Skip to content

Re-considering the shadow system argument

📅 26 July 2025 in: shadowsystems, casa

Origins#

The following was sparked by this blog post from Col Beer titled: "Is it time to reignite the argument around shadow systems?"

My answer is yes, but unlikely to happen.

In part because re-consideration should probably be given to more than just IT. IT practices are reflective of broader institutional corporate practices. A broader change is very difficult. Even just focusing on just IT would require accepting a different mindset (e.g. Col's interesting (more below) mention Complexity Leadership (Uhl-Bien et al, 2007)). Adopting a different mindset is difficulty and unlikely.

Thoughts on Col's post#

The first paragraph cites personal experience with learning and teaching within Australian tertiary education highlighting lots of examples of shadow systems in L&T. Most/all of which was killed off by the IT department, but a far percentage never became visible to the IT department (or anyone else beyond the immediate user(s)).

Col and I have some shared experience, but also varied experience at different institutions. In a previous post I analysed ~50 shadow systems I'd used during my time in higher ed against some online course quality indicators.

Are there different/similar experiences elsewhere?

  • How widespread is this common experience with L&T shadow systems?

  • Are there examples of shadow systems that have been accepted by IT departments?

    • Perhaps distinguishing between those accepted by IT departments and those that have become stand alone products adopted at other institutions. e.g. SRES

  • Does Australia's more corporatised approach to management and governance of higher education make a difference? Is the experience different in other countries?

  • How is this experience of shadow systems changing with the changing nature of digital technology? For the longest time the rise of shadow systems have been enabled the increasingly easy development affordances offered by contemporary digital technologies. An early example being the rise of spreadsheets and after that a sequence of technologies: Visual Basic/Access; Robotic Process Automation (RPA); and, more recently the rise of GenAI and vibe coding.

The second paragraph briefly summarises the perceived risks and benefits of shadow systems. But suggests a need for more than just prohibition.

What else might there be than prohibition?

  • What are the potentially possibilities? Prohibition, selective support, active enablement...?

Later a broad difference between rules-based and values-based approaches is mentioned.

Dimensions of shadow system permissions

I'm wondering about the specific possibilities either of these approaches might create. Rather than blanket and total prohibition. Dimensions might include

  • Permission - who says who can engage in Shadow Systems. Ranging from open slather, through local unit approval, to some sort of centralised permissions approach. All using different criteria.
  • Support - what support is provided by the organisation for shadow systems, possibilities might include: none, support of "innovation" trials, through to active design of organisational digital technology ecosystems that enable end-user-programming
  • Scope - what is allowed to be developed as a shadow system. A system designed for personal use, use within an individual organisational unit, across multiple units, the entire organisation, or for sharing beyond the organisation?

Next organisational management practice which separates technology consideration from educational concerns is compared to what is known about the best outcomes for teaching. Which is argued as being the consideration of the whole. Consider the symbiosis, not the parts separately.

Can this be done (effectively) in a teleological design approach? How contextual?

There are two broad approaches to technological design: ateleogical-versus-teleological. All organisational approaches use teleological design. An approach that starts with the goal, gathers requirements, and selects between choices before implementation. Such an approach tends to strip away messy complications about context in such decisions. The specifics of pedagogical approach to which a technology is applied or the specific needs of a single third year course in electrical engineering are some of the context that is ignored/lost within such an approach.

Symbiosis requires consideration of such context. But where to draw the line? Do you have to consider the specific context of each course? If not, do you consider a broader context? Where do you draw the line?

Col also weaves in the challenge of shadow governance as part of the problem. e.g. committee stacking. A method by which people with specific perspectives ensure their perspective carries the day. For Col, it is practice like this that create an imbalance between considerations of education and technology.

Leading to the recommendation that it is time for organisations to reflect on the balance of perspectives in decision making. Also a suggestion to reflect on shadow governance mechanisms. Most interestingly, the what controls "need to be hard (rules-based compliance) and what controls can be soft (principles-based and discretionary)" Complexity leadership (Uhl-Bien et al, 2007) is offered as a framework that may help.

The value of complexity leadership

Perhaps the key is identifying three types of leadership required by complex adaptive systems (CAS) (e.g. any largish contemporary organisation) (Uhl-Bien et al, 2007, p. 305)

  1. Adaptive leadership - learning activity arising from the interactions within a CAS in response to tensions.
  2. Administrative leadership - formal managerial tasks planning/coordinating activities to achieve organisational goals.
  3. Enabling leadership - catalysing conditions for adaptive leadership to thrive and manage the entanglement between bureaucratic (administrative) and emergent (adaptive) organisational functions.

From this perspective, shadow systems emerge because contemporary organisations only engage in administrative leadership. Enabling leadership is missing, hence adaptive leadership occurs in the shadows (e.g. shadow systems).

Col closes with the question of how does IT governance - especially a rules-based approach - resonates with questions of academic freedom.

Is teaching an academic or institutional activity/responsibility? Is the problem more than just technology governance?

Fairly common within Australian higher education L&T circles for at least 20 years has been ideas like the university owns all teaching material and this is the institution's course not some specific academic's course. Over the same timeframe, the core of quality assurance at Australian universities has been outcomes driven. Learning outcomes are established for courses and programs and quality is judged by how effectively they are achieved.

While Col's definition of academic freedom includes "teach...without unreasonable interference" one assumes that the requirement to teach to the accepted learning outcomes is not unreasonable interference. The teaching has to occur within some organisational constraints. Similarly, the institution's chosen LMS and other technologies are part of the organisational constraints.

Perhaps the shadow system question isn't just a question of technology governance? Perhaps corporate governance? Perhaps the issue is the assumptions underpinning a requirement for a corporate governance approach?

Is it the absence of enabling leadership and proper support for adaptive leadership?

References#

end-user-programming

ateleogical-versus-teleological

Uhl-Bien, M., Marion, R., & McKelvey, B. (2007). Complexity leadership theory: Shifting leadership from the industrial age to the knowledge era. The Leadership Quarterly, 18(4), 298–318. https://doi.org/10.1016/j.leaqua.2007.04.002