07 · Leading Agile at Portfolio Level¶
At the portfolio level, the unit of work isn't a story or a sprint — it's an investment decision across many products and teams, competing for finite capacity. Agile principles still apply, but the tools change: backlogs become investment themes, and "sprint planning" becomes capital allocation.
Portfolio-level agile concepts¶
| Team-level concept | Portfolio-level equivalent |
|---|---|
| Product backlog | Portfolio backlog of epics/investment themes |
| Sprint planning | Quarterly (or PI-based) capacity allocation across themes |
| Definition of Done | Portfolio-level exit criteria (an initiative is "done" when it hits a measurable business outcome, not just ships) |
| Retrospective | Portfolio-level inspect-and-adapt: which bets paid off, which didn't, and why |
Prioritizing at portfolio scale: WSJF¶
Weighted Shortest Job First (from SAFe) prioritizes by economic value divided by effort, favoring high-value, low-effort work and explicitly penalizing large, slow-to-deliver bets that tie up capacity for a long time.
| Component | Question it answers |
|---|---|
| User/business value | How much value does this unlock? |
| Time criticality | Does the value decay if delayed (e.g., a market window)? |
| Risk reduction / opportunity enablement | Does this reduce risk or unlock other future work? |
| Job size (denominator) | How much capacity does this consume? |
WSJF score = (value + time criticality + risk reduction) ÷ job size. Rank by score, not by whoever has the most political capital to push their initiative up the list.
Governing without micromanaging teams¶
| Portfolio-level lever | What it controls | What it must NOT control |
|---|---|---|
| Which initiatives get funded/staffed | Investment allocation | How the funded team runs its sprints internally |
| Exit criteria for an initiative | What "success" means at the business level | The team's technical implementation choices |
| Quarterly re-prioritization | Whether an initiative keeps its funding into the next period | Fixing scope irrevocably for a whole year |
A worked example¶
A portfolio committee is choosing among four proposed initiatives with a fixed capacity of 40 team-quarters for the next year: a large platform rewrite (high estimated value, very high effort, no time pressure), a compliance-driven feature (moderate value, moderate effort, hard external deadline), a customer-requested integration (high value, low effort, no deadline pressure), and an internal tooling improvement (low direct business value, low effort, reduces risk of a known recurring incident).
Scoring each with WSJF surfaces that the customer-requested integration (high value, low effort) and the compliance feature (moderate value, hard deadline, so effectively high time-criticality) score highest, while the platform rewrite — despite being the initiative most executives are excited about — scores lowest, since a large denominator (job size) dominates. Rather than funding the rewrite first because of its visibility, the committee funds the two highest-WSJF items fully, allocates a smaller protected slice to the tooling improvement (justified by risk reduction rather than raw value), and defers the platform rewrite to next quarter's re-prioritization once the higher-scoring bets have delivered.
How It Actually Works¶
WSJF's real function is forcing a hidden variable — the opportunity cost of time — into a decision that would otherwise be made on visibility and political capital, and the platform-rewrite scoring last is the direct mechanical consequence of what the formula deliberately penalizes.
Why job size sits in the denominator, not as a separate factor to "consider." Two initiatives with identical total value produce very different economic outcomes if one takes one quarter and the other takes eight — the fast one returns its value (and frees capacity for the next bet) seven quarters sooner, which compounds across everything that capacity could have been used for in the meantime. Placing job size in the denominator makes this compounding effect explicit and comparable across very differently-shaped initiatives, rather than leaving "but it'll take years" as a vague, easily-overridden caveat attached to an otherwise impressive value story — which is exactly the caveat that gets waved away when the initiative also happens to be the one executives are most excited about.
Why time criticality has to be scored separately from raw value, not folded into it. Two initiatives can unlock identical dollar value, but if one's value decays with delay (a market window, a regulatory deadline) and the other's doesn't, treating them as equally urgent misallocates capacity — the decaying-value initiative loses real value for every quarter it waits, while the stable one loses nothing by waiting. This is precisely why the compliance feature outranks the platform rewrite despite lower raw value: its effective urgency (hard external deadline) is a cost of delay the rewrite simply doesn't carry, and WSJF's separate time-criticality term is what makes that difference visible in the ranking instead of buried in qualitative discussion.
Why "exit criteria, not implementation" is the correct governance boundary, and violating it recreates enterprise micromanagement. A portfolio committee dictating a team's internal technical choices is making a decision with far less context than the team actually has about its own codebase and constraints — this is the identical authority-mismatch failure as Module 03, Level 4's stage-gated CAB reviewing every release regardless of risk, or Module 05, Level 4's coach becoming a mandatory approval step: a higher layer inserting itself into decisions the lower layer is better positioned to make. Setting the exit criterion (a measurable business outcome) and then stepping back is what lets the portfolio layer govern what gets funded without recreating the very micromanagement agile principles were adopted to remove at the team level.
Exercise¶
Score these three initiatives with WSJF (use a 1-10 scale for each component, effort 1-10 for job size) and justify your numbers: (1) a new market-entry feature with high value but a 6-month delivery window and no external deadline, (2) a security patch required by a regulator in 60 days, moderate effort, (3) a UX polish item, low effort, no deadline, modest value. State which you'd fund first under limited capacity and why, and name one thing the portfolio committee should NOT dictate to the team once an initiative is funded.