What is the Portfolio Optimization Analysis?

The Portfolio Optimization Analysis helps you decide how to allocate your limited capacity across multiple bets. Instead of picking a single “#1 bet,” it recommends a balanced portfolio that maximizes expected value while managing risk, overlap, and diversification constraints.

If Bet Impact tells you what looks high leverage, Portfolio Optimization helps you answer: what mix of bets should we run together — and how concentrated should we be?

What Value It Gives

• Better allocation – Move from “top 3 bets” to an intentional mix of investments.
• Risk-adjusted planning – Balance upside with uncertainty so you don’t overcommit to fragile plans.
• Diversification by design – Keep healthy coverage across categories (growth, platform, retention, etc.).
• Constraint-aware decisions – Respect required bets, prohibited bets, dependencies, and allocation bounds.
• Clear trade-offs – Make “more return vs more risk” explicit with portfolio-level risk metrics.

Common Use Cases

• Quarterly planning – Allocate effort across growth, retention, and platform without betting everything on one thesis.
• Team capacity splits – Decide how much capacity goes to product growth vs reliability vs lifecycle.
• High-variance roadmap – Reduce the chance that one failed assumption derails the whole quarter.
• Dependency-heavy work – Ensure platform bets are funded when feature bets depend on them.
• Leadership alignment – Turn “we should do everything” into a portfolio with explicit weights and constraints.

How Portfolios Are Optimized

Portfolio Optimization models each bet as an investment with:

• Expected return (your estimated impact)
• Risk (how uncertain/variable the outcome is)
• Optional correlations (which bets tend to succeed/fail together)

Then it searches for a set of allocation weights (0–1) that sum to 1.

Under the hood (methodology)

1. Build a risk model Segflow converts bet risks and optional correlations into a covariance matrix (a map of how bets interact).

2. Optimize for return vs risk The model balances expected return against portfolio variance using a configurable risk aversion setting:

   • lower risk aversion → more concentrated, higher-upside allocations
   • higher risk aversion → more diversified, lower-volatility allocations

3. Apply constraints and guardrails The optimizer can enforce:

   • max/min number of bets
   • required/prohibited bets
   • category diversification bounds
   • dependency floors (if a bet is selected, its dependencies are funded)
   • min/max allocation per bet (caps or floors)

4. Compute portfolio risk metrics The result includes risk-adjusted metrics (Sharpe, VaR, CVaR) to help you interpret how conservative vs aggressive the portfolio is.

Interpreting the portfolio table

Each selected bet includes:

• Weight: the recommended share of your portfolio focus/capacity (sums to 1 across bets)
• Contribution: how much expected return this bet contributes (weight × expected return)
• Risk contribution: how much this bet contributes to portfolio risk given correlations
• Effort share: effort weighted by its allocation (useful for comparing “capacity consumed” across bets)

Example: Building a Balanced Quarter (Illustrative)

Scenario: You have 10 candidate bets. Some are high-upside growth bets, but uncertain. Some are low-risk platform investments. You want a portfolio that can still succeed if one growth bet underperforms.

How you’d use Portfolio Optimization:

1. Start with a scored list of bets (often from Bet Impact Analysis).
2. Add basic risk estimates (or use uncertainty proxies) and categorize bets (growth, retention, platform).
3. Add constraints (e.g., require the reliability bet; cap any single bet to 30%; keep at least 2 platform bets).
4. Run the analysis at a moderate risk aversion setting, then compare with a more aggressive and a more conservative run.

What you do next:

• Use the recommended weights as a starting allocation for resourcing and sequencing.
• Cross-check the portfolio against your strategy: do you have the right diversification and coverage?
• Re-run when assumptions change (new data, new bets, shifting constraints).

What You Provide

Required Inputs

• Bets – The set of candidate initiatives you’re considering.
• Expected return – An impact estimate per bet (often derived from Bet Impact scores).
• Risk – A proxy for uncertainty/variance per bet (e.g., uncertainty bands, historical volatility, or a calibrated heuristic).
• Effort – Resource requirement (person-weeks, points, or normalized effort units).

Optional Inputs (Advanced)

• Correlations – If some bets are likely to succeed/fail together, provide correlations to avoid false diversification.
• Categories – Labels like growth/platform/retention to enforce diversification.
• Dependencies – Ensure supporting bets are funded when dependent bets are selected.
• Allocation bounds – Floors for required bets and caps to avoid overconcentration.
• Constraints – Max bets, min bets, prohibited bets, diversification ranges.
• Risk settings – Risk aversion, confidence level for VaR/CVaR, risk-free rate for Sharpe.

Strategic Scoring Fields

These optional fields integrate outputs from other Segflow models to create strategically-informed portfolios:

• confidenceScore (0-1) – From BetImpactModel. Discounts expected return based on estimate confidence. A score of 0.7 means 70% of the expected return is used in optimization.

• driverImportance (0-1) – From KeyMetricDriverModel. Weights bets by how much they affect key business drivers. Higher values prioritize bets on high-importance metrics.

• loopGain – From GrowthLoopModel. The raw amplification factor from growth loop analysis (>1 = amplifying loop, < 1 = decaying). Used directly as a multiplier on expected return (capped at 3x).

• costOfDelayPerDay – From CostOfDelayModel. Higher values indicate more time-sensitive bets that lose value by waiting.

Cost of Delay Constraints

When bets have high cost-of-delay scores, you can use these constraints to prioritize them:

• codPriorityThreshold – Bets with costOfDelayPerDay above this threshold receive priority treatment.

• codMinimumWeight (default: 0.1) – Minimum portfolio weight for bets that exceed the CoD threshold. Ensures high-urgency bets aren’t under-allocated.

Example: Setting codPriorityThreshold: 1000 and codMinimumWeight: 0.15 ensures any bet costing over $1000/day in delay gets at least 15% portfolio weight.

What You Get Back

Core Outputs

• Recommended allocation – A weighted portfolio across bets (weights sum to 1).
• Contribution breakdown – Which bets drive expected return and which drive risk.
• Diversification view – Allocation by category (e.g., growth vs platform).
• Constraints applied – The portfolio alongside the constraints you requested.

Risk Metrics

• Expected return – The portfolio’s weighted expected value.
• Standard deviation – A portfolio-level uncertainty/risk proxy.
• Sharpe ratio – Return per unit of risk (higher is better risk-adjusted).
• VaR / CVaR – Conservative downside metrics at your chosen confidence level.

How to Interpret Results

• Treat weights as guidance, not autopilot – Use them to shape resourcing and sequencing, then apply judgment.
• Compare portfolios by risk appetite – Run conservative vs aggressive risk aversion and see what changes.
• Watch concentration – If one bet dominates, ask if you’re comfortable with that dependency on one thesis.
• Use diversification intentionally – Category limits help ensure you don’t starve platform work or overinvest in a single area.
• Correlations matter – Two “different” bets can still be the same underlying risk; correlations make that visible.
• Don’t over-trust precision – The value is directionally better allocation and clearer trade-offs, not perfect optimization.

Best Practices

• Start with Bet Impact to quantify expected value; use Portfolio Optimization to allocate across bets.
• Use Cost of Delay to sequence within a portfolio when timing matters.
• Keep risk estimates consistent (even if coarse) so comparisons are fair.
• Define categories that reflect real strategic buckets your org cares about.
• Re-run on a cadence (weekly/biweekly during planning) and after major assumption changes.

Summary

The Portfolio Optimization Analysis helps you move from ranking bets to building a portfolio. Use it to allocate capacity across initiatives in a way that balances expected value, risk, and strategic constraints — so your roadmap succeeds even when reality deviates from plan.