing.
· assuming that process quality will lead to product or service quality
· underestimating the effect that individual personalities have in business success
· underestimating the effect that chance plays in business success.
See also the “Human Factors in Hierarchical Organisations” page at http://avancier.website.
Enterprise architecture Organisation design
Processes involve roles which are instantiated in assignments performed by human actors who report to organisation units
Enterprise architecture teams concern is formalised business processes that create and use information.
The very rarely design an organisation’s management structure, a task usually left to business managers.
EA frameworks recognise that a management structure usually evolves faster than its processes. So, what to do?
First, define an abstract version of the management structure – often called a business function hierarchy or capability map
Then, relate other views of the enterprise to that more logical and stable structure.
The capabilities/functions may be mapped to the organisation’s management structure.
But that management structure is a secondary concern to most EA teams.
EA teams don’t lead organisation design or cultural change, or recruit employees, or write their job specifications.
These things are usually the responsibility of others - business managers, line managers, human resources and business change teams.
EA treats an enterprise as a collection of repeatable determinate processes
What if the majority of business activity is not orderly, repeatable or determinate?
That doesn’t mean EA should abandon its aims or approach; it only means that EA is not a theory of everything;
It can be supported by other approaches that focus more on the human condition.
Consulting and service provider organisations are often project oriented.
In these organisations, employees have a reporting line in their own organisation.
But for most work, they report to a project manager employed to manage work done for a customer, or directly to a customer.
So they have an alternative management reporting line to the chief of the customer’s business.
This leads to what is called “the consultant’s dilemma”
Consultants are destined to walk the thin line between what is best for their company and what is best for the client.
Footnote: Creative Commons Attribution-No Derivative Works Licence 2.0 30/12/2015 12:28
Attribution: You may copy, distribute and display this copyrighted work only if you clearly credit “Avancier Limited: http://avancier.co.uk” before the start and include this footnote at the end.
No Derivative Works: You may copy, distribute, display only complete and verbatim copies of this page, not derivative works based upon it.
For more information about the licence, see http://creativecommons.org