Why Digital Transformation Fails: It’s Not the Technology, It’s the Middle Management

Paul Thomas • May 4, 2026

Share this article

Boards approve digital transformations. IT delivers them. Consultancies design them. So why does McKinsey research consistently show that around 70 per cent of large-scale transformation efforts fail to meet their original objectives?


Not because the technology was wrong. The platforms are mature. The integrators are competent. The business cases were defensible at the point of approval.


Digital transformations fail in the middle layer of an organisation. The heads of function, the operational leads, the team managers. The people who are expected to absorb the change while still hitting last year’s targets. That layer rarely gets the support to do both.


The Top Floor Decides. The Shop Floor Adapts. The Middle Carries the Cost.


Look at how most transformations are structured. A board sponsor sets the vision. A programme team designs the future state. IT builds the platform. Training is rolled out to end users.


Now look at where the friction lands. The head of operations is being asked to learn the new system, retrain her team, hit quarterly numbers on the old system while the new one rolls out, and represent the change positively to her people. None of which she was resourced for.


She is not resisting. She is overwhelmed. And when overwhelmed middle management quietly de-prioritises the transformation in favour of the day job, adoption stalls. Without adoption, the benefits case collapses. McKinsey’s own transformation practice names a lack of engagement across the organisation as one of the most consistent reasons transformations underperform.


Three Reasons the Middle Layer Gets Left Behind


1. The transformation is treated as a technology programme

If the project plan reads like a system implementation, with phases for design, build, test, go-live, and hypercare, there is no workstream that owns the human change. The assumption is that training plus communications equals adoption. It doesn’t.


Real adoption requires people to change how they do their work. That requires time, support, and permission to slow down temporarily. None of which appear on a typical implementation plan.


2. The middle layer wasn’t involved in design

Programmes that go straight from board sponsor to technical design skip the layer that knows how the work actually gets done. The resulting solution is theoretically correct and operationally awkward. The people being asked to use it can see the friction immediately, and they were never consulted about it.


Once that gap is visible, trust in the transformation drops. Workarounds appear. Shadow processes emerge. The shiny new platform is technically live and operationally bypassed.


3. There is no slack in the system

Organisations run lean. The same managers who own the day-to-day are the ones being asked to own the change. There is no spare capacity to absorb the cost of transition. So either the day job gives, or the change gives, and the day job almost always wins, because that’s what their compensation is tied to.


If the programme didn’t plan for backfill, secondments, or temporary reinforcement at the operational layer, it didn’t plan for adoption.


How to Land Digital Transformation in the Middle


Treat adoption as a workstream, not an afterthought

Give it a senior owner. Give it a budget. Give it a plan that runs alongside the technical build, not after it. Adoption isn’t something you sprinkle on at go-live. It is built into the programme from week one.


Involve middle management in design, not just delivery

The people who run the operation should help shape the operation’s future. Not in a token workshop. In the design decisions that affect their daily work. The cost of involving them is a few weeks of their time. The cost of not is two years of slow adoption.


Resource the transition honestly

If the change is large enough to need a programme, it is large enough to need temporary capacity at the operational layer. Interim project managers, secondments, backfill: whatever it takes to give middle management room to lead the change rather than survive it.


Measure adoption, not implementation

Going live is a milestone, not an outcome. The real measure of transformation is whether the new way of working has displaced the old one twelve months after go-live. If it hasn’t, the platform may be running, but the transformation hasn’t happened.


The Pattern We See Most Often


Halfway through a transformation, the programme is technically on track. The platform is being built. The board reports are green. But there are quiet signals from the operation. Questions that aren’t getting answered. Training dates being pushed. Change champions going silent. The middle layer is starting to disengage.


That is the moment to intervene. Not at go-live. Not in hypercare. Now. The cost of fixing adoption mid-programme is a fraction of the cost of fixing it after launch, when habits have set and the platform has already missed its first benefits milestone.


If your digital transformation is technically on track but operationally quiet, that quiet is the warning sign. We help mid-market organisations land transformation where it has to land: in the layer that has to live with it.

Free Project Assessment & Audit, Tailored to You

Not sure where your project really stands? We assess what's happening, and give you a clear and honest picture at no cost. Every engagement starts with understanding your specific situation, because no two organisations are the same.

Totally free. No commitment needed.

Recent Posts

Hand circling a date on a desk calendar with a pen beside a laptop
By Paul Thomas July 16, 2026
Planning a PMO setup? This guide walks through the realistic 3-6 month timeline phase by phase, plus a free assessment to find your own number.
By Paul Thomas July 15, 2026
Most consultancies won't answer the cost question directly. Ask, and you'll typically get "it depends" and a form to fill in. It does depend, but that's not an excuse to be vague about what it depends on. If you're a COO trying to build a business case for the board, three variables actually set the price, and none of them are secret. What actually drives the number Programme size is the first. A single at-risk workstream costs less to stabilise than a portfolio of twelve, scope drives everything downstream. A workstream with one team, one sponsor, and one clear deliverable is a contained diagnostic exercise. A twelve-workstream portfolio with overlapping dependencies and multiple sponsors is a different order of problem entirely, because stabilising one workstream in isolation without understanding its dependencies on the other eleven risks solving the visible symptom while the underlying cause resurfaces somewhere else three months later. Embed duration is the second: recovery engagements typically run three to six weeks, and a straightforward stabilisation, credible plan, honest reporting, a functioning team, costs less than a six-week turnaround with practitioners embedded across multiple workstreams. The difference between the two isn't really about calendar time so much as coverage, how many people need to be physically present, in the room, for how many of those weeks, before the organisation can run the plan on its own. Seniority is the third, and the one most likely to be quietly reduced to save money. We only deploy practitioners with 15+ years of experience, never graduates running a template, and that's a deliberate cost floor, it's also why the fix tends to hold. It's worth being honest about why seniority costs more and is still worth it: a junior consultant working from a template can produce a plan that looks credible on paper. A senior practitioner who has seen a dozen versions of this specific failure pattern before knows which parts of that plan won't survive contact with the actual organisation, and can say so in week one rather than week five. What that seniority actually costs, in public numbers You don't have to take a consultancy's word for what senior delivery talent costs in the UK market. The Association for Project Management's 2025 Salary and Market Trends Survey, the UK's largest annual study of the profession, puts the average project management salary at £52,500, up 10% on 2023, with practitioners in consultancy and energy/utilities sectors averaging £62,500, per the APM Salary Survey 2025 . Embedded recovery work commands more than that average because it's compressed: fifteen-plus years of pattern recognition delivered on a timeline measured in weeks, not spread across a year. There's a useful way to think about what's actually being purchased here. A permanent hire's salary buys a year of their time, spread across whatever the organisation needs that year. A recovery engagement's cost buys weeks of concentrated attention on one specific problem, from someone whose entire value is having seen that problem before. The hourly economics look expensive next to a permanent salary until the comparison is corrected for what's actually being delivered in that window. The cost that never makes it onto an invoice Every cost conversation focuses on what recovery costs. The number that's harder to see, and usually larger, is the cost of inaction. A project drifting for another quarter doesn't just slip its date, it compounds in several distinct ways. The team burns out chasing an unrealistic plan, because nobody has told them the plan is unrealistic, so they keep trying to hit dates that were never achievable in the first place. The best people quietly start looking elsewhere, because skilled practitioners can generally tell when a programme has lost its way faster than the sponsors funding it can, and they don't wait around to find out how it ends. Stakeholder trust erodes in a way that outlasts the project itself, a sponsor burned once by an over-promised, under-delivered programme brings that scepticism into the next funding conversation, and the next, long after the specific project is closed. And the eventual fix costs more because the drift has to be unwound before recovery can even begin, every extra week of a bad plan being followed is a week of decisions made on false assumptions that now need to be found and corrected. By the time a board asks how much it will cost to fix, the more useful question is how much it has already cost to wait. Why nobody publishes a rate card A three-week reporting fix and a six-week multi-team turnaround are different projects with different costs, and any headline number would be wrong for most people who read it. That's not evasion. It's the same reason a surgeon doesn't quote a price before the scan — not because the price is a secret, but because quoting it before the diagnosis would mean quoting the wrong number to almost everyone who asked. Where to start Cost climbs with the number of embedded practitioners, the complexity of stakeholder alignment, and how far a programme has already drifted before anyone calls for help. It does not climb because of methodology, tooling, or brand name. Agile Minds gives every prospective client a free, two-week assessment before any number gets discussed. If you need a figure you can actually take to the board, we should talk .
Blue infographic chart with numbered hexagons and a pencil on a desk
By Paul Thomas July 14, 2026
Not every project needs the same delivery model. This step-by-step guide shows you how to build the right one for yours, then offers a free assessment.
Jaguar and Land Rover logos on a gray building facade
By Paul Thomas June 1, 2026
How AgileMinds built JLR a working agile framework through a real product, not a workshop. A 16-week app and guilds that kept it alive after we left.
Halfords storefront with glass entrance, orange logo, and parked bicycles outside
By Paul Thomas June 1, 2026
How AgileMinds delivered Halfords' unified ecommerce platform on Salesforce Commerce Cloud. Two brands, one basket, over 100 integrations, nine months.
Three people in red helmets and jackets stand outdoors at dusk, looking serious.
By Bernadine Dela Cruz May 29, 2026
How AgileMinds helped E.ON take a virtual power station from proof of concept to commercialised, award-winning solution in six months.
Stressed person at desk with laptop, head in hand, reviewing charts in a bright office
May 24, 2026
Uncover the hidden costs of project team burnout, like knowledge loss & attrition. Contact us for strategies to mitigate these issues.
Team collaborating around a laptop with papers, notebooks, coffee, and a smartphone on a desk.
By Paul Thomas May 4, 2026
Most failing projects send warning signs months before the board notices. Here’s how to spot them, and how to recover before it’s too late. Free assessment.
Hands holding a tablet with a to-do checklist and red markings on a desk, lit by blue and warm light
By Paul Thomas May 4, 2026
More than half of PMOs are shut down within two years. Here are the five reasons mid-market PMOs fail, and what to do differently. Free PMO assessment.