Funding Experiments, Not Production Lines
A digital portfolio should purchase evidence before it purchases certainty.
The Capital Allocation Problem Disguised as a Technology Problem
Digital investment is still being governed as though it were a conventional capital programme: define the destination, specify the solution, calculate the return, approve the budget and hold delivery to the original promise. That sequence is comforting because it converts uncertainty into a document. It is also increasingly detached from the nature of the work.
A factory, branch network or core-system replacement can often be justified through a relatively stable relationship between scope, cost and benefit. A digital proposition begins somewhere else. Customer behaviour is only partly understood. Competitors can alter expectations in months. The value of data emerges through use. A service that appears marginal in isolation may become important when connected to another service, while an apparently compelling idea may fail as soon as real users encounter it.
The central portfolio challenge is therefore not simply choosing the right projects. It is creating a funding system capable of learning which opportunities deserve to become investments.
Most organisations claim to support experimentation while retaining governance designed to eliminate uncertainty before funding is released. The result is predictable. Teams learn to present discovery as certainty. Sponsors inflate benefits to clear approval thresholds. Review boards compare initiatives using numbers that look consistent but rest on assumptions of radically different quality.
When uncertainty is punished at the approval gate, it does not disappear. It is merely hidden inside the business case.
This matters because digital disruption is not producing a shortage of ideas. It is producing a shortage of credible mechanisms for moving from idea to evidence, and from evidence to scaled commitment.
Why the Annual Business Case Distorts the Digital Portfolio
The traditional business case performs several useful functions. It forces a sponsor to articulate an objective, identify costs, consider alternatives and accept accountability. The problem begins when it is treated as a prediction contract rather than a decision instrument.
For a digital opportunity, the weakest evidence is often available at the moment when the most precise forecast is demanded. Teams are asked for five-year benefits before they have observed five weeks of customer behaviour. They are required to select a delivery solution before testing the problem. They must attach confidence to adoption assumptions that have never met a user.
This creates three distortions.
First, it rewards initiatives that can tell the most persuasive story, not those that can learn fastest. A large programme with a polished financial model can appear more investable than a modest experiment designed to test the model’s most fragile assumption.
Second, it makes stopping politically expensive. Once an initiative has secured a substantial budget, acquired a programme identity and announced its benefits, evidence that challenges the original case is treated as a delivery problem. The organisation has invested not only money but reputation. Governance then shifts from asking whether the proposition remains valuable to asking how the approved plan can be defended.
Third, it fragments the portfolio into self-contained projects. Yet digital value frequently depends on shared capabilities: common identity, reusable data, consistent service design, integration standards and platforms that allow several propositions to move faster. Project-by-project approval tends to duplicate these capabilities because each case must stand alone.
The apparent discipline of the annual cycle can therefore produce an undisciplined portfolio: too many commitments, too little learning and infrastructure funded only when a sufficiently large project can carry it.
From Project Selection to Investment Progression
A more useful model treats funding as a sequence of commitments. The organisation does not attempt to decide everything at once. It decides what must be learned next, how much that learning is worth and what evidence would justify a larger commitment.
This is not an argument for weak governance or an unrestricted innovation fund. It is an argument for governance that becomes more demanding as evidence accumulates.
At the earliest stage, a team should be funded to establish whether a material customer or operational problem exists. The required evidence may include direct observation, service data, process analysis and a clear account of who experiences the problem and why it matters. The output is not a detailed solution plan. It is a better-founded decision about whether the opportunity deserves further attention.
A second commitment should test the critical assumptions behind a proposed response. Can the relevant data be obtained? Will customers change their behaviour? Can the service operate within legal, security and operational constraints? Is the organisation capable of supporting it? At this stage, a prototype that invalidates an assumption may be more valuable than one that attracts praise.
Only after these uncertainties have been reduced should the portfolio make a significant scaling commitment.
- Explore: establish the problem, strategic relevance and evidence gap.
- Test: examine the assumptions most capable of destroying value.
- Prove: demonstrate that the proposition works in a limited but real setting.
- Scale: commit substantial funding when adoption, economics and operational viability are credible.
These stages should not become another rigid lifecycle with new labels attached to old approval habits. Their purpose is to match the size of the commitment to the strength of the evidence.
“A digital portfolio should purchase evidence before it purchases certainty.”
That principle changes the conversation at investment committees. Instead of asking whether a speculative return is correct, decision makers ask what is known, what remains uncertain, what has been learned since the previous commitment and which assumption the next tranche of funding will test.
Governing Options Rather Than Defending Predictions
Incremental funding is sometimes criticised as indecisive. In reality, it can produce sharper decisions because it exposes the cost of waiting, learning and stopping.
Every early-stage digital initiative is an option. A relatively small investment buys the right, but not the obligation, to make a larger investment later. The value lies partly in the opportunity itself and partly in the information obtained. This makes failure more intelligible. A controlled experiment that disproves a major assumption has not necessarily wasted money; it may have prevented a far larger misallocation.
The portfolio must nevertheless distinguish useful learning from perpetual experimentation. Teams should not receive repeated tranches simply because they remain busy. Each commitment needs explicit evidence thresholds and an expiry point.
A credible investment review might ask:
- Has the problem become more or less important?
- What direct evidence supports that judgement?
- Has customer or competitor behaviour changed?
- Which critical assumption has been tested?
- What did the team expect?
- What actually happened?
- What is the next irreversible commitment?
- Can it be delayed until stronger evidence exists?
- What opportunity would be displaced by funding it?
- What would cause the organisation to stop?
- Is that condition observable?
- Does the sponsor genuinely accept it?
The final question is often the most revealing. Many organisations define success criteria but avoid defining termination criteria. Without them, staged funding becomes staged approval: every review releases the next tranche because stopping still feels like admitting failure.
The Portfolio Is More Than a Collection of Experiments
Digital investment cannot be governed entirely at initiative level. A portfolio must also make deliberate choices about common capabilities and strategic concentration.
Some investments will never produce an attractive standalone return because their value is distributed across multiple services. Data quality, identity management, integration capability, security engineering and reusable service components are examples. If they must compete as ordinary projects, they will be chronically underfunded or repeatedly rebuilt.
The portfolio therefore needs at least three distinct investment conversations:
- Propositions: customer or operational opportunities tested through evidence.
- Platforms and shared capabilities: foundations whose value appears across several propositions.
- Core renewal: changes required to reduce fragility, cost or constraints in existing technology.
Combining all three in a single ranking creates false comparisons. A customer proposition can describe visible benefits; a shared capability often enables benefits elsewhere; core renewal may chiefly reduce exposure to future failure. They require a common strategic frame but different evidence.
The portfolio should also resist the temptation to distribute digital funding evenly across every business unit. Strategic investment is not an entitlement programme. If an organisation believes digital disruption is changing the basis of competition, it must be willing to concentrate resources where learning and advantage can compound.
That may mean funding several related experiments around a small number of customer journeys, data assets or market opportunities rather than approving isolated initiatives across the enterprise. Concentration enables teams to reuse capabilities, compare evidence and build momentum. Dispersion produces activity without strategic weight.
A Different Contract Between Finance, Technology and the Business
None of this can be delegated to a digital team. It requires a different contract between finance, technology and business leadership.
Finance must protect the organisation from uncontrolled commitment without demanding fictional precision. Technology leaders must expose architectural dependencies and long-term constraints without turning every proposal into a platform programme. Business sponsors must remain accountable for outcomes while accepting that the route to those outcomes will change as evidence develops.
The portfolio function sits at the centre of this contract. Its role is not to standardise every initiative into comparable paperwork. Its role is to make uncertainty visible, establish the conditions for progressive commitment and ensure that learning in one initiative changes decisions elsewhere.
This also requires a more useful rhythm than the annual planning round. Strategy may still set annual direction, but investment decisions should occur often enough to respond to evidence. A quarterly portfolio review, supported by shorter learning checkpoints, can redirect resources without turning every change into an exception to the plan.
The measures must change accordingly. Delivery against scope remains relevant for established commitments, but early-stage propositions need measures of evidence gained, assumptions retired, customer response and time to a credible decision. Portfolio health should include the rate at which weak ideas are stopped, not merely the percentage of projects reported green.
Beyond the Comfort of Approval
Digital disruption is often described as a technology challenge. For established organisations, it is equally a challenge to the machinery of investment. New tools and channels matter, but they will not create strategic renewal if capital remains locked into long promises made with weak evidence.
The answer is not to abandon business cases. It is to make them living records of an investment thesis: what the organisation believes, what evidence supports that belief, what would disprove it and why the next commitment is proportionate.
This approach will feel less certain at the beginning because it refuses to disguise assumptions as facts. It should create greater confidence over time because commitment grows with evidence rather than with organisational momentum.
The uncomfortable question for every portfolio board is simple: how many current digital investments would still receive their next tranche if approval depended on what has been learned rather than what was promised?