Lumen Software: Tailoring MoP to Continuous Delivery

MoP distinguishes principles from practices and from documentation. The five principles are universal and should be kept without modification: senior management commitment, governance alignment, strategy alignment, portfolio office, energised change culture. Lumen is weak on the last of these, and the compliance tightening is making it worse.

The practices should be kept in substance but scaled in form. Prioritisation, benefits management, resource and capacity management, risk management and stakeholder engagement all prevent failure modes that agile delivery does not remove. Two teams delivering work that had been de-prioritised twice is precisely a portfolio failure: the absence of an authoritative, current prioritised view meant teams optimised locally and the organisation paid twice.

What should be scaled down or redesigned: the annual fixed plan becomes rolling quarterly prioritisation with a rolling-wave plan; 45 mandatory detailed business cases become tiered cases, with a one-page case for the £15k experiments and full cases reserved for significant or irreversible bets; the 12-page template becomes a dashboard plus a decision paper; universal stage gates become gates proportionate to spend and risk -- a £15k experiment and a regulated data migration do not merit the same control.

'Drop the process' fails because the problems Lumen wants to escape -- plan obsolescence, slow approvals -- would be replaced by problems it cannot currently see: duplicated spend across product lines, capacity over-commitment, and no link between investment and strategy. The one genuinely valuable artefact, the spend cap the CFO relies on, also disappears.

Separate the envelope from the allocation. The CFO's control is an annual funding envelope set by strategic theme, plus a held reserve of roughly 15% for reallocation. That caps spend and gives finance a single number to hold. Allocation within the envelope is a quarterly portfolio board decision, chaired by the CEO, with delegated authority to theme owners. Teams then reprioritise within a theme without a new approval, because the money and the decision rights are already in place.

The supporting mechanics: fund capacity and themes rather than named initiatives, so pivoting does not require a business case each time; write down explicit thresholds (for example, items below £50k with no regulatory, security or data implications are decided by the theme owner and reported at the next board; anything above, or irreversible, comes to the board); replace universal stage gates with a single significant-investment gate; and substitute the monthly variance reports with one live dashboard. Benefits are tracked at theme level, not per initiative, because at Lumen's cadence the initiative is a transient container.

Two features are non-negotiable or the model fails. First, an explicit stop mechanism each quarter -- without one, the portfolio only accumulates and the reserve is consumed by drift. Second, decision rights written down and visible, so that a team leader knows what they can decide alone. This is MoP tailoring done properly: fewer, clearer, better-informed decisions rather than fewer controls, or more.

The principle says a portfolio should actively foster a culture that supports change: visible leadership, clear accountabilities, empowered teams, recognition of good work, honest and blame-free discussion of failure, and an appropriate balance between control and autonomy. Where control is experienced as arbitrary or disproportionate, it produces compliance and disengagement rather than assurance. The portfolio office's response -- monthly variance reports, gates for every initiative, a 12-page template -- increases the control burden and signals distrust, which is the opposite of what the principle asks for. It is also self-defeating: a 12-page template for a £15k experiment will be completed badly or late, and the information quality falls.

Testing it rather than assuming it matters, because attrition has many causes. Useful evidence: exit interviews coded for whether portfolio processes were cited; a short pulse survey distinguishing 'portfolio processes help me decide' from 'portfolio processes slow me down'; the proportion of team time spent producing portfolio information versus delivering; whether the same data is requested more than once in different formats; and comparison of attrition between teams whose work is heavily inside the portfolio and teams largely outside it. If the pattern holds across these, the portfolio is a contributor.

The remedy is process reduction plus a role change for the portfolio office, from compliance police to decision support -- fewer artefacts, one source of data, thresholds that let small decisions be taken locally, and a public stop list that demonstrates the portfolio says no as well as yes. That shift needs executive sponsorship, because the office almost certainly tightened up in response to an expectation from somewhere; without changing that expectation, any relaxation will be reversed within two quarters.