Giddens in IS: Why the Same System Produces Different Organizations

There is a question that IS researchers argue about constantly, usually without naming the argument directly. Does technology change organizations, or do organizations shape how technology gets used? Technological determinists say the first thing. Voluntarists say the second. For most of my time in grad school I noticed that neither answer felt right, because every real case I looked at was obviously both. Technology constrains what people can do, and people use technology in ways its designers never imagined, and both things are true simultaneously. Giddens had the vocabulary for this in 1984, almost a decade before information systems became a mainstream research topic.

Giddens (1984) built structuration theory around what he called the duality of structure. The claim is not complicated to state, but it takes a while to really absorb. Structures, meaning the rules and resources that organize social life, are simultaneously the medium through which people act and the outcome of that action. People draw on existing structures to do things. By doing those things, they reproduce or modify those structures. There is no structure without action that enacts it, and there is no action that is not shaped by structure. The two constitute each other in an ongoing recursive loop. This is why Giddens called it "duality" rather than "dualism." A dualism has two separate things in opposition. The duality of structure has one thing that is two aspects at once.

He specified three modalities that link structure to agency in any moment of interaction. Signification is the dimension of meaning. People draw on interpretive schemes to make sense of what is happening and to communicate, and in doing so they reproduce or transform shared systems of meaning. Domination is the dimension of power. People draw on resources, both authoritative ones like control over people and allocative ones like control over material things, and in doing so they reproduce or transform power relations. Legitimation is the dimension of moral order. People draw on norms and apply sanctions, and in doing so they reproduce or transform what counts as appropriate behavior. Every social interaction operates across all three modalities simultaneously. You are never just exchanging information or just exercising power. You are always doing all three at once.

What Orlikowski did in 1992 was take this framework and use it to rethink what technology is in an organizational context. Her argument was that technology is both a product of human action and a medium of human action. It is a product because people design, build, and configure it through social and organizational processes. It is a medium because once it exists, it shapes what users can do and how they can do it. This is the duality of technology, which mirrors Giddens' duality of structure directly. Orlikowski introduced the concept of technology-in-practice to make this concrete. The structural properties of a technology are not fixed at design time. They are enacted through recurrent use. Two teams using the same system can develop entirely different technologies-in-practice because their patterns of use differ, and those patterns differ because the structural contexts they are working in differ.

The practical implication of this is significant and still underappreciated. Orlikowski's framework predicts that you cannot evaluate a technology without knowing the context in which it will be used, because the technology-in-practice that emerges from use is the actual phenomenon you care about, not the technology as designed. Organizations that assume a system's effects are determined by its features are doing technological determinism without knowing it. And organizations that assume people freely choose how to use technology and that the technology itself is neutral are doing the opposite mistake. The duality says both influence is happening.

The contrast with technological determinism is worth dwelling on. Technological determinism, in its strong form, says that technology shapes society in predictable ways based on its technical properties. The printing press caused the Reformation. The internet caused social fragmentation. In IS, this shows up as assumptions that implementing ERP standardizes processes, that deploying collaboration tools increases knowledge sharing, that adding AI reduces decision time. The technology causes the outcome. Giddens and Orlikowski both reject this. They argue that the same technology can produce different outcomes depending on the structural conditions of its adoption. The printing press was adopted into social structures that were already changing. ERP is adopted into organizations that have existing power distributions, interpretive schemes, and legitimation norms that will shape how the system gets used.

I have been trying to understand ERP implementation failures through this lens and the pattern is striking. The literature on ERP failure is full of cases where a system that worked well in one organization produced disaster in another. The usual explanation is "poor change management" or "lack of user training" or "inadequate customization." The structurational explanation is more precise. The system inscribed assumptions about workflows, approval hierarchies, and data structures that matched one structural context and not another. When the organization tried to use the system, the three modalities played out differently. The signification modality: people did not share the interpretive schemes the system assumed. The domination modality: the power resources the system allocated did not match the existing authority structures. The legitimation modality: the norms embedded in the system's workflows were seen as illegitimate by key groups. The technology-in-practice that emerged was not the one the implementers planned for.

What I find compelling about the structurational approach to technology is that it explains not just failure but the particular texture of adaptation that follows adoption. Organizations rarely use a system the way it was designed, and the adaptations are not random. They follow the lines of existing structural conditions. Finance teams create workarounds that preserve the approval authority they already have. Sales teams configure CRM dashboards to track metrics their managers care about, not the ones the vendor thought were important. IT departments develop shadow scripts that connect systems the platform vendors assumed would not be connected. These adaptations are the structural conditions of the organization reproducing themselves through and around the technology. The technology is the medium, and the existing structure is asserting itself through the use.

Orlikowski also showed that the technology-in-practice that emerges from adaptation can become a new structural condition that shapes subsequent use. The workaround that finance created becomes the way finance does things. New hires learn the workaround as standard operating procedure. After a few years, nobody remembers that it was a workaround. It is just how the system works in this organization. That is the recursive loop Giddens described, now operating through technology. The structure shaped the use, and the use became part of the structure.

The tension I sit with when I apply this framework is the question of where intentional change fits. Giddens' recursive loop can sound like determinism with extra steps if you are not careful. If structure reproduces itself through action, and action is always conditioned by structure, how does anything change? Giddens addressed this by arguing that agency always involves the capacity to act otherwise. Structural conditions constrain but do not eliminate discretion. There are always cracks. Orlikowski's work on technology-in-practice similarly allows for transformative uses that break from existing structural conditions, not just reproductive uses that reinforce them. The question for IS researchers is when and why does use become transformative rather than reproductive, and what structural conditions make that possible.

I do not think we have a fully satisfying answer to that question yet. But the structurational framework is the right place to ask it, because it refuses to assign causality to technology alone or to human agency alone. Both are in play, all the time, and the outcome is always the product of their interaction in a specific structural context.