19.1 Against planning as a product
In The Heart of Enterprise Beer asks what holds several levels of recursion together, and he knows his answer will surprise: the binding process “is none other than PLANNING. Many a treatise has been written about planning; but I do not know of any that contends that it is actually the ‘glue’ of organizational cohesion” (HOE 1979: 336). His treatment of planning is short, but a large part of his theory of organization rests on it. To make the claim plausible Beer first has to clear away the ordinary picture of planning, and he states that picture precisely. Treatises on planning, he writes, usually assume “that planning is an activity carried out by professionals called planners, that this activity results in products called plans, and that these plans will then be either implemented or not by those in authority” (HOE 1979: 336). He asks the reader to check this against experience: “(Please reflect for a moment: is that not indeed the common belief?)” (ibid.). Between this common belief and the view he is about to defend there is, in his words, a cognitive gap: “It is not so much a gap, as a chasm” (ibid.). The common picture has three parts: a group of people (planners), an object (the plan), and a later step (implementation by someone else). Beer takes each part apart. He starts from where planning actually happens: “What needs to be reiterated is that planning happens only when there is an act of decision. […] It follows that the only planners are managers, namely those people who are entitled to commit resources” (HOE 1979: 337). Between these two sentences he gives the reason: the act of decision “commits resources now, so that the future may be different from what would otherwise have simply happened to us” (ibid.). A document describing a desired future is therefore not yet planning. Planning occurs when someone with the authority to commit resources does so. Beer had already made the point in the summary at the start of Part Three of The Heart of Enterprise: planning “needs to be seen as a continuous, managerial, process of decision, whereby allocations of investment are made now, so that the future may be different” (HOE 1979: 276). The professional planner may help, but he or she does not plan. Beer draws on two points that, he says, others made before him. The first “is that planning is a continuous process”, which “conflicts violently with the stereotype of a plan based on the next month, the next year, the next five years, or ‘the year 2000’” (HOE 1979: 336). He adds, briefly: “Nature does not have a calendar” (ibid.). The second is the one just stated, that planning is an act of decision.
19.2 Plans must abort
From these two points Beer derives a third, which is his own. Decisions have to be adjusted continuously, “because rational expectations and probable scenarios are both constantly changing, so that the decision taken yesterday is very probably and ostentatiously wrong today” (HOE 1979: 337). What drives the change is the arrival of information. Because information keeps arriving — “‘information is what changes us’” — “plans must continually abort. Produce a plan, dated so-and-so, which has today been implemented in every detail, and you will see a stupid action” (HOE 1979: 337). The quoted phrase is Beer’s own, from Platform for Change (see chapter 18.5). The structure of the argument is simple. A plan fixed on a given date embodies the information available on that date. By the time it has been fully carried out, more information has arrived. If none of it has altered the plan, it has been ignored. A plan implemented in every detail is therefore evidence that the organization did not learn while implementing it. Beer states the consequence without softening it: “There is nothing more hair-raising than to see an implemented plan — unless it is to see honours and awards heaped on the idiots who went through with it” (ibid.). The summary at the start of Part Three says the same in more measured terms: “since at any moment more information is available than at the previous moment, it is madness to implement plans — those considered as professional products — at some date later than they were conceived. They will necessarily be mistaken” (HOE 1979: 276). And it adds the cybernetic formulation: “In cybernetic terms, plans should be continuously adaptive to fresh information” (ibid.). Beer gives two examples. The dome of St. Paul’s “would have collapsed without the last-minute emergency action that put a vast chain around its circumference to hold it together”, and Concorde, whatever its technological and aesthetic merits, “loses a great deal of money for every mile” (HOE 1979: 338). He is careful about what they show: “The argument is not against St. Paul’s or against Concorde” (ibid.). The dome stands because someone changed course at the last minute; realized projects come out of continual adjustment, and plans carried out as drawn are the exception: “planning is a continuous act of adaptive decision that continually aborts” (ibid.). Beer then sets out three counter-theses, each answering one element of the common picture (HOE 1979: 338):
- “planning is not an activity carried out by professionals called planners; it is an act of decision carried out by managers empowered to commit resources to a different future.”
- “Planning is not an activity resulting in products called plans: it is a continuous process, whereby the process itself — namely that of aborting the plans — is the pay-off.”
- “The plans do not have to be implemented or not by those in authority: what the authority does constitutes the plan — and its realization.” The second counter-thesis goes furthest. It puts the value of planning in the repeated act of abandoning the plan. The third removes the separation between plan and implementation altogether: what is done is the plan.
19.3 Planning as the “glue” of cohesion
With this redefinition Beer returns to his opening question. “And in that planning process the manager dispenses the cohesive ‘glue’. The institution hangs together because, and only because, plans are constantly aborting” (HOE 1979: 338). Planning is thereby tied to the Law of Cohesion (see chapter 13). The summary at the start of Part Three states the connection: the planning process “is perceived as the cohesive ‘glue’ that holds the embedments of viable systems in place”, and this examination “leads to the ennunciation of the Law of Cohesion for multiple recursions of the viable system” (HOE 1979: 276). The following reasoning is my reconstruction, not Beer’s own derivation. Each level of recursion is a viable system with its own intentions. What holds the levels together is a process in which intentions are continually adjusted to each other, rather than a fixed structure of dependency. Planning, understood as continuous decision, is that process. If plans did not abort, the levels would drift apart, each pursuing a plan fixed at a moment when the others were in different states. The chain around the dome of St. Paul’s is a physical picture of the same thing: cohesion supplied by a late decision. Beer’s own derivation of the Law of Cohesion runs through variety rather than through the timing of plans. He builds it on his diagram of the planning process between two levels of recursion (Figure 61, HOE 1979: 353) and states it as: “The System One variety accessible to System Three of Recursion x = the variety disposed by the sum of the metasystems of Recursion y” (HOE 1979: 355). Chapter 13 treats the law itself. Several rules follow from recursion. The first concerns uniformity: “no uniform planning rubric can be imposed” (HOE 1979: 340). Beer gives the reason: a uniform rubric “would very seriously impair the autonomy of the System One management, whose own needs (horizontal) are quite different from the needs of the metasystem (vertical)” (ibid.). Chapter 6 of The Heart of Enterprise made the same argument about freedom: the metasystem should reduce the variety of the operational units only as far as cohesion requires (see chapter 17.2). Beer notes that imposing a planning rubric “in the supposed interests of homogeneous planning across the institution” is nonetheless “normal practice”, and that in his observation this is where the troubles usually start (ibid.). The second rule concerns the content of plans: “Plans are the embodiment of the intentions subscribed to at each level of recursion: since these intentions are constantly changing […], they must continually abort” (HOE 1979: 342). What, then, should be measured and reported? Beer’s answer is narrow: “The only characteristic worth detecting, measuring, reporting, or doing anything about, is incipient instability in the system” (ibid.). It matches his description of the manager’s job in a well-designed system: “It is to be alert to incipient instability — and to take action about that” (HOE 1979: 276). Cyberstride had been designed in Chile to do exactly this: filter the routine and raise an alert when an index moved unusually (see chapter 16.3). The third rule concerns System Two. Different levels and units will describe their intentions in different terms. System Two must not settle this by issuing “a book of rules”, because “there is no need for a uniform language, but only a need for System Two to speak all the languages of System One” (HOE 1979: 344). Beer gives a second reason in the same sentence: any such manoeuvre “would be perceived as the introduction of a planning rubric on the command axis” (ibid.). System Two’s role, as he restates it, is anti-oscillatory: it “exists to damp oscillation” (ibid.). The Second of the five principles written for Chile makes the same point (see chapter 16.6).
19.4 The same argument for society
Six years before The Heart of Enterprise, in the last Massey Lecture of 1973, Beer had already applied the argument to public planning. He observes that planning has acquired a bad name, and asks why: “The reason surely is that our plans are not adaptive, and the institutions charged with making them grind on with their implementation long after it has become obvious to everyone who will be affected that the plans are inappropriate” (DF 1974, Lecture 6). He ties this to the diagnosis of the whole series: “We are back to the unviability of the institution again” (ibid.; see chapter 17.1). His remedy is the one the later book gives, stated more briefly: “Planning should be continuous and adaptive. Societary plans should continuously abort, and be recast, before they give birth to a monster” (ibid.). He then lifts from planners a burden usually thought to rest on them, the burden of prediction. If plans abort continually, “there is no need to base them on the predictions that no-one can correctly make in any case, but only on the analysis of an unfolding situation in which every decision constrains future variety” (ibid.). Earlier in the same lecture he had disclaimed any intention to forecast: “I do not believe that we can predict the future. I believe instead that we can describe the present with perspicuity, if we use the proper instruments, and that this same present constrains future variety” (ibid.). Here the lecture series links back to its theme of freedom. The future “is something we use our freedom to determine, rather than something that is lurking out there, and will happen to us” (ibid.). Beer compresses this into one sentence: “We can make, rather than prophesy, the future” (ibid.). On this view planning is one of the forms freedom takes, provided that the plans are allowed to abort. The two texts, six years apart, state one position. Planning is decision; decision is continuous because information keeps arriving; plans therefore abort, and the aborting is the pay-off. For an enterprise this is the source of cohesion between levels of recursion. For a society it is the difference between a future that is made and one that is merely suffered.
References
- DF — Beer, Stafford (1974). Designing Freedom. London: John Wiley. Cited by lecture from a transcript edition.
- HOE — Beer, Stafford (1979). The Heart of Enterprise. Chichester: John Wiley & Sons.