Q&As on Organizational System Design

Q: What are typical failure modes in designing organizations as systems?

A: The two most serious errors are also among the most frequent: leaving ambiguity in the value proposition that becomes the organizational design point, and failing to validate the benefits of the value proposition with the customer. Other failure modes include: over or under specifying the levels of decomposition of the value proposition into subordinate component roles; making roles too granular or not granular enough; populating roles with the wrong people; maintaining command and control governance, and reluctance to re-design when unanticipated change requires it.

Q: What is the difference between a process design and a system design?

A: Unlike process designs, which sequence action, system designs have no time dimension. They structure the effects of action, and do so in a way that guarantees synergy and alignment of organizational parts around a common purpose. They provide explicit knowledge of how the parts interact to achieve that purpose. Lacking a system design, massive “bolt-on” inefficiencies flourish: suboptimization, mismatched priorities, incompatible local objectives, and destructive internal conflicts. The organization becomes a Petri dish for breeding unintended consequences.

If customer preferences are sufficiently predictable, and the rate of innovation in capabilities sufficiently incremental, process designs can help….. by showing how an existing organization might operate more efficiently. But process descriptions do not shed light on how the organization should be structured. For instance, the engineering drawings for an automobile, like an architect’s plans for a house, are systems designs. They say nothing about the sequence of tasks required to either build or use the object being designed. What they do show, that process designs don’t, are the requisite constituent parts, how they relate to one another, and the function of the completed car or house.

Equally significant in environments of rapid change, a process design requires more predictability: the designer must specify in advance the inputs, the outputs, the contingencies, and the sequence of actions and choices to be made. A system designer need only specify the effect to be created, and the exchange of outcomes between elements.

Q: Why should I design my organization as a system if it already is one?

A: There is indeed a system somewhere in your company that creates customer effects. But because it wasn’t designed, it is virtually impossible to find; and if found, would undoubtedly look like a Rube Goldberg production. There are inevitably redundancies, irrelevancies, conflicts, and suboptimization that a knowledgeable system designer would never tolerate, much less specify. To clarify how a system can exist even when buried in a kludge, consider this analogy:

Think of a car that gets you back and forth to work most days, but only one cylinder is working, the head gasket leaks, the brake pads are shot, the odometer is missing, you can’t shift into third, and you are getting 5.7 miles per $3.00 gallon of gas. Further imagine that there are no designs for the car, only an operator’s manual and an organization chart that shows, for example, that the brake pads report to the wheels, who report to the axles and are “matrixed” to the tires and the generator. (The radiator, which you might expect to report to the engine block, actually reports to the differential in order to satisfy span-of-control policies). How would you go about fixing this car? Possibly by trying to replace each defective part with a “best of breed” substitute, only to find after the third part is replaced that the car no longer runs at all, because the new parts don’t work with each other or with the existing parts….and you couldn’t know this in advance because you don’t have a design that shows how the parts are supposed to work together. The jalopy that got you to work some of the time was a system – a lousy system – but a system nevertheless because it was capable, on good days, of performing its essential function of getting you where you wanted to go. But if you were asked to diagram the system, you couldn’t—because you wouldn’t know which parts are essential, nor how they relate to one another.

Q: Even in today’s world of rapid and discontinuous change, there are still outcomes that are predictable and stable. Why shouldn’t these be designed as processes to maximize efficiency?
A: They should. But they will usually not show up on a R&A design, which is concerned with outcomes, not the activities –ad hoc or procedural—that created them. However, if the designer wishes, terms and conditions that constrain HOW the outcome is to be created can be specified and associated with a particular (always negotiated) outcome.

If customer preferences are sufficiently predictable, and the rate of innovation in capabilities sufficiently incremental, process designs can help an existing organization operate more efficiently. So an individual accountable for a particular outcome may decide to deliver it using a process design, or automating it. The accountability remains the same; but the actions to take it can be left up to the individual in the role.

Q: Matrixed organizations link people whose work is related. Why is this unsystematic?
A: A typical organizational design depicts a series of vertical hierarchies of authority linked by a dozen or so end-to-end processes and “matrix” lines that show communication and/or approval relationships with other units. This is an explicit recognition that interdependencies exist. But many of the matrixed individuals have incompatible metrics and compete with one another for such things as budget and revenue/profit credit, causing them to suboptimize around their own objectives. Suboptimization is the hallmark of anti-systemic behavior.

Q: Is it necessary to get rid of the traditional organizational reporting structure when implementing a R&A design?

A: No. An analogy is the creation of a varsity football team comprised of players from multiple fraternities and scholastic disciplines. Winning football strategies are “X and O designs” that depict the accountability of players in positional roles. Their social and scholastic associations should be irrelevant when they are on the field of play.

Q: How can an organization designed as a system be embedded in a larger organization that isn’t?

A: By defining the relationships to other organizational entities in terms of outcomes (not activities), and requiring that changes to these outcomes and/or to the conditions of satisfaction associated with producing them are negotiated — not unilaterally imposed. For example, if a budget cut is necessary, modifications of the outcome(s) committed must be renegotiated.

Q: Can customers and third parties be incorporated in an R&A design, and should they be?

A: Co-developing value with customers is a concept that goes back at least to the “prosumer” anticipated by Alvin Toffler in his 1982 book The Third Wave. But for many organizations it remains more of an exhortation than a competence….an objective with little conceptual or practical understanding of how to realize it.

The rule, rather than the exception, is that customer roles and capabilities are essential to realizing a beneficial effect. The same is true of third party contributions. Therefore, the decisions made in decomposing the effect (value proposition) should incorporate the outcomes required of all entities. After the decomposition is complete, a determination of who is best qualified to occupy a given role is made. The result is an organizational systems design whose components are resourced from a variety of institutions. A good metaphor for this is a varsity football team whose roles are populated by players belonging to a variety of fraternities and other campus institutions.

There is obviously great competitive advantage in being the “value architect” of an extended-enterprise organizational design. That is why acquiring Role and Accountability design skills is identified as a core competence of post-industrial managers.

Q: What is necessary to make an organizational system design adaptable?

A: Modularity — the ability to dynamically reconfigure the organization into responses specific to the current situation is the key to adaptability. This is achieved in Sense & Respond organizations by 1) using negotiated commitments to outcomes as the way to link, unlink and re-link individuals in customer-supplier roles; and 2) using a common and rigorous commitment management protocol for the negotiations. The protocol must enable real-time tracking of the status of all commitments, and ensure that organizational policy constraints are propagated and complied with as conditions of satisfactory completion of all outcomes.

© 2010 Stephan H. Haeckel. All rights reserved.