The case for standardisation
Jay R. Galbraith’s information-processing account of organisation design provides the strongest grounding for the standardisation position. Galbraith’s argument is that rising task uncertainty and interdependence increase the information-processing burden an organisation must carry, and that deliberate design responses — including standardised, self-contained processes that reduce the need for constant real-time coordination — are what allow an organisation to keep functioning as that burden grows. A tightly standardised process, in this view, is not merely a compliance mechanism; it is a way of reducing how much information needs to move between people and sites in the first place, freeing coordination capacity for the genuinely novel problems that still require it. Under this logic, a growing manufacturer that allows too much local variation is quietly increasing its own coordination burden at exactly the point where it can least afford to.
Larry Greiner’s research on organisational growth offers indirect support for the same position from a different angle: his account of organisations moving through a control crisis — the point at which growth outpaces what informal coordination can manage — is frequently resolved, in his own framework, by exactly the kind of formalisation and standardisation the efficiency case argues for. An organisation that has outgrown informal coordination has, in Greiner’s terms, a structural reason to standardise, independent of any argument about information processing specifically.
The case for flexibility
Erik Hollnagel, Robert L. Wears and Jeffrey Braithwaite’s research on the distinction between work-as-imagined and work-as-done provides the strongest grounding for the flexibility position, and it is not a minor counterpoint — it is a direct challenge to the premise that more complete standardisation reliably improves outcomes. Their central finding, drawn from safety research across complex operational environments, is that “things do not go well because people simply follow the procedures and work as imagined. Things go well because people make sensible adjustments according to the demands of the situation.” Under this logic, a standardised process is always, necessarily, incomplete — it describes a typical case, and a growing, diversifying manufacturing operation encounters more atypical cases, not fewer, as it scales into new products, new sites and new conditions the original standard never anticipated. The local adaptation that a standardisation-first view treats as a deviation to be eliminated is, in this account, frequently the mechanism actually keeping the operation functioning.
It is worth being precise about what this position does and does not claim. It does not argue that documentation or standards are worthless, nor that any local deviation is automatically justified. It argues specifically that the growing distance between a written standard and an increasingly varied real operation is not a solvable gap to be closed through ever more detailed documentation — some portion of that gap will always exist, because no document written in advance can specify the correct response to every condition a scaling, diversifying operation will eventually encounter, and treating every deviation as an error to be trained out risks removing precisely the judgement that handles what the document could not.
Why the disagreement persists
The disagreement is genuine rather than semantic because the two bodies of research are describing different failure modes, both real. Galbraith’s concern is coordination collapse — an organisation that cannot process the information volume its own growth has created. Hollnagel, Wears and Braithwaite’s concern is brittleness — a system so tightly specified that it cannot absorb the variation reality inevitably presents, and which breaks, rather than adapts, when conditions depart from what was anticipated. A manufacturer genuinely at risk of coordination collapse and a manufacturer genuinely at risk of brittleness look similar from the outside — both show inconsistent results across sites — but the evidence-backed response to each is close to opposite, which is precisely why leaders cannot resolve this question with a general rule and need to look at their own specific situation.
There is a further reason the two bodies of research are difficult to reconcile into a single position: they were developed to address different kinds of problem in the first place. Galbraith’s framework grew out of studying how organisations structure themselves to handle rising volumes of routine coordination — a question about capacity. Hollnagel, Wears and Braithwaite’s framework grew out of studying safety-critical environments where the cost of a system failing to adapt to an unanticipated situation could be catastrophic — a question about resilience under genuine novelty. A growing manufacturer experiences elements of both problems simultaneously, which is exactly why neither framework, applied alone, offers a complete answer — each was built to see one half of the picture clearly, at the cost of the other. Neither camp would claim its own research settles the question for an organisation facing both pressures at once, which is itself a useful piece of honesty for a leader to carry into the decision rather than expecting the literature to resolve it on their behalf.
What it means for a leader is that the question “should we standardise more or trust local judgement more” does not have a single correct answer supplied by the research — it has two correct answers depending on which failure mode the organisation is actually closer to, and the more useful question may be which of the two risks — coordination collapse or brittleness — currently looks more present in the organisation’s own recent experience.
Where this shows up
People working in manufacturers will probably recognise this pattern first. The wider Asset and Production-Based grouping and the other sectors page show how it presents elsewhere.
Questions worth considering
- Which currently looks more present in the organisation’s recent experience: coordination strain from too little standardisation, or brittleness from too much?
- Where people deviate from the written procedure, are those deviations treated as errors to be trained out, or as judgement handling something the document could not anticipate?
- How would the organisation tell which is which?
Your team is facing similar challenges?
Subscribe to The Riddle Ledger
Perspectives are sent as they are published, with no promotional email, and you can unsubscribe at any time.