Scoring Model

A Scoring Model is a decision rule that does three things:

1. 1)Decomposes your Utility Function into measurable criteria (Expected Return, Execution Risk, Time Horizon, strategic fit)
2. 2)Weights each criterion to reflect how much it matters to your P&L goals
3. 3)Scores every competing option against those criteria and sums the weighted scores into a single number

The output is a ranked list. Option A scores 7.8, Option B scores 6.1, Option C scores 8.3 - now you know where to Allocate first.

You already know Expected Value gives you a probability-weighted dollar figure. You already know your Utility Function encodes what you actually value beyond dollars. The Scoring Model is the instrument that makes both operational - something you run every quarter when deciding where Budget goes.

Operators face a specific problem: you always have more plausible investments than Budget to fund them. This creates opportunity cost pressure - every dollar allocated to Project A is a dollar not allocated to Project B.

Without a Scoring Model, allocation decisions happen one of two ways:

• •Loudest voice wins. Whoever pitches most passionately gets funded. This optimizes for presentation skill, not P&L impact.
• •Recency bias. The last project you heard about feels most urgent. Your Allocation drifts toward whatever crossed your desk this morning.

A Scoring Model forces you to evaluate every option against the same criteria with the same weights. It makes your decision rule explicit and auditable. When your CFO asks why you funded the churn model over the pricing model, you can point to the scores instead of saying "it felt right."

This matters even more in PE-Backed environments where Capital Allocation is scrutinized quarterly. Your Scoring Model is how you demonstrate capital discipline.

Step 1: Choose criteria from your Utility Function

Your Utility Function tells you what you value. Translate that into 4-6 measurable criteria. For an AI investment portfolio, common ones are:

Step 2: Assign weights that sum to 1.0

Weights reflect your risk appetite and Time Horizon. An Operator focused on near-term EBITDA Optimization might weight like this:

• •Expected Return: 0.30
• •Execution Risk: 0.25
• •Time to Value: 0.25
• •Strategic fit: 0.10
• •Revenue vs. Cost Reduction: 0.10

An Operator building a Compounder with a longer Investment Horizon might put 0.25 on strategic fit and 0.10 on Time to Value - the inverse.

Your weights ARE your Utility Function made numeric. Two Operators with different weights will rank the same projects differently, and both can be rational.

Step 3: Score each option 1-10 on every criterion

Use anchors to keep scores consistent:

• •10: Best realistic outcome (e.g., 5x ROI, near-zero Execution Risk)
• •7: Strong (e.g., 2-3x ROI, team has done this before)
• •4: Mediocre (e.g., 1x ROI, significant unknowns)
• •1: Poor (e.g., negative Expected Value, unproven technology)

Score each project before looking at the weighted totals. This prevents you from reverse-engineering scores to justify a decision you already made.

Step 4: Multiply and sum

For each project: Total Score = Σ (weight_i × score_i)

Rank by total score. Fund from the top until Budget runs out.

Use a Scoring Model when:

• •You have 3+ competing options for the same pool of Budget or capacity. Two options you can compare directly. Eight options need a system.
• •The decision involves multiple criteria that trade off against each other. If only Expected Return mattered, just rank by Expected Value and you are done.
• •You need to defend the decision to someone else - your CFO, your board, your PE sponsors. The model is the audit trail.
• •You will repeat this decision quarterly or annually. The upfront cost of building the model pays off when you reuse it every cycle.

Skip it when:

• •One option clearly dominates on every criterion (Dominant Strategy - no model needed)
• •The decision is trivially small relative to your P&L
• •You lack data to score meaningfully - in that case, invest in the Value of Information first

A Scoring Model sits at the junction of Expected Value and Utility Function - it takes the dollar estimates from Expected Value calculations and filters them through the preference weights your Utility Function defines. Without Expected Value, you have no inputs to score. Without a Utility Function, you have no basis for setting weights. Downstream, your Scoring Model feeds directly into Capital Allocation and Capital Budgeting decisions - it is the mechanism by which you turn a pile of proposals into a funded Portfolio. It also connects to Sensitivity Analysis, which stress-tests whether your ranking is robust to uncertainty in the weights themselves. When your Scoring Model produces tight rankings (projects within 0.5 points of each other), that is a signal to invest in Value of Information - gather more data before committing Budget rather than trusting a razor-thin margin.