The problem: ten competing agendas and no single decision engine
Most companies don’t lack ideas. They lack a mechanism to rank them.
So the default system becomes:
- the loudest person wins,
- “urgent” crowds out “important,”
- and capital leaks into half‑finished initiatives.
The fix is not more spreadsheets. It’s governance with teeth.
Capital allocation is strategy with receipts
Strategy is what you say you’ll do. Capital allocation is what you actually did—measured in dollars, time, and attention.
Many capital allocation frameworks start with decision governance (who decides, when, with what evidence).
The investment committee: what it is (and what it isn’t)
It is:
- a forum that owns the rules,
- ranks the top initiatives,
- and enforces stop/continue decisions.
It is not:
- a meeting where every project gets negotiated in public,
- or a place where the CFO “does math” while everyone else vibes.
The CFO’s job is often better framed as designing discipline:
- definitions,
- thresholds,
- data integrity,
- and the cadence that makes decisions repeatable.
“One queue” beats ten shadow portfolios
A workable model is simple:
- Single intake (no drive-by “can we just…”)
- Standard one‑pager business case
- Scoring rubric
- Ranked queue
- Funding decisions + stage gates
- Post‑completion review
If an initiative can’t get through intake, it doesn’t deserve stealth funding.
Buckets: a portfolio architecture that fits most companies
Buckets prevent “everything is top priority.”
A minimal taxonomy:
- Run: keep the machine operating (compliance, uptime, core ops)
- Improve: efficiency, margin, cycle time
- Grow: revenue expansion, new channels, new products
- Risk: controls, resilience, security, regulatory exposure
- Option bets: small experiments with capped downside
No mysticism required—just clarity.
Stage gates and tranche funding (capital earns its next dollar)
Instead of funding a project “because it’s strategic,” fund it because it met evidence thresholds.
- Stage gate: the proof required to move forward (definition, owner, baseline, feasibility)
- Tranche: partial funding released per gate
This forces learning early and avoids “sunk cost cosplay.”
(These concepts show up in many business‑case and allocation frameworks: governance, stage gates, and tranche-based asks are common ways to reduce waste.)
The scoring rubric (Contour-style, but portable)
Pick 6–8 dimensions. Keep it boring.
Examples:
- Strategic fit (explicitly tied to priorities)
- Economic impact (range, not a fantasy point estimate)
- Time to impact
- Risk / reversibility
- Dependencies (people, systems, vendors)
- Data confidence (how much is measured vs guessed)
- Operating load (ongoing maintenance cost)
If “data confidence” isn’t scored, you’ll fund the best storyteller.
Post-completion review: close the loop
A portfolio becomes intelligent when it remembers:
- what we assumed,
- what happened,
- what we’d do differently next time.
This doesn’t need theatrics. It needs a template and a calendar event.
The test: can you say “no” cleanly?
If you can’t kill projects, you don’t have an allocation system. You have a museum of unfinished intentions.
A good committee produces:
- fewer initiatives,
- clearer owners,
- explicit trade-offs,
- and a queue that matches reality.
Not perfection—just a decision engine that doesn’t depend on mood.