Major financial-services software organization: Improving predictability across an internal project portfolio

Client context

A major financial-services organization had a large internal software-development environment responsible for delivering business-critical IT projects.
The organization had many capable people and a substantial project portfolio, but leadership faced a recurring problem: delivery dates were difficult to predict with sufficient confidence.
The business impact was significant.
When project delivery dates were uncertain, the organization could not reliably plan business releases, allocate resources, coordinate dependencies, manage stakeholder expectations or make confident portfolio decisions.

Sponsor challenge

The sponsor challenge was to improve predictability across the software project portfolio.
The relevant questions were:

Why are project schedules not sufficiently reliable?
Which parts of the project portfolio process create uncertainty?
Are estimates, plans, dependencies, quality criteria and progress signals strong enough?
Do project managers and middle managers share a professional language for managing delivery?
How can leadership improve predictability without simply adding more reporting?
How can the organization raise the professional level across many projects, not only fix individual troubled projects?

This was not simply a scheduling problem.
It was a portfolio governance and capability problem.

Advisory contribution

As part of the Whitebox advisory work, I was centrally involved in applying and operationalizing an assessment-based portfolio approach.

The approach combined two executive decision dimensions:

Expected value — the expected impact of an initiative on organizational maturity, performance or strategic goals.
Predicted success — the assessed likelihood that the initiative would actually succeed, using ImprovAbility parameters related to initiation, project execution, deployment and organizational foundation.

This created a simple but powerful portfolio view: initiatives could be discussed not only by ambition or political relevance, but by their expected value and their realistic chance of success.

Results and relevance

The organization achieved a higher level of professionalism in its project organization, and delivery schedules became significantly more reliable.
The work helped create a stronger shared language around project capability, planning discipline, quality, progress insight and portfolio predictability.
It also created an internal group of strong project managers who could support others and act as carriers of improved practice across the organization.
For my current owner-side governance advisory, this case is highly relevant because it demonstrates the ability to improve predictability in a complex internal software-development portfolio.
It validates competences in:

Portfolio governance
Software-development capability improvement
Predictability analysis
Project and portfolio process improvement
CMMI-based maturity and capability development
Training of senior management, middle management and project managers
Development of internal improvement agents
Translation of delivery uncertainty into management action
Raising professional standards across a project organization
Improving the sponsor’s ability to rely on schedules and portfolio commitments

Core lesson

In large software organizations, poor predictability is rarely caused by weak schedules alone.
It is usually the result of interacting factors: weak estimation practices, unclear quality criteria, immature progress signals, unstable dependencies, inconsistent project leadership practices and portfolio processes that do not reveal uncertainty early enough.
The sponsor’s task is not simply to demand more accurate dates.
The sponsor’s task is to create the governance, capability and professional discipline that make reliable dates possible.

Competences validated

Improving portfolio predictability by strengthening the professional system that produces schedules, commitments and delivery confidence.