It is tempting to treat a stalled digital transformation as a technology problem: the wrong platform, the wrong vendor, not enough automation. The evidence keeps pointing somewhere else. I recently read an interesting paper: ‘Organizational design for digital transformation: A systematic literature review’ by Thome, Hanisch, Oehmichen and de Faria (2027 – Still not formally published), a synthesis of 109 studies published between 2000 and 2025. Here are some key findings that stood out to me as someone who spends most of his working life coaching in matrix organisations, helping people align leadership, structure, and culture in the same direction (See my Matrix leadership series)
Most digital transformations do not work
The paper cites failure rates of between 66 and 84 per cent for transformation initiatives. It also flags something I have not seen named this directly before: the published research has a strong positivity bias. Success stories get written up. Failures mostly do not, with a handful of famous exceptions such as Kodak, Nokia and General Electric that get recycled as cautionary tales because almost nothing else is documented in comparable detail. That matters for anyone using published case studies to build a business case. The evidence base you are drawing on is skewed toward what worked. It says very little about what quietly failed in rooms that looked identical.
Three tensions, not three problems
The authors organise digital transformation into three stages, each built around a tension leaders have to keep managing rather than a problem they get to solve once. Their own diagram of this is worth seeing before I unpack it, since the arrows and loops do work that a bullet list cannot.
Figure 3 from Thome et al. (2027), reproduced under a CC BY 4.0 licence.
Read it round and round rather than left to right. Leadership, governance and culture align on a direction. Strategy, organisation and resources translate that into coordinated execution. Operating model and business model institutionalise what works, then feed back into leadership as conditions change and the loop starts again. Three points stood out most once I sat with each stage in turn.
Align. Digital leadership, digital governance and digital culture establish direction and readiness. The tension here is control versus autonomy: enough governance to keep initiatives coherent, enough autonomy for teams to actually experiment.
Execute. Digital strategy, digital organisation and digital resources translate intent into coordinated action. The tension is stability versus adaptability: structures stable enough to coordinate complex work, flexible enough to keep realigning as conditions shift.
Transform. Digital operating model and digital business model institutionalise new ways of creating and capturing value. The tension is optimisation versus reconfiguration: improving what already works while staying willing to rebuild it.
What I find useful about this framing is that it treats these as polarities rather than dilemmas with a correct answer. That is a familiar idea to anyone who has worked with matrix structures, where the pull between central coordination and local autonomy never resolves, it just gets managed better or worse at any given moment. This paper makes the same case for digital transformation more broadly: it is not a sequence of decisions you get to finish, it is an ongoing balancing act across three linked polarities at once.
A complexity leadership counterpoint
The paper’s own theoretical foundations gesture toward complexity leadership theory, particularly in how it treats digital leadership, but the framework as a whole still speaks the language of design. Get the leadership, governance, culture and structure right, and the tensions become manageable. Complexity leadership theory, the version associated with Uhl-Bien, Marion and McKelvey, argues something a little different: coherent behaviour in a complex adaptive system does not come from specifying the design correctly, it emerges from the interplay between three leadership functions operating at once. Administrative leadership sets structure and allocates resources, which is essentially what this paper calls digital governance. Adaptive leadership is the informal, often unplanned creativity that shows up when people who do not report to each other have to solve a genuinely new problem together. Enabling leadership sits between the two, and its whole job is managing that interface: keeping administrative structure from choking off adaptive activity, and keeping adaptive activity connected enough to the formal organisation that something useful comes back out of it.
Read this way, control versus autonomy stops being a dial that governance design sets once and holds. It becomes ongoing enabling work: noticing where productive friction is happening between hierarchy and self-organising activity, protecting that friction rather than resolving it, and deliberately feeding what emerges back into the formal structure. That is a harder, less satisfying job than “designing the right governance model”, and it is closer to what leadership actually feels like inside a live transformation.
Governance increasingly lives inside the technology, not around it
One example from the paper has stuck with me. Uber’s algorithm allocates rides, evaluates performance and enforces behavioural constraints on drivers, all without a conventional employment relationship. The authors call this digital governance: control exercised through embedded rules and automated coordination rather than through hierarchy, contracts or a manager’s judgement. For anyone working on organisation design, that is a genuine shift, not a cosmetic one. Where authority used to sit visibly with a role or a reporting line, it can now sit inside a system that nobody in the room fully controls day-to-day. Questions of fairness, accountability and trust do not disappear when governance goes digital. They just move to a different, less visible location, and leadership development conversations about accountability need to catch up with that.
Organisation, operating model and business model are three different levers
The paper makes a point that I think gets lost constantly in practice: digital organisation, digital operating model and digital business model are routinely used as though they mean the same thing. They do not. Digital organisation is about structure and decision rights. The operating model is about how work actually gets done, day to day, process by process. The business model is about how value gets created and captured in the first place. A transformation can restructure the organisation chart, redesign every workflow, and still fail to change the business model at all, or vice versa. Naming which of the three is actually broken, rather than treating “digital transformation” as one undifferentiated effort, seems like useful diagnostic discipline before committing a change programme to a direction.
The open question underneath all of it
The paper closes by asking whether its own framework, and organisational design theory generally, survives contact with agentic AI. Emerging “agentic firms” feature minimal hierarchy and AI systems that plan, execute and coordinate work with limited human oversight. The authors are candid that existing organisational design thinking, their own included, assumes human cognition and human social interaction sit at the centre of coordination. That assumption is precisely what agentic AI puts under pressure.
I do not think anyone, including the paper’s authors, has a settled answer to this yet. It strikes me as one of the more important open questions for leadership development to be sitting with right now rather than waiting for it to resolve itself.
Reflection questions for leaders and coaches
A few questions I have been sitting with since reading this, and that seem worth putting to any leadership team midway through a digital transformation:
- Which of the three tensions, control versus autonomy, stability versus adaptability, optimisation versus reconfiguration, is actually live in your organisation right now, and who owns managing it on an ongoing basis rather than deciding it once?
- When you say “digital transformation” internally, are you actually talking about organisation design, operating model change, business model change, or some blend of all three without realising it?
- Where has governance quietly moved into a system or a tool rather than staying with a person, and does anyone still feel accountable for it?
- Where is productive friction already happening between formal structure and self organising activity in your organisation, and is anyone protecting it, or is it being designed away as soon as it appears?
- What would your case for this transformation look like if you assumed, going in, that the base rate of success is closer to one in three than four in five?
Final thought
The technology is rarely the constraint. The constraint is usually whether leadership, governance, culture, structure, capability, strategy, operating model and business model are pointed in the same direction at the same time, and whether anyone is holding the tensions between them rather than pretending they resolve. That is organisation design work, and it is squarely leadership work, long before it is an IT project.
References
Thome, T., Hanisch, M., Oehmichen, J. and de Faria, P. (2027) ‘Organizational design for digital transformation: A systematic literature review’, Journal of Business Research, 218, Article 116508. Available at: https://www.sciencedirect.com/science/article/pii/S0148296326005436 (Accessed: 5 September 2026).
Uhl-Bien, M., Marion, R. and McKelvey, B. (2007) ‘Complexity leadership theory: Shaping the future of leadership from the industrial age to the knowledge era’, The Leadership Quarterly, 18(4), pp. 298 to 318.
Text by myself & header image by AI based on my style (somewhat)!
Leave A Comment