Why Cloud Transformations Fail — What Boards Should Watch in Major Technology Investments and Delivery Programmes
By Raviprasad Bangalore | Senior Vice President & Head of Technology, Comviva (Tech Mahindra) | India's Top 10 AI Trailblazers — ET India
The Uncomfortable Truth About Cloud Transformation
Boards approve cloud transformation programmes with confidence. The business case is compelling — cost optimisation, agility, scalability, competitive positioning. The management team is enthusiastic. The technology partner is credible. The timeline looks achievable.
Three years later, the programme has consumed twice the budget, delivered half the promised outcomes, and the organisation is more dependent on its technology vendors than it was before it started.
This pattern is not rare. It is the norm.
Gartner estimates that through 2025, more than 80% of enterprises will overshoot their cloud budgets. McKinsey research shows that while cloud adoption is near-universal among large enterprises, only a fraction are capturing the full value they projected at the outset. The gap between cloud ambition and cloud outcome is one of the most consistent and expensive patterns in enterprise technology investment.
I have spent more than a decade leading cloud transformation programmes — from strategy through execution — across Fortune 500 clients in healthcare, insurance, telecom, and retail. I have managed cloud portfolios exceeding $125M, led organisations of 800+ cloud professionals, and delivered transformations that produced real, measurable outcomes: 50–60% cost reductions, 90% turnaround-time improvements, 15–20% revenue uplift in selected programmes. I have also seen — up close — why cloud transformations fail. Not in theory. In practice.
This article is written for boards. Not to make cloud transformation sound harder than it is — but to give independent directors the specific, grounded oversight lens that most cloud investment proposals do not provide.
Why Cloud Transformations Fail — The Real Reasons
Failure Reason 1: The Business Case Was Built on Benchmarks, Not Reality
Most cloud business cases are constructed using industry benchmarks — average cost savings achieved by comparable organisations, productivity improvements from analyst reports, efficiency gains from vendor case studies. These benchmarks are real. But they represent outcomes achieved by organisations with specific starting conditions, specific execution capabilities, and specific governance disciplines.
When those conditions do not exist in the organisation approving the investment, the benchmark becomes a fiction that the programme is measured against for years — creating a persistent gap between reported progress and actual value delivery.
I have seen programmes where the cost optimisation projections were based on hyperscaler benchmark data that assumed a level of application modernisation the organisation had not yet achieved — and had no realistic plan to achieve within the programme timeline. The savings never materialised. The budget overrun did.
Failure Reason 2: Migration Was Confused With Transformation
This is the single most common and most expensive mistake in cloud programmes. Organisations move their existing applications, data, and processes to the cloud — and call it transformation. It is not. It is relocation.
True cloud transformation means redesigning applications and processes to take advantage of cloud-native capabilities — elasticity, automation, microservices architecture, DevOps operating models. Lift-and-shift migration — moving existing systems to the cloud without redesigning them — typically increases costs rather than reducing them, because cloud infrastructure is priced for dynamic, optimised workloads, not for running legacy applications in a new location.
In my experience leading cloud transformation at scale, the programmes that delivered the most value were the ones that invested in modernisation alongside migration — redesigning applications for cloud-native operation rather than simply relocating them. That takes longer and costs more upfront. But the alternative — migrating without modernising — produces a cloud bill that is higher than the on-premise bill it replaced, with none of the agility benefits that justified the investment.
Failure Reason 3: Vendor Dependency Was Underestimated and Underpriced
Cloud transformation programmes almost always involve significant dependency on technology vendors — hyperscalers, system integrators, platform providers, and specialist tooling vendors. That dependency is rarely fully priced into the business case, and its governance implications are almost never discussed at board level.
The dependency problem manifests in three ways. First, pricing: hyperscaler pricing models are complex, and egress costs, data transfer fees, and reserved instance commitments can produce cloud bills that bear little resemblance to the projections in the original business case. Second, lock-in: organisations that build deeply on a single hyperscaler's proprietary services — databases, AI services, serverless platforms — find that switching costs are prohibitive, giving the vendor significant pricing power at renewal. Third, capability dependency: organisations that outsource cloud engineering entirely to a system integrator often find that they have not built internal capability — and when the engagement ends, they cannot operate or evolve what was built.
Failure Reason 4: Governance Was Treated as a Programme Management Function
Cloud transformation governance is not the same as programme management. Programme management tracks milestones, manages budgets, and reports status. Governance asks whether the programme is still solving the right problem, whether the assumptions in the business case are still valid, and whether the organisation is building the capability it needs to sustain the outcomes after the programme ends.
Most cloud transformation programmes have strong programme management. Very few have strong governance. The result is that programmes stay on track — against a plan that stopped being relevant 12 months ago — while the strategic value they were designed to deliver quietly disappears.
Failure Reason 5: The Organisation Was Not Ready
Technology transformation programmes fail at the human layer more often than the technology layer. Cloud transformation requires new skills, new ways of working, new operating models, and new accountability structures. Organisations that invest heavily in technology and lightly in organisational change consistently underperform against their business cases.
The specific capability gaps that cause the most damage are cloud financial management (FinOps), cloud security operations, and DevOps engineering. These are not skills that can be hired quickly or developed through a training programme. They take time to build — and programmes that do not start building them early enough find themselves with cloud infrastructure they cannot optimise, secure, or evolve.
What Boards Should Watch — Seven Specific Oversight Areas
1. The Business Case Assumptions
Every cloud transformation business case rests on a set of assumptions — about the current state of the organisation's technology landscape, about the pace of migration and modernisation, about the cost of cloud infrastructure at scale, and about the organisation's ability to execute. Those assumptions are almost never made explicit in the board presentation.
The board's first oversight responsibility is to require that the assumptions are surfaced and stress-tested before approval. Specifically: what is the assumed migration pace, and what evidence supports it? What is the assumed cost per workload in the cloud, and how was that calculated? What is the assumed level of application modernisation, and what is the plan to achieve it?
Assumptions that cannot be clearly articulated are assumptions that have not been properly examined. That is a governance red flag — not a reason to reject the investment, but a reason to require more rigour before approval.
2. The Lift-and-Shift vs. Modernise Split
Ask management directly: of the applications and workloads being migrated, what percentage will be lifted and shifted without modification, and what percentage will be modernised for cloud-native operation? What is the plan and timeline for modernising the lift-and-shift workloads after migration?
This question matters because lift-and-shift workloads are the primary driver of cloud cost overruns. An organisation that migrates 80% of its workloads without modernisation has not transformed — it has relocated. And the cloud bill will reflect that.
In my experience, the most successful cloud programmes maintained a disciplined modernisation agenda alongside migration — not as a future phase, but as a concurrent workstream. Boards should ask whether that discipline is built into the programme plan.
3. Cloud Cost Governance
Cloud costs are fundamentally different from traditional IT costs. Traditional IT infrastructure is largely fixed — you buy servers, you depreciate them, you know what they cost. Cloud infrastructure is variable — costs scale with usage, and without active management, they scale faster than the business value they generate.
Cloud cost governance — often called FinOps — is a discipline that most organisations underinvest in, particularly in the early stages of cloud transformation. The result is cloud bills that grow faster than projected, with limited visibility into which workloads are driving the cost and whether those costs are justified by the value delivered.
Boards should ask: what is our cloud cost governance framework? Who is accountable for cloud cost optimisation? How are we measuring cost per unit of business value — not just total cloud spend? Is there a FinOps capability in place, or is cloud cost management being handled reactively?
4. Vendor Concentration and Exit Strategy
Ask management to map the organisation's vendor dependencies across the cloud programme — hyperscalers, system integrators, platform providers, and specialist tooling vendors. For each significant dependency, ask: what is the switching cost if we need to change vendors? What is our exit strategy? What capability are we building internally to reduce dependency over time?
This is not an argument against using vendors — cloud transformation at scale requires strong vendor partnerships. It is an argument for understanding and governing the dependency, rather than discovering it at contract renewal when the leverage has shifted entirely to the vendor.
I have seen organisations that built deeply on a single hyperscaler's proprietary AI and data services — without a clear exit strategy — find themselves facing 40–50% price increases at renewal with no realistic alternative. That is a governance failure that started at the programme design stage and was never surfaced at board level.
5. Internal Capability Building
Ask specifically: what internal capability is this programme building, and how will we measure it? At the end of the programme, will we be more capable of operating and evolving our cloud environment independently — or more dependent on external vendors than we are today?
This question is particularly important for organisations that are outsourcing significant portions of their cloud engineering to system integrators. Outsourcing execution is a legitimate programme strategy. Outsourcing capability building is a long-term governance risk.
The programmes I have seen deliver the most sustained value were the ones that treated capability building as a first-class programme objective — not a secondary benefit. They established internal cloud centres of excellence, built FinOps capabilities, developed cloud security operations teams, and created engineering communities that could sustain and evolve what the programme built.
6. Security and Compliance in the Cloud
Cloud security is not the same as on-premise security. The shared responsibility model — where the hyperscaler is responsible for the security of the cloud infrastructure and the organisation is responsible for the security of what it puts in the cloud — is widely misunderstood, and that misunderstanding creates real exposure.
Boards should ask: has our security and compliance team been involved in the cloud programme design from the outset — not just as a reviewer at the end? Do we have a cloud security architecture that addresses identity and access management, data encryption, network security, and compliance monitoring? Have we mapped our cloud deployments against our regulatory obligations — data residency, data sovereignty, sector-specific compliance requirements?
In regulated industries — BFSI, healthcare, telecom — these questions are not optional. Regulators are increasingly scrutinising cloud deployments, and organisations that have not built compliance into their cloud architecture from the start face significant remediation costs and regulatory risk.
7. The Value Realisation Plan
This is the question that is most consistently absent from cloud transformation governance. The investment case projects outcomes. The programme plan tracks milestones. But who is responsible for ensuring that the projected outcomes are actually realised — and what happens if they are not?
Value realisation requires a specific plan: which outcomes will be measured, how they will be measured, who is accountable for delivering them, and what the decision criteria are for scaling, adjusting, or stopping the programme if outcomes are not materialising.
Without that plan, cloud transformation programmes have a tendency to declare success based on technical milestones — migration complete, workloads in the cloud, infrastructure decommissioned — while the business outcomes that justified the investment remain unmeasured and undelivered.
The Board's Role — Independent Oversight, Not Technical Review
I want to be clear about what I am not suggesting. Boards do not need to become cloud architects. They do not need to evaluate technology choices or review infrastructure designs. That is management's job.
What boards do need is an independent oversight capability — the ability to ask the right questions, interpret the answers, and recognise when the answers are incomplete, inconsistent, or optimistic beyond what the evidence supports.
The seven oversight areas above are not technical questions. They are governance questions. They can be asked by any director who has taken the time to develop an independent mental model of how cloud transformation programmes work — and how they fail.
The directors who provide the most value in technology investment oversight are not the ones who know the most about technology. They are the ones who ask the most precise questions about assumptions, accountability, capability, and outcomes — and who are willing to require better answers before approving the next phase of investment.
A Pattern Worth Recognising
In nearly three decades of building and leading technology transformation programmes, I have observed a consistent pattern in the ones that fail.
They start with a compelling business case. They are approved quickly because the opportunity is real and the competitive pressure is genuine. They are governed as programme management exercises rather than strategic investments. The assumptions in the business case are never revisited. The vendor dependencies are never mapped. The internal capability is never built. The value realisation plan is never written.
And when the programme ends — or when the budget runs out — the organisation has cloud infrastructure it cannot fully optimise, vendor dependencies it cannot easily exit, and a cloud bill that is higher than the on-premise costs it replaced.
The board that approved the investment never had the information it needed to prevent that outcome. Not because the information did not exist — but because nobody asked for it.
These seven questions are how you ask for it.
Raviprasad Bangalore is Senior Vice President & Head of Technology at Comviva (Tech Mahindra), recognised among India's Top 10 AI Trailblazers by ET India. He has led cloud transformation programmes across Fortune 500 enterprises in healthcare, insurance, and telecom, managing cloud portfolios exceeding $125M and organisations of 800+ cloud professionals. He advises boards and enterprises on technology governance, cloud transformation strategy, and AI oversight.