Organization design

In plain terms

Every organization has a design, whether or not anyone chose it. Someone decided how work is grouped, who reports to whom, who can decide what, how information moves, what gets rewarded and who gets hired. Organization design is the attempt to make those choices on purpose, and to make them fit together.

The most visible output is the organization chart. Much of what matters is not on it: decision rights, the flow of information between units, and what people are rewarded for. A new chart that leaves those untouched may change less than it seems to.

Where it came from

Structure follows strategy. In Strategy and Structure (1962), Alfred Chandler examined four large U.S. industrial firms from the 1900s to the 1940s and the executives who devised the decentralized, multidivisional structure. The Harvard Gazette credits the book with helping to spawn the field of corporate strategy and with making "strategy precedes structure" a staple of corporate management (Chandler, 1962; Harvard Gazette, 2007).

No single best structure. Tom Burns and G. M. Stalker's The Management of Innovation (1961) contrasted mechanistic and organic forms of management. A later paper summarizes their argument: in dynamic economic sectors, firms with organic structures are more effective than those with more mechanistic ones (Burns & Stalker, 1961; Sine et al., 2006). Paul Lawrence and Jay Lorsch published work on differentiation and integration in complex organizations in 1967, as an article and as the book Organization and Environment (Lawrence & Lorsch, 1967a, 1967b).

Design as a set of levers. Jay Galbraith published Designing Complex Organizations in 1973. His firm says he developed the Star Model in the 1960s as a framework for analyzing organizations (Galbraith, 1973; Galbraith Management Consultants, n.d.).

The contingency argument has also been questioned. Sine, Mitsuhashi and Kirsch (2006) argued that Burns and Stalker's proposition does not hold for new ventures in turbulent, emerging sectors, and hypothesized that more formal structure helps such ventures.

How it works

Galbraith's star model. His firm describes five categories of design policy (Galbraith Management Consultants, n.d.):

1. Strategy, which sets direction.
2. Structure, which determines where decision-making power sits.
3. Processes, which govern the flow of information.
4. Rewards, which provide motivation and incentives for the behavior wanted.
5. People, the selection and development of people in line with the other policies.

The firm's summary of the model's main implication is that organization design is more than structure. The McKinsey 7-S and the Burke-Litwin model make a similar point with different lists.

One practical order of work, offered as a working sequence rather than a researched method, is to be clear on the strategy and the work, decide how to group the work, decide decision rights and the processes that connect units, align rewards and roles, and then plan the move from the old design to the new one. That last step is a reorganization, which the research linked there shows can carry people costs.

What the research says

The Evidence Library holds no trial comparing one organizational design with another. It does hold several studies on parts of design: layers, participation and team structure.

Layers can slow ideas going up. In a publicly listed European consumer-goods firm of about 50,000 employees, each additional manager at the second-highest level was associated with roughly 10 percent fewer employee ideas passed upward. In a follow-up lab experiment with 60 students, a sense of lacking control, far more than fear of judgment, explained the holding back (Reitzig & Maciejovsky, 2015, an Atlas study rated medium after a full-text read).

Felt participation is more strongly linked to outcomes than formal structure. A meta-analysis of 60 field studies of democratic enterprises found that employees' own sense of taking part in organizational decisions was more strongly related to job satisfaction (ρ = .25), work motivation (ρ = .36) and a supportive climate (ρ = .44) than formal participation structures or employee ownership were (Weber et al., 2020, rated Medium from its abstract). The studies were cross-sectional.

Team design. A meta-analysis of team design variables found that task design variables, such as task significance, had consistent positive relationships with team performance, and that some structure variables, such as degree of self-management, were also related to performance (Bell, 2004, rated Medium from its abstract). Field links were mostly correlational.

Who starts a change. In 1,795 responses from 468 Dutch organizations early in a change, employees backed change most when middle managers started it and top managers carried it out, yet that pattern was the rarest, about 3 percent of cases (Heyden et al., 2017, an Atlas study rated medium after a full-text read). See top-down vs. bottom-up change.

Where the evidence is thin

1. Design against design. Organizations do not adopt designs at random, so there are no trials of, say, a functional against a divisional structure. The studies here look at features of design, one at a time.
2. Frameworks. The star model, 7-S and Burke-Litwin are frameworks for thinking. None has a graded test showing that using it produces a better design.
3. The move itself. In a systematic review of 39 longitudinal studies (de Jong et al., 2016, abstract read), employee well-being mostly declined after restructuring. A better design could still cost more than it gains if the move goes badly, but no graded study here tests that. See reorganization.

How it maps to the Unfrozen Loop

1. Week 0
   Before you start
2. 1
   Surface both ways

   In the firm studied below, each extra manager at one level of the hierarchy was associated with fewer employee ideas passed upward. Design is one place to look for how far bad news and ideas have to travel.

3. 2
   Make sense together, top included
4. 3
   Start it close to the work

   In the democracy meta-analysis below, how much employees felt they took part in decisions was more strongly related to their satisfaction and motivation than the formal participation structure was. A design that starts close to the work gives people a real say, not only a seat.

5. 4
   Review on a rhythm, then hold the pace
6. 5
   The middle’s own weather
7. Ring
   Safe to speak, and speaking changes something

Highlighted parts are the ones this idea speaks to. The mapping is my reading, not a tested finding. See the whole Loop and the evidence for each part.

Common misreadings

"Organization design means drawing the org chart." Galbraith's framework treats structure as one of five policy areas (Galbraith Management Consultants, n.d.). Rewards, processes and people are part of the design too.

"There is a best structure." Burns and Stalker argued that structure should suit the setting, with organic structures more effective in dynamic sectors (Sine et al., 2006). Sine and colleagues challenged even that for new ventures.

"Flat is always better." One firm's data link more layers with fewer ideas passed up (Reitzig & Maciejovsky, 2015). That is a reason to look at how ideas travel in your design, not proof that removing layers helps everywhere.

Using it well

1. Start from the work, not the boxes. Write down the three or four things the organization must do well for its strategy to succeed. Check each design choice against them.
2. Map decisions, not just reporting lines. For the ten most common decisions in a unit, note who decides, who must be consulted and how long it takes.
3. Count the layers an idea must climb. Pick a recent frontline suggestion and trace how many people it passed through before someone could act on it.
4. Plan the transition as carefully as the design. See reorganization for what the research says about the period after the announcement.