Category Archives: 3 Asymmetries

The different ways of managing the supply-demand relationship.

On stratification

by Philip Boxer PhD

Why the interest in stratification?
A colleague, Simon Western, referred me recently to Actor-Network Theory and the work of Bruno Latour in the context of a conversation about the behaviour of health care networks.

  • His point was that the value of this approach was in the way it focused on ‘following the interactions’, including in a network anything that “modifies a state of affairs by making a difference”.[1] Physical objects that constrain or enable interactions are thus just as important in defining a network as are human actors.  His observation was that standardisation, while restrictive if it became overly prescriptive and bureaucratic, could be liberating, “underpinning all successful business networks”.  Just enough standardisation was the key, for without minimal standardised interfaces, nothing happened.  His examples were:
    • Ryan air- the biggest airline in Europe, grown from nothing on the basis of standardisation creating behaviour change from all stakeholders, to grow a huge network of passengers and air travel;
    • McDonalds- standardised processes, fueled by entrepreneur franchise ownership + fantastic supply chain networks
    • Facebook and Twitter – phenomenal growth through simple standardised frames for individuals to fill with their own personalised content
  • I commented that I thought standardisation per se was insufficient in understanding what made networks effective, the issue being to understand their stratification.  Thus standardisation operated within some of the strata of a stratification and then in different ways depending on the nature of the relationship to demand that the network as a whole was mobilising.[2]
  • His response was that my use of strata and layers spoke more of an engineering project, geology or construction site rather than the fluid complexity of networks. Following Latour’s understanding of networks, his point was that “standardisation and structure are actants within networks, but not the architecture of them.”

So here we were with what looked like a disagreement – follow the networks of interactions and the worlds they construct versus how are the worlds of networks built? In what follows, I explore the ways in which there is no disagreement between these positions.

Follow the networks of interactions and the worlds they construct
In relation to what does a network form?  A useful place to start here is the social object.  An example of a social object would be the condition of a patient[3], but the condition encountered as an event – some singular moment in which there is something about the condition that disrupts existing understandings and/or irrupts in a way that insistes on being attended to [4].  A social object represents a particular affective relation to a situation in which some aspect of the situation itself is experienced as complex, question-generating, endlessly unfolding and incomplete.  The social object is to be distinguished from a ‘real’ object, being like a flag around which people may gather allied in relation to the situation that the flag signifies.[5]  Its efficacy in serving as a social object depended on there being a fit between the nature of its incompleteness and the individuals’ own experience of lack – an identification between some aspect of an individuals’ unconscious lack and the imaginary form given to it by the social object.[6] This ‘gathering around’ takes the form of a network of interactions that includes not only people as actants, but tools, technologies, ways-of-thinking and anything else that enables a current state of affairs to be modified by the differences it makes in the interactions. In the case of our patient, it is hopefully a gathering around the cause of addressing his or her condition.

Latour introduced the notion of punctualisation as a way of thinking about how actants are related to as ‘black boxes’[7]:

the way scientific and technical work is made invisible by its own success. When a machine runs efficiently, when a matter of fact is settled, one need focus only on its inputs and outputs and not on its internal complexity. Thus, paradoxically, the more science and technology succeed, the more opaque and obscure they become.

The definition of these black boxes ‘punctualise’ the actants between which interactions are taking place, and when there is some breakdown in the interactions, such black boxes are ‘depunctualised’ in the sense of being opened up themselves as a network. It is this relationship of embeddedness of networks as elements of networks that is described as a ‘stratification’.

Latour further distinguished a ‘real’ object “not by virtue of being tiny and fundamental, but by virtue of having an intrinsic reality that is not reducible to its sub-components or exhausted by its functional effects on other things.”[8] This is consistent with a view of all objects and systems as forms of novel emergence, even though it is convenient for many of such objects to be considered ‘objective’ in the sense of their existence being inter-subjectively agreed as independent of the observer .[9]

So now we have networks, each one formed by interacting actants allied by a relation to a social object the relation to which operates as the (final) cause of the network.  Such networks are stratified by the ways in which they are constituted in relation to embedded networks that are ignored so long as the network as a whole performs as expected in relation to its cause.

How are the worlds of networks built?
The ambiguity in this heading is intended.  We are interested both in finding ways of describing the way networks are punctualised into strata, and also interested in how new forms of punctualisation become possible in pursuit of new kinds of effect.  Latour identified a second type of  ‘intentional’ object that “has no interior of its own, but exists purely on the interior of some other object”.[8] The descriptions of interacting actants from which stratified networks are constructed have this intentional nature.[10] Looked at like this, it is possible to see why Simon was concerned.  In terms of the following diagram, by seeking to identify the ways in which structures might shape the ways in which actants interacted with each other, we would also be creating new understanding of the actants within the network itself. [11] How so?

actantsSuch structure structures an actant’s way of understanding his or her interactions with the network. The actant is subject to this structuring, over-determining the way that they attribute ontic status to their constructions.[12] Using this approach, three kinds of depunctualisation can be articulated in the ways in which these structuring structures structure the way actants’ constructions are made. These depunctualisations are particular to the actant’s subjection to the structuring structure that they embody.

These depunctualisations are referred to as asymmetries, producing a stratification of six types of embeddedness.  When this stratification is projected into the actant’s constructions, they articulate the embeddedness of underlying technologies in relation to the social objects of actants embodying demands within their contexts-of-use[13]:

  • The first asymmetry, separating a product/service from the technology embedded in it.
  • The second asymmetry, separating a solution delivered to a customer from the business organisation embedded in its processes of delivery.
  • The third asymmetry, separating the customer’s experience of the solution within their context-of-use from the customer’s demand.

When the delivery of a product/service, solution or experience is by a network that can be identified with a single organisation, its embedded behaviors describe a theory-in-use that may or may not correspond to what members of the organisation say they are doing.[14] When the relationships between these embedded behaviors become fixed by supply-side interests, they are fixed by an accountability hierarchy. [15]  The delivery of a particular customer’s experience is more likely, however, to be identified with a number of organisations operating as a network, describable also as a ‘system of systems’. [16] For such networks to function effectively, there has to be sufficient agility in the relationships between its embedded systems for them to be capable of being aligned dynamically in response to accelerating tempos in the emergence of new forms of demand. [17]

No disagreement?
Why should we not want to think about the ways in which networking is made impossible by the ways in which it is not possible to punctualise? The engineer in me wants to find ways of overcoming such impossibilities. But my colleague is right in pointing out that what always comes first must be the desire motivating the formation of the network.

[1] Bruno Latour (2005) Reassembling the Social: An Introduction to Actor-Network-Theory, Oxford University Press. p71
[2] Standardisation can be applied both to the way supply is coordinated and also to the way demand is defined in relation to the client/customer’s context-of-use. These two forms of standardisation have to be managed over a governance cycle. The examples of Ryan Air, McDonalds, Twitter and Facebook belong to particualr parts of that cycle in which there is competition on ‘customisation’ and/or ‘cost’.
[3] Accountable know-how in relation to the patient’s particular pathway is about much more than cost. The process of healthcare must be (and must be expected to be) a collaboration around a social object (the patient’s experience) in the full sense of the word. See Learning about Clinical Commissioning from the USA
[4] ‘Event’ is being used here “the problem of irregularity and indetermination, of the unforeseen and the unforseeable, of the eventually subversive and the disruptive.” See Parker, I. and D. Pavon-Cuellar (2014). Lacan, Discourse, Event: New Psychoanalytic Approaches to Textual Indeterminacy. New York, Routledge.
[5] This understanding of a social object is written about more fully in The social object – distinguishing Kleinian, ‘real’ and Lacanian objects. In explaining the basis of these social objects, Karin argued that the ‘real’ object came to serve as a social object to the extent that it supported a being-in-relation, mutuality or reciprocity between individuals on the basis of enabling temporal synchronisation or on the basis of establishing a shared temporal immediacy – individuals able to collaborate around a shared task, or individuals able to be present to each other in some situation (in contrast to the more familiar spatial synchronisation and immediacy of a face-to-face meeting). Furthermore, to the extent that this mutuality was experienced, it was experienced as a ‘We’-ness embedding the individual in a larger context, but derived from the nature of the shared situation rather than from an institutional affiliation.
[6] This understanding of the relation to ‘lack’ is developed further in a conversation on the refusal of (symbolic) castration. It is to be understood not in the sense of something unconsciously known but not yet brought to consciousness (an interpretive unconscious), but as something radically unknowable in relation to the unconscious per se – a real unconscious (in the sense of the Lacanian Real, not in the sense of ‘real’ reality). See Soler, C. (2014[2009]). Lacan – The Unconscious Reinvented. London, Karnac.
[7] Taken from Bruno Latour (1999) Pandora’s hope: essays on the reality of science studies. Cambridge, MA. Harvard University Press.
[8] Quoted from Harman, G. (2009). Prince of Networks: Bruno Latour and Metaphysics (Anamnesis). Melbourne, This is teh definition of novel emergence.  See later posting distinguishing novel emergence from hierarchy.
[9] This leads to an understanding of embedded strata of novel emergence as an effect of the interest and capabilities of the observer rather than an inherent property of that-which-is-observed. See Ryan, A. (September 2006). “Emergence is coupled to scope, not level.” Complexity – Complex Systems Engineering 13(2): 67-77. See the following series of postings.
[10] The ‘intentional’ nature of these objects may be understood as themselves networks that, in addition to their synchronic characteristics as a network, also have diachronic characteristics to do with the tempo at which they exhibit their behaviours, corresponding to a timespan of discretion. See Timespan of discretion and the double alignment of ‘know-how’.  For more on the significance of tempo, see [17] below.
[11] Category theory is one medium in which the relationships between these actants may be thought about. See A Categorial expression of Demand Asymmetry.
[12] The nature of such ontic assumptions are described in Describing what is going on (wigo)
The Oxford English Dictonary defined ‘ontic’ as follows: “Of or pertaining to knowledge of the existence or structure of being in a given entity.” Thus any ‘realist’ assertion of ontology is mediated by the ontic assumptions being made by the observer-entity making the assertion i.e. an ontology is built by an entity making ontic assumptions. The 4-quadrant model gives us a way of thinking about what kind of ontic assumptions the entity is making. The concept of the strategy ceiling further elaborates on the way these ontic assumptions are held by an entity in the form of stratified relations between the enterprise and demand.
[13] For more on these layers see 3 Asymmetries.
[14] These behaviors relate to each other in the form of a stratification of nested contexts which places the supply-side behaviors of the enterprise in relation to the demand-side contexts with which it interacts. Where such a relationship does not exist, we may say that the strategy ceiling of the enterprise prevents it. See The strategy ceiling.
[15] East-West dominance means having a business agile enough to support the particular relationships of embeddedness needed to sustain a relationship to the distinct forms of demand arising at its edges. Under these conditions, the 6-layer stratification is no longer usefully thought of as a hierarchy, but rather as a particular structuring of the alignment between supply and demand. Note that it is only by including the third asymmetry that the stratification can no longer be thought of as hierarchy. See When is a stratification not a universal hierarchy?
[16] Such networks involve distributed collaboration in a complex system-of-system multi-enterprise context over which there is no single source of control. See Enterprise Architecture for Complex System-of-Systems Contexts.
[17] The tempo at which an enterprise creates new uses for its systems is different from that of its acquisition or systems development processes. For example, the military continues to confront the issue of how fielded systems can support the agility needed by its deployed forces. This problem of diverging tempos applies to a variety of large-scale, software-reliant enterprises-such as those found in healthcare and digital communications. See Building Organizational Agility into Large-Scale Software-Reliant Environments.

Boundary Perimeter Edge

by Richard Veryard
We can use the three asymmetries to appreciate different strategies for security and trust, such as deperimeterization. First we need some definitions: Boundary refers to a discontinuity in a physical system, Perimeter to a discontinuity in a social system, and Edge to a discontinuity in systems of meaning. As with the asymmetries, these build on each other, so a perimeter includes a ‘virtual’ boundary, and an edge includes a ‘virtual’ perimeter. Thus where we place boundaries, perimeters and edges reflect where we place the three asymmetries. It also determines the way we are able to approach security and trust.

For example, deperimeterization can be understood as an effect of the third asymmetry. A traditional perimeter defence assumes that rights and obligations (social) coincide with certain physical divisions (boundaries). Deperimeterization means it is no longer feasible to align the levels of security with the social boundaries, because the social system is itself losing its cohesion under the influence of the third asymmetry.

Assuming symmetry means being able to run something as a closed system – the way it interacts is wholly defined by the supply-side, so control is possible. With the breaking of the first symmetry, the use of the technology is defined by its outputs, and not its internal functioning. But we can still apply a fortress approach to this, so long as we can wholly define the boundary across which the outputs are to be provided. The metaphor here is the fortress.

With the second symmetry being broken, our business changes from being defined by its outputs to being defined by its ability to organise business processes that deliver solutions. But the supplying business organisation is still in control of this, although the complexity of what is ‘inside’ is greatly increased by its now explicitly socio-technical nature. The fortress metaphor is still possible here, but understood now in terms of a dynamic frontline (e.g. Nato warfare across Europe).

It is with the third symmetry being broken that we get the necessity for defence in depth (they can strike from anywhere), asymmetric threat (they can play by their own rules), and agile/manoeuvrist conflicts that require power-to-the-edge and synchronization at the edge. This is the environment in which collaborative composition is necessary because of the complexity of the demand environment which you are trying to interact with. (The military metaphor here would be “operations-other-than-war” where you have to work with the inhabitants etc.) It is this latter third symmetry-breaking that creates the de-perimeterization effect.

Why is a stratification not a universal hierarchy?

by Philip Boxer

In describing the 3 asymmetries, Richard establishes a six-layer stratification relating underlying technologies to ultimate contexts-of-use. Thus in the case of orthotics, if we approach it from the point of view of a manufacturer of orthotic footwear, these layers look like increasingly general descriptions of the contexts within which the underlying technology will come to be used: technology=soles, product=footwear, business=footwear-to-order, solution=fitted footwear, customer demand=orthopedic patients, customer experience=difficulties in mobility. Is this therefore not just a hierarchy moving from the particular of the technology to the general of its uses?

If, as a supplier, we want to take a symmetric view of demand, then this is true – ‘orthopedic patients with difficulties in mobility’ is a general definition of the footwear manufacturer’s market.

But in distinguishing the third asymmetry we define the relationship to demand as being to a particular context-of-use that demands a particular form of orchestration and composition of services and products in order to satisfy it. Thus if we take up the perspective of the customer experience=the patient’s experience of living with their condition through its life, then the customer demand=that treatment for my condition that will have the greatest impact on my through-life experience at this time, and the solution=the treatment that is fitting for the current situation within its through-life context. Not surprising, then, that one of the major issues faced when insourcing clinicians employed by the manufacturers was how to reflect the through-life dimension of performance in the way the clinical service was contracted.

In finding the edge, we describe the particular form of orchestration and composition needed in response to asymmetric demand in terms of a wedge of services that needs its own four-colour model of how it is aligned to demand. Thus for our patient, the customer experience is in the black quadrant, the customer demand and its particular solution in the red quadrant, the business and its product(s) in the white quadrant, and the technology in the blue quadrant. East-West dominance means having a business agile enough to support the particular white-red organisation needed to sustain a relationship to the distinct forms of demand arising at its edges. But now the 6-layer stratification can no longer be thought of as a hierarchy, but rather as a particular structuring of the alignment between supply and demand – something more horizontal than vertical.

The way we understand the four-colour model is therefore central to the way this alignment is defined. In order to be able to construct it, three distinctions have to be made:

  • Internal//External: what is internal to the way we do business vs what is not. This distinguishes the provider of the insourced clinical service from the environment into which the service is being provided.
  • Viability//Identity: the way things work vs what determines the shape of the way things work. Clinicians learn about how orthoses are made and how they can be used on the musculo-skeletal system (the way things work), but the particular ways these are shaped depend on the patients’ characteristics and the way the manufacturer chooses to do business.
  • Addressed//Ignored: the domain of reality (later called the domain of relevance) being addressed vs not addressed. The domain defined from the point of view of the manufacturer is going to be much narrower than that defined from the point of view of the patient’s needs. The way the domain is defined is fundamental to governance-at-the-edge, and implicates the ‘I’ of the beholder. Thus when demand is assumed to be symmetric, the ‘I’ can be the view from the top/centre of the supplying business. But when it is assumed to be asymmetric, the ‘I’ must be defined collaboratively through the way the relationship at the edge is itself constructed.

The 3 asymmetries and their associated economies correspond to the relationships between the blue-white, white-red and red-black quadrants, accounting for the particular way the quadrants are held in relation to each other. By including the third asymmetry, the stratification can no longer take the form of a universal hierarchy, but instead must be particular to the relationship to demand. It is this which presents the business with its double challenge, and the necessity to shift from an object-oriented to a subject-oriented approach to modeling the relationship of the enterprise to the demands of its clients.

rcKP – value propositions at the edge

by Philip Boxer

How are we to think about the services offered by a business taking power to the edge? One way is in terms of the three asymmetries.

As pointed out here, the economies associated with the first two asymmetries can be secured under conditions of North-South dominance. This means that their profit potential is defensible because the knowledge associated with creating them is asymmetric on the side of the business: the business has something that both its competitors and client-customers do not.

In contrast, the third asymmetry requires East-West dominance capable of delivering an appropriate degree of intensity in the relation the business has to the client-customer’s value deficit, defined in terms of the client-customer’s effects ladder. This intensity reflects the degree to which the business is engaging with the asymmetric nature of the client-customer’s demand – asymmetric on the side of the client-customer. From this we can derive four kinds of value proposition, the first two of which assume no relation to the client-customer’s value deficit:

  • r-type: The presumption is that demand is symmetric, and therefore the value proposition is to replicate the offered product or service in as many variants and forms as can be profitably sustained, based on its ability to capture economies of scale. (e.g. pharmaceutical products, telecoms equipment)
  • c-type: The value proposition is to offer a combination of products and services that can be dynamically customised in relation to the customer’s demand, based on the ability to capture economies of scope. (e.g. providing injections, or telecoms connections). The value proposition is in delivering a customised product/service where and when it is demanded.

The other two services all have varying degrees of involvement with the client-customer’s context-of-use:

  • K-type: the value proposition is to offer the know-how needed dynamically to orchestrate and synchronize the use of products and services in collaboration with the client-customer in solving some part of a larger problem that the client-customer is currently experiencing. (e.g. managing an episode of care, or the way connectivities can be made available to a business in particular types of situation).
  • P-type: the value proposition is to offer the ability to work with the client-customer on some area of pain that they are currently suffering, in order to find a way of making it tractable. (e.g. diagnosing what kind of treatment is needed, or defining what kinds of connectivities a business needs).

We can combine these into a diagram that shows the different types of value proposition as a cycle which may or may not end up in the r-type zone:


In each cycle, there is an initial (P-type) proposition that develops with the client-customer a way of addressing its need, experienced by the client-customer as a value deficit (red circle). As the client-customer learns this way of organizing its demands for itself, this may become a (K-type) service managing how particular aspects of those overall needs are being addressed (yellow diamond), these ‘particular aspects’ being the customer situations within the effects ladder generated by the P-type proposition.  This will happen if there is some aspect of the K-type know-how that is also defensible. This K-type know-how may become a (c-type) service that the client-customer includes as part of how it takes up this K-type know-how for itself (green triangle). Or the service may end up becoming commoditized and defined purely in terms of its means of production, becoming an (r-type) service (blue square).1

The precise dynamics of these cycles, and the mix of rcKP value propositions offered by a business, will depend on the particular demand and competitive conditions encountered. What I have described, however, are the different kinds of service needed depending on the way in which a client-customer is choosing to internalise learning (or not) as it responds to some aspect of its own particular value deficit.

It follows that if a business is to be able to sustain power at its edge, then the services it offers will involve some mix of K-type and P-type services. The interesting thing about this mix is that its dynamic and collaborative nature makes the services necessarily relational, creating (at least) two-sided markets in the relationships a business has with its client-customers.

[1] More is written about the rcKP cycle in Understanding Value Propositions and Effects Ladders.

Enterprise IT

by Richard Veryard

In his Confused of Calcutta blog, JP Rangaswami (now CIO of BT’s Services Division) picks up a definition of Enterprise Architecture from Andrew McAfee: “IT that specifies business processes”.

JP argues that

“… enterprise systems work well only when there are rigorous standardised processes; they work well when these rigorous standardised processes are industrial strength, with external frames of reference; they work well when the number of processes is kept to an absolute minimum, and where process divergence and diversity is avoided.”

There may well be situations where it makes sense for the business process to be specified (=overdetermined) by IT. But there are other situations where this doesn’t make sense at all, because of the asymmetrical relationship between business and IT.

Are there intelligent ways of implementing enterprise IT without constraining (=overdetermining) the business, thus respecting the need for agility? We believe there are. But it’s probably not going to happen if it is left up to those “respectable” consultants who are the target of JP’s criticism.

Creating Economies of Alignment

by Philip Boxer
As we develop our understanding of the three asymmetries, it is helpful to associate them with three corresponding forms of economy that their management generates. The first two of these correspond to N-S dominant forms of governance, while generating the third necessarily involves E-W dominance:

  • Scale: The ability to create additional output from an existing capability, reducing average unit cost. (i.e. producing more output from the same technology infrastructure).
  • Scope: The ability of a business to extend the scope of its operations across different markets reducing average operating costs. (i.e. covering more markets with the same business process infrastructure).
  • Alignment: The ability to create additional ways of organising the business relationship with a customer over time, reducing the average cost of alignment of business operations to the dynamics of each customer relationship. (i.e. managing more distinct customer relationships over time supported by the same infrastructure).

Why take power to the edge? Because it generates economies of alignment in the management of a relationship over its life.

The three asymmetries

by Richard Veryard
Demand Asymmetry means that the forms of demand are increasingly specific to the context in which they arise.

The first asymmetry involves separating out technology from the supply of specific products. This requires modelling of possible behaviors that can be supported (so Microsoft or car manufacturing has to modularize itself in support of families of technology use).

The second asymmetry requires separating out business models that can organize supply from the solutions that are on offer. This requires modelling of the possible forms of business geometry (so rail maintenance or retail services have to use a franchise model to allow the variation in business organization to accommodate the variety of ways in which the service needs to be implemented).

And the third asymmetry requires separating out the different contexts of use within which a demand arises. This requires modelling of the possible forms of demand within the contexts in which they arise (so that financial or care services are having to take up the way the through-time wealth/conditions are managed in a way that responds to different forms of context-of-use). The result is a stratification that describes six layers of organisation through which underlying technology is brought into relation with ultimate context-of-use:

It is worth considering what happens if these asymmetries are ignored.[1]

  • In the first asymmetry, this means defining the product by the technology. This is typical of the early stages in the emergence of new technologies. (Do you remember how we used to have to use mobile phones?).
  • In the second asymmetry, this means defining the solution for the customer by the way the business is organized. (Do you remember how large businesses used to relate to their customers before CRM?).
  • And in the third asymmetry, the solution to the problem presented by the customer is assumed to be what the customer actually needs. (Have you ever received a prescription from the doctor that turns out only to treat the symptom?).

Source: Metropolis and SOA Governance. Part 1: Towards the Agile Metropolis

[1] There is a fourth ethical dilemma, therefore, that is the dilemma of whether or not to hold the other three dilemmas open…

Distinguishing the 3 Asymmetries

by Philip Boxer
Charlie, Referring to your last blog on responding to diversity in value models, you are of course right that it is not possible to consider the supply-side without at least implying a particular relationship to the demand-side. What is interesting is the supply-side logic that governs the way in which the supply-side looks at the demand-side. Thus a car manufacturer ‘projects’ onto the demand-side all kinds of assumptions about what it is interested in. In asking whether it is useful “to look at the demand-side variation independently of the supply-side”, I was asking under what circumstances the demand-side logics governing use could be considered independently of supply-side logics… arguably any market where there is a significant degree of standardisation and commodification is one where this separation of logics is taking place. Thus managing mobility separates from the choice of automobile or airline, and managing connectivity separates from the choice of particular digital platform (qua iPod, mobile, handheld etc).
It was for this reason that Richard and I tried to separate out the three asymmetries in the Governance paper. Clearly the dilemmas and trade-offs you raise dominate when dealing with the first asymmetry (the technology is not the product) – as you graphically illustrate with the example of the SUV. This is also true when dealing with the second asymmetry (the business model is not the customer’s solution), but to a lesser extent. Thus with the second asymmetry, the low fares/low cost business model established an alternative solution to that offered by United, American etc, confronting those airlines with the necessity of using a varying business process model depending on the characteristics of the market segment being addressed. It is managing this second asymmetry that has dominated much of the flattening/restructuring/ outsourcing of businesses in the last 25 years or so (and perhaps the remarkable increases in the productivity of US businesses).
And you are right to point out that the equilibrium established in relation to any given asymmetry reflects the economies of scale and scope that limit how much diversity can be spanned – modified of course by the constraints placed on competition. It is managing this constraint that your paper addresses so well, and which ultimately necessitates the dominance of supply-side logic.
But it is with the third asymmetry (the customer’s demand is not the customer’s experience) that something new happens. Thus the customer’s demand may be necessarily dynamic (the way they use ICT in support of their business), or necessarily emergent (the way the patient’s chronic condition develops over time), or just idiosyncratic (each client wants their investments managed in a way uniquely suited to their family situation). In each case, there is a need to orchestrate available commoditised services and products and add a new level of interoperation through the way they are composed with the experience/life/situation of the client-customer in a way that is particular to both the time and the relationship – adding a new level of business model that itself becomes real-time (real-time business). This presents a new challenge, particularly in the way the role of software architecture needs to be understood, because from the point of view of any one service or product, these new uses are emergent.
The metaphor I like is that of heart transplants – however well the transplant operation is conducted, and however good the nursing, if the immunological response of the host body is not managed in a way that is particular to that body, the transplant will fail. In the same way, managing the first two asymmetries is necessary but not sufficient for managing the third asymmetry…
It won’t surprise you to know that our work involves supporting the needs of real-time business – defining the strategies, identifying and mitigating the risks, creating the governance. But it is a difficult terrain to distinguish. How are we doing!
next in this thread

Responding to diversity in Value Models

by Charlie Alfred
Philip, Thank you very much for your comments on separating the supply-side from the demand-side. I agree with the points you raised, and have a few observations to share:
1. You observed that my article approaches the problem from the provider-side. Given the nature of my job, my original goal was to teach people I worked with about how to do a better job of software architecture. This is the main reason that the article takes this point of view.
My belief is best expressed by Dr. Russell Ackoff, former professor at the Wharton School of Business. He discusses the sibling processes of synthesis and analysis. He observes that analysis has been the predominant model of thought since the Industrial Revolution. Take a whole, break it down, and study the pieces in isolation.
Ackoff suggests that while analysis helps you to explain how things work, it doesn’t give you enough context to understand why they need to work that way, and what might cause them to change. To do this, you must study how a subject (system) functions in its larger context(s). This begins to expose which expectations, obstacles, and constraints are imposed on the subject.
In short, as a result of this research, I’ve come to the conclusion that an architect needs to:
a) start with synthesis to understand context (both shared and diverse),
b) use analysis to formulate approaches to challenges, analyze trade-offs, and mitigate risks, and
c) use synthesis to consider the feedback effect of the solution on the contexts, etc.
Ackoff’s writings have also convinced me that the scope of a system must expand to include all elements with significant inter-dependencies. The little bit of thinking that I’ve done on the subject of value suggests that the consumer and supplier sides cannot really be considered independently. This is the subject of my next observation.
2. You mention the two perspectives of value: consumer-side and supplier-side and raise the question about whether they can (and should) be considered independently or must be considered together.
As a general rule, I think that they cannot be considered in isolation:

a) In the long run, there is no supply-side value where there is no demand-side perceptions of value. In cases where demand-side utility curves are strong and inelastic, and suppliers have major market power, “long run” can be quite a long time. For example, OPEC can ignore the needs of its customers for price, supply, and to a lesser extent quality. However, this has the side effect of stimulating research into alternate sources. I agree that diversity on the demand side tends to strongly influence market segmentation in more competitive markets. While this happens a lot in dynamic markets, it also happens in ones that were thought to be stable. Southwest Airlines low fares/low cost approach has been extremely effective against United, American, and Delta.

b) Supply-side innovations often change utility curves on the demand side, by changing consumer’s perceptions of what is possible. Today, I can’t write a paper or report without having my web browser open and Google ready to run. My iPod Nano is so small and holds so much music that I want to take it to work. If I’m away from home and need to make a phone call, I pull my cell phone out of my pocket. 10 years ago, I was quite content without any of these things. It wasn’t that I was unaware of the benefits of instantaneous research, portable entertainment, or accessible communication. The main reason I was content with less than that, was that I had no idea that major improvement was possible. However, once exposed to the innovations, my utility curves (and hence, my value expectations) were reshaped for ever.

c) When faced with significant diversity in value perceptions by consumer, a provider can be faced with a difficult juggling problem. If the provider attempts to craft individual solutions, its development and operating costs can go up. If the provider attempts to create a solution that can be tailored to each context, it risks creating something that doesn’t fit any of them very well. Consider a sport utility vehicle. Buyers who like to drive it off-road want a lot of ground clearance. Suburban moms like the fact that a higher clearance gives them better visibility of the road. However, a higher clearance also tends to mean a higher center of gravity, which increases the risk of rollover in sharp turns. This is not much of a plus for the driver who wants the all-wheel drive for driving in snowy and icy climates.

In summary, the consumers and providers both win when the provider understands the value models and challenges well enough to combine the right set of diverse needs into a solution, while omitting the ones with incompatible challenges.

3. I agree completely with your point about Porter’s use of horizontal and vertical linkages. To me, the notion of linkages was key, and differentiating between those that are internal and external was less significant.
I find it more interesting that the “architectural decisions” of the firm (value change) are what determine its cost and differentiation drivers. For example, an airline like United has chosen a hub/spoke architecture and a variety of airplane types. This lets it have very frequent departures and match aircraft sizes to the flight lane. OTOH, Southwest standardizes on a small number of aircraft and prefers direct flights. This lets it reduce the costs of training and maintenance, and reduce the cost of operating large hubs. The different business architectures appeal to diverse sets of passengers, because the passengers have different utility curves. In effect, linkages tie directly back to the value models of the consumer(s) and provider(s) and the constraints
that the environment enforces on its own (e.g. how far can a 727 fly without refueling?). As a related note, I love Christensen’s concept of discontinuous innovation.

In summary, I believe that necessity is the mother of invention, and value model diversity on the demand side will force the supply side to better satisfy it. It may not be a painless transition, because a lot of firms haven’t learned how to:
a)Forget about their own agenda, and immerse themselves in their customer’s context
b)Learn how to be jugglers and figure out which things can be juggled together and which cannot.
Let me know what you think.
next in this thread

Value-Driven Architecture

by Charlie Alfred

by Charlie Alfred
I am the author of an article on Value-Driven Architecture that was published in the same issue of Microsoft Architecture Journal as the one that you wrote on SOA Governance. I finally had the opportunity to read your article during the past couple of weeks, as well as the earlier one you reference (Metropolis), written by Pat Helland.
It was very interesting to see how you both took the concept in different directions. Pat focused on the service-provider side, and emphasized the benefits of standardization and high-speed transportation (networks). You took a more systemic view and emphasized how diversity often is not limited to service
provider implementations, but instead reaches more deeply to affect the service API’s as well as the workflow management (or other control structures).
Someone at the 2004 Software Product Line Architecture conference in Boston made a very interesting observation:
“The ROI of a software product line comes from leveraging the commonality. However, you cannot achieve this effectively until you also identify and manage the variability.”
I believe this quote sums up a lot of the philosophy difference between Pat’s article and yours. Pat seems to be making the “benefits of leveraging the commonality” argument, while you do a very good job of articulating how difficult “managing the commonality” can be.
Both of these notions are essential and relate back to one of the central themes of the article I wrote on “Value Driven Architecture.” In order to know where to accommodate diversity and where to leverage commonality, you must have a way of identifying and contrasting context-specific challenges
(where each context consists of a set of like-minded stakeholders who are affected by equivalent obstacles and constraints). By contrasting the key challenges and their priorities side-by-side, the architect begins to clarify which challenges are similar enough to leverage with a standardized approach, and which ones require more diverse approaches.
Again, thanks for the thought-provoking article.
next in this thread