Digital Transformation Business Case Finance Will Approve

Build a digital transformation business case finance will approve: full-lifecycle cost models, priced risk, phased ROI with funding gates, and stakeholders aligned before the meeting.

Digital Transformation Business Case Finance Will Approve

Most IT programmes overrun their budgets by a margin the organisation can absorb. That sounds survivable. The problem is the tail: a minority of projects overrun so badly that they wipe out the whole business case, and nobody can say in advance which ones they will be.

Your CFO may not know the paper, but they know the pattern. A digital transformation business case that survives scrutiny is one that shows you know it too, and have priced it in.

Get a costed, phased transformation plan your finance team can actually interrogate before you commit budget.

Pressure-test your business case

Why business cases stall in the finance review

Appetite for technology spend is rarely the blocker. A Gartner survey of 301 finance leaders found that 77% planned to increase technology spending, and the same release frames the CFO’s job as driving technology investment discipline. Money exists. Confidence in your numbers is what has to be earned.

Most cases fail on the same four questions. What does this really cost over its life? What happens if it goes wrong? When does value arrive, and who is accountable for it? Why this option rather than a cheaper one?

A useful structure for answering all four comes from an unlikely place: HM Treasury. Its Green Book organises every proposal around five linked perspectives: the strategic, economic, commercial, financial and management cases. Strip away the public-sector language and you have a checklist that maps to your stakeholders. Strategy for the executive sponsor. Economics and finance for the CFO. Commercial for procurement. Management for the PMO and IT. If our board-ready business case guide covers the document, this article covers the modelling underneath it.

Cost modelling: the build quote is the smallest number

The supplier’s build estimate is the line everyone scrutinises, and usually the one least likely to sink you. Overruns tend to come from the costs nobody put in the spreadsheet.

A credible model covers the full lifecycle over three to five years, split into one-off and recurring costs:

  • Discovery and design: research, service blueprinting, information architecture, design system work.
  • Build and integration: front end, CMS or DXP configuration, and every integration with CRM, ERP, PIM, identity and payments. Integrations are where estimates break, because the legacy API documentation is often wrong.
  • Migration: content audit and transfer, data cleansing, URL mapping and 301 redirects to protect organic search equity.
  • Platform and infrastructure: licences for platforms such as Adobe Experience Manager, Sitecore or Contentful, plus hosting, CDN and observability tools. Check which are usage-based. Headless CMS pricing tied to API calls or seats can climb sharply as you scale.
  • Dual running: the months when the old and new platforms both cost money. Almost always underestimated.
  • Internal cost: backfill for the subject-matter experts you will borrow, training, and change management.
  • Decommissioning: retiring the old stack, archiving data and ending contracts, some of which carry notice periods.
  • Run and optimise: the post-launch team that keeps improving the product. Without it, the asset decays and the benefits case decays with it.

Present each line as a range, with a confidence rating and the assumption behind it. A single point estimate signals false precision, and finance teams know it. An assumptions log (“integration with the order management system assumes the existing REST API supports real-time stock”) turns arguments about numbers into arguments about facts, which are much easier to settle.

If the programme involves retiring older systems, the dependency mapping in our legacy modernisation guide is the input your cost model needs.

Price the tail, not the average

Risk sections in most business cases are a table of red, amber and green ratings. That is not risk framing. It is decoration.

The Green Book is blunt on this point: appraisals should be explicitly adjusted for optimism bias, by increasing cost and duration estimates and decreasing benefit estimates. Doing this voluntarily, before finance does it for you, is the single strongest credibility signal a sponsor can send.

Contingency is not padding. It is a priced statement about which risks you have identified, how likely they are, and what they would cost. A finance reviewer trusts a sized, explained contingency far more than a lean budget.

In practice, that means three things. First, quantify the top risks as cost and schedule ranges, not adjectives. Second, show the risk of doing nothing: a CMS approaching end of support, a platform your team can no longer hire for, rising security exposure. The status quo has a cost, and it belongs in the comparison. Third, define kill criteria. Name the conditions under which the programme pauses or stops at the next gate. Paradoxically, a plan that admits it might stop is easier to approve.

Sequencing is also a risk lever. Replatforming and redesigning at once multiplies unknowns; our piece on sequencing a replatform goes deeper on when to split them.

Phased ROI: when does the value actually show up?

Transformation value follows a J-curve. Costs land early, benefits arrive late, and the gap between them is where sponsors lose their nerve and finance loses patience.

McKinsey’s research on transformation implementation is instructive here. The average transformation realizes just over half of its value within the first 18 months, while top-quartile companies move much faster. The same research finds that the implementation phase is where the largest share of value leakage occurs, 35 percent on average. Speed to first value is not a nice-to-have. It is where the return is won or lost.

So structure the projection as phases, each with its own funding gate:

  • Phase 0, discovery: funded separately and small. Its output is a validated cost model and baseline metrics. It produces evidence, not revenue.
  • Phase 1, foundation plus one journey: the platform, the design system and one high-value user journey (checkout, quote, onboarding) in production. This is your first measurable benefit.
  • Phase 2, scale: remaining journeys, markets or brands, reusing the foundation at lower marginal cost.
  • Phase 3, optimise: continuous experimentation and performance work, where compounding gains live.

Separate benefits by type, because finance will discount them differently. Hard savings (retired licences, ended maintenance retainers, reduced hosting) are bankable. Operational gains (time to publish content, release frequency, deployment lead time) are measurable but need converting to cost. Revenue gains (conversion rate, average order value, lead quality) are the largest and the least certain, so show them as conservative, expected and optimistic scenarios.

None of this works without a baseline. Measure current conversion by journey in your analytics platform, current Core Web Vitals from Chrome UX Report field data, and current publishing and release cycle times before a line of code is written. A benefit you cannot compare against a baseline is a claim, not a projection.

Stakeholder alignment happens before the meeting

An approval meeting should confirm a decision, not make one. If the CFO is seeing the model for the first time in the room, you have already lost control of the conversation. These are the steps that do the work beforehand:

Approval is also the start of accountability. The benefit owners and gates you define here should carry straight into your governance model, so the case keeps being tested against reality after funding.

1

Map the real decision-makers:

list who approves, who can veto and who merely gets consulted. Procurement, security and the data protection officer often hold quiet vetoes.

2

Pre-wire finance on the model:

walk a finance business partner through the cost structure and assumptions weeks before submission. Let them change things. Ownership of the model is the goal.

3

Give IT its own case:

show the target architecture, the security review path and the run model. Technology leads approve plans they can operate.

4

Name benefit owners in the business:

each projected benefit needs an accountable owner outside the programme team, who signs off the baseline and the target.

5

Present options, not a verdict:

show do nothing, do minimum, the preferred option and an ambitious option, each costed. Choice builds trust in the recommendation.

6

Run a red-team review:

ask a sceptical senior colleague to attack the case. Every question they raise is one the board will not ambush you with.

Telling a competent supplier from a plausible one

Your delivery partner’s numbers will sit inside your business case, so their estimating discipline becomes your credibility. A competent supplier gives you ranges with stated assumptions, lists exclusions explicitly, and pushes for a paid discovery phase before committing to a fixed build price. A plausible one gives you a confident single figure on the first call.

At Moburst Digital Experience, discovery is the first stage of our process for exactly this reason: it is where cost assumptions get tested before design starts. Whoever you hire, the questions in our guide to evaluating an enterprise agency will help you separate the two.

Your next step: before you write a single slide, fund a short discovery phase whose only deliverables are a ranged cost model, an assumptions log and measured baselines. Everything else in the case gets easier once those three exist.

Frequently asked questions

What should a digital transformation business case include?

It should cover the strategic rationale, a costed comparison of options including doing nothing, a full lifecycle cost model, a quantified risk assessment with contingency, phased benefit projections with named owners and baselines, and a delivery and governance plan. HM Treasury’s Five Case Model is a practical structure for organising these elements.

How far ahead should ROI projections go?

Three to five years is typical, because platform licences, run costs and benefits compound over that period. Show the projection phase by phase so reviewers can see when each benefit is expected and which funding gate it depends on.

How much contingency should a transformation programme carry?

There is no universal percentage. Size contingency from your quantified risk register: estimate the likelihood and cost impact of each major risk, then explain the total. A justified contingency is more persuasive than a round number.

Should discovery be funded separately from the main programme?

Yes, in most cases. A small, separately approved discovery phase produces a validated cost model, tested assumptions and baseline metrics, which makes the main funding request far more defensible and reduces the risk of committing to an unrealistic fixed price.

Who should own the benefits in a transformation business case?

Benefits should be owned by accountable leaders in the business, not by the programme team. Each owner signs off the baseline and the target, and reports against them after launch through the programme’s governance.

Turn your transformation case into a funded programme

You now know what finance looks for: ranged lifecycle costs, priced risk and phased value with named owners. Moburst Digital Experience runs the discovery that produces those numbers, so your case goes to the board backed by evidence rather than estimates.

Talk to our strategy team