Rule of 60 vs Cost Per Outcome: The New Portco KPI

Vista Equity Partners popularized the "Rule of 60" metric for valuing SaaS companies—a formula weighting growth, margins, and retention. It became gospel for assessing software company health. Now Vista and others are applying it to AI-driven workflows, and it is wrong for PE operating partners. The Rule of 60 measures financial health. It does not measure operational efficiency. For AI-powered work, the metric that matters is cost per outcome: the all-in cost to execute one unit of work (one resolved support ticket, one adjudicated claim, one loan application) relative to the prior manual process.

This article explains why cost-per-outcome is the true operating partner metric and walks through how to implement it across your portfolio.

Why Rule of 60 Misses the AI Operating Leverage

The Rule of 60 combines three metrics: magic number (ARR growth rate), net revenue retention, and gross margin, weighted equally. It's elegant for SaaS because these three inputs fully capture the health of a software business. For a software company, they tell you everything: are we growing fast (magic number), retaining customers (NRR), and doing it efficiently (margins)?

But for AI-driven operational workflows, the Rule of 60 is a red herring. Here's why.

Consider an insurance underwriting operation that deployed AI to process applications. In year one, the team handled 50,000 applications at a manual cost of $65 per application ($3.25M total cost). In year two, with AI assistance, they handled 75,000 applications at a cost of $42 per application ($3.15M total cost). The Rule of 60 would say: marginal improvement. Revenue is flat (no top line), costs are slightly down. Magic number is low.

But the operating reality is this: the same team is 50% more productive. The cost per work unit dropped 35%. The operation is now scalable. This is exactly what an operating partner wants to see.

The Rule of 60 does not capture this. Cost per outcome does, immediately.

Cost Per Outcome: The Right Operating Metric

Cost per outcome is defined as: total all-in cost of the AI-assisted workflow divided by the number of work units completed, compared to the prior manual cost per work unit. The "all-in" cost includes visible API spend, hidden inference, vector database storage, observability, human review, retry penalties, and the fully loaded cost of any dedicated AI ops person.

The formula is simple:

Cost Per Outcome = (Total Monthly AI Spend) / (Monthly Work Units Completed)

For the insurance example:

• Total AI spend: $3.15M per year / 12 months = $262.5K per month
• Work units: 75,000 applications / 12 months = 6,250 per month
• Cost per outcome: $262.5K / 6,250 = $42 per application

Compare that to the prior manual cost of $65 per application. The delta is $23 per application. On 75,000 annual applications, that's $1.725M of economic value creation.

That is the conversation your operating partner should be having with the CFO and COO.

The Three Reasons Cost Per Outcome Beats Rule of 60

First, it is outcome-focused, not financial-focused. The operating partner's mandate is to create value in the portfolio company. Value, in the context of AI-powered work, is measured in efficiency—how much work can the team do with the same headcount and capital? Cost per outcome answers that directly. Rule of 60 obscures it by treating the operation as a black box.

Second, it is benchmarkable across portcos. The moment you have cost per outcome for your healthcare claims operation, your insurance underwriting operation, and your staffing client onboarding operation, you can ask: "Which operation is generating the most AI-driven margin expansion? Where should we double down?" You cannot ask that question with Rule of 60. Rule of 60 is an absolute metric, not a relative one.

Third, it drives the right behaviors. When you tell a COO "your magic number is 1.5x," the COO thinks about pricing and top-line growth. When you tell a COO "your cost per outcome should drop 20% in the next 12 months," the COO thinks about workflow optimization, vendor consolidation, and automation depth. The latter is where AI value actually gets created.

A Worked Example: Claims Processing

Let's model a 50-person claims processing operation for a regional health insurer, processing 280,000 claims per year.

Year 1 baseline (manual process):

• Annual claims processed: 280,000
• Cost per claim: $32 (fully loaded labor, systems, overhead)
• Total annual cost: $8.96M
• Cost per FTE per month: $14,800 (280,000 claims / 50 FTEs / 12 months = 467 claims per FTE per month)

Year 2 (with AI-assisted adjudication):

The claims team piloted an AI agent on 30% of incoming claims for six months, then rolled out to 100%. The AI agent pre-adjudicates claims, flags for human review if confidence is below a threshold, and routes complex cases to specialists.

• Annual claims processed: 280,000 (same volume)

• Headcount: 52 FTEs (hired 2 additional people to manage the AI system and quality)

• AI spend breakdown:

  • OpenAI API (inference): $8,400/month
  • Vector DB and observability: $2,100/month
  • Human review time and QA: $6,200/month (19 claims per FTE per month that need review)
  • Dedicated AI ops person: $18,500/month (fully loaded)
  • Cloud infrastructure and retries: $1,800/month
  • Total AI spend: $37,000/month = $444,000/year

• Claims processed per FTE per month: 280,000 / 52 FTEs / 12 months = 449 claims per FTE

• Cost per claim with AI: $444,000/year + $7.592M labor + $968K overhead = $9.004M / 280,000 claims = $32.16 per claim

Wait—that doesn't look like value creation. Let's recalculate.

Actually, here's the key: the human review cost embedded in AI ops is $6,200/month, not the full cost of 52 FTEs manually adjudicating claims. The right calculation is:

Manual process cost: 50 FTEs × $14,800/month × 12 months = $8.88M

AI-assisted cost:

• 50 FTEs doing AI-assisted work (majority pre-adjudicated by system) at lower cognitive load: $7.2M
• 2 FTEs managing AI system and quality: $370K
• AI infrastructure and vendor: $444K
• Total: $8.014M

Cost per claim with AI: $8.014M / 280,000 = $28.62 per claim

Value created: $32.00 - $28.62 = $3.38 per claim × 280,000 claims = $946,400 per year

The operating partner's thesis: "We've reduced cost per claim by 10.6% while maintaining the same volume and headcount level. The AI system is sustainable (we hired dedicated ops), and the margin expansion is real and auditable." If you can scale this to three more insurance operations in your portfolio, the total value is $3.8M annually.

That is the conversation Rule of 60 cannot have.

Implementing Cost Per Outcome Across Your Portfolio

Start with your largest or most AI-intensive portcos. Work with the CFO and COO to:

1. Establish a baseline. What is the manual cost per work unit today? For a claims operation, it's claims per FTE per month and cost per claim. For customer support, it's tickets resolved per agent per month and cost per ticket. Get this dialed in; it's your control number.

2. Build the AI spend model. Before deployment, estimate the all-in monthly cost including APIs, infrastructure, observability, human review, and people. Do not just use the vendor's API cost. Use the AI Cost Iceberg framework to anticipate the hidden costs.

3. Track cost per outcome monthly. Once the AI is live, measure it every month. Most operations take 3–4 months to reach stable cost per outcome. By month 6, you should have a clear trend line.

4. Benchmark across portcos. Once you have 3–4 portcos with stable cost per outcome metrics, compare them. Why does healthcare claims cost $28.62 per claim while insurance underwriting costs $41 per application? What can you learn from the most efficient operation?

This sits at stage 4 of the 5-Stage AI Cost Maturity Curve: AI spend is tied to work items, with clear cost-per-outcome KPIs that drive operating decisions.

Why Vista's Rule of 60 Still Works (But For Different Purposes)

To be fair to Vista: the Rule of 60 works brilliantly if your AI-driven operation is generating revenue, not just reducing cost. If an AI sales agent is prospecting and closing deals, or an AI insurance pricing engine is unlocking new policies, then top-line growth matters enormously and Rule of 60 applies. But for the majority of AI deployments in mid-market PE—operations that are about efficiency, not revenue—cost per outcome is the true measure of value creation.

The best operating partners track both. They use Rule of 60 to understand the financial trajectory of the business. They use cost per outcome to drive behavior change and measure operational excellence in their AI initiatives. For more on embedding this KPI into portfolio-wide governance, see the article on AI strategy for PE-backed CFOs. Operating partners running this analysis across a portfolio can request the PE Operating Partner Field Guide and the cost-per-outcome KPI tracking template.