The Conversion Rate Your Board Has Never Seen

A 5,000-person company spending $150,000 in average total compensation makes a $750 million annual bet on human capital. That bet has no conversion metric.

There is no gauge that tells the board what percentage of that $750 million actually converts into proportional enterprise value. No instrument that separates the signal from the noise. No measurement that distinguishes between a company that is converting at 40% and one converting at 25%. Both companies will report similar engagement scores, similar eNPS, similar retention rates. The dashboard will look fine. The waste will be invisible.

We have spent twenty years building the instrument that makes it visible. This document explains what it measures, how the math works, and what we have found.

Why everything your CHRO measures is a proxy for something they can't see

The standard Human Capital ROI formula is (Revenue - Operating Costs) / Human Capital Costs. It produces a ratio. Most companies run between 1.2 and 2.5. The number is real. It is also structurally useless.

An HCROI of 1.8 could mean every employee converts uniformly at 1.8x. Or it could mean 25% of the workforce converts at 5x while the other 75% converts at 0.7x. The aggregate conceals the distribution. And the distribution is where every actionable decision lives.

Engagement scores have the same defect. We measured the correlation between reported engagement and actual value conversion across multiple organizations. The number: r = 0.31. Engagement explains less than 10% of the variance in conversion. Ninety percent of the signal is coming from somewhere engagement surveys cannot see.

Retention rates, time-to-fill, cost-per-hire, eNPS, training completion rates. All of these are lagging indicators of surface phenomena. None of them measure the thing that determines whether a company's human capital investment is compounding or decaying: the conversion efficiency of the system itself.

What we actually measure

The word "system" is doing the heavy lifting in that definition. OCE does not measure people. It measures the physics between the people and the work. A company can have exceptional talent and run at 28% conversion if the architecture around that talent is poorly configured. The waste is never in the people. It is in the conversion physics between the people and the work.

This is a different claim than "we have a talent problem." Most companies that think they have a talent problem actually have a conversion problem. The talent is already inside the building. The system is failing to metabolize it.

The four layers where conversion fails

Conversion loss does not happen in one place. It accumulates across four layers, each compounding the losses of every other. We discovered this structure empirically, not by designing a framework and fitting data to it, but by measuring organizations where the obvious explanations kept failing and following the residual variance until the actual sources of loss became visible.

You are introducing noise at the point of entry

Standard interview-based hiring operates at approximately r = 0.2 predictive validity. This means the hiring process explains roughly 4% of the variance in future job performance. The remaining 96% is noise, luck, and post-hoc rationalization.

No board would accept a capital allocation process that predicts 4% of variance in returns. Yet every board accepts exactly this for the single largest capital allocation category on their P&L.

We measure the predictive validity of the client's actual hiring outcomes at 6, 12, and 18 months against performance data. The gap between observed validity and the theoretical ceiling of r ~ 0.65 (achievable with optimally designed work-sample assessments) is Layer 1 conversion loss. In our deployments of TALOS, we have consistently closed this gap to r = 0.55. The practical effect: 40% fewer false-positive hires entering the system. Every false positive eliminated is conversion capacity recovered downstream.

Your best people are in the wrong seats and no one can see it

Once talent enters the system, the second conversion layer is the match between individual capability and the role where that capability produces maximum value. This is not the same as "right person, right seat." It is more precise and more uncomfortable than that.

We map capability distributions across the workforce using behavioral assessment data correlated against output metrics. We then model the theoretical optimal allocation: the configuration that would maximize total value output given the existing capability pool. The delta between actual allocation and optimal allocation is the Layer 2 loss.

The finding that recurs: 15 to 25% of the workforce in a typical organization sits in roles where their primary capability is either underutilized or misaligned with the value the role is designed to create. These individuals are not underperforming. Their managers rate them as adequate or above. The system has placed a $180K chess piece on a square where it controls two diagonals instead of seven. Nobody notices because nobody is measuring the board.

Your org chart is a fiction. The real network is invisible.

This is the layer most HR functions cannot see at all. Even with the right people in the right roles, conversion collapses if information does not reach decision-makers at the speed the business requires.

We deploy organizational network analysis to map actual information flow: not the formal hierarchy, but the real topology through which knowledge, coordination, and decisions move. In every single deployment, we identify bridge nodes: individuals who disproportionately control information transfer between clusters. These people are invisible to the org chart. When they leave, get overloaded, or are bypassed by a restructuring that doesn't know they exist, entire sections of the organization lose conversion velocity. Often permanently.

This is why 85% of restructurings fail. Not because of "change resistance." Because the restructuring severed bridge nodes it couldn't see.

We also measure decision latency and what we call alignment theater: coordination overhead that produces agreement but not decisions. In our C-suite assessments, executive teams spend 20 to 30% of their bandwidth on alignment theater. That is a direct, quantifiable tax on the organization's conversion rate. It does not appear on any dashboard. It is not captured by any engagement survey. It is a structural loss, visible only through network measurement.

Your reward system is optimizing for the wrong behaviors. The math proves it.

The final layer measures the gap between what the organization's incentive architecture actually reinforces and what produces value. This sounds like it should be obvious. It is not. It is the most common source of systemic conversion failure, and the most invisible, because the people inside the system believe their incentives are aligned. They are not measuring. They are assuming.

We map actual behavioral reinforcement patterns using compensation data, promotion velocity, recognition distribution, and departure triggers. We then correlate these against value-producing behaviors as measured by output data. The divergence between "what the system rewards" and "what produces value" is the Layer 4 loss.

The canonical case: a moonshot laboratory where teams clung to failing projects because the informal incentive structure punished project termination. Career advancement was tied to project longevity, not project quality. Millions burned monthly on zombie projects that everyone privately knew should die. We redesigned the incentive architecture to reward early, honest termination. Result: $40M+ in sunk costs recovered. The people did not change. The talent did not change. The conversion physics changed.

A more common case: a consumer brand where the top five individual performers by metrics were simultaneously the top five sources of team attrition. The incentive system rewarded individual output. It did not measure the team damage produced by the behavior that generated that output. Three of the five were net-negative value contributors when system-level effects were included. The Narcissism Index was built to make this visible.

The multiplication problem

Here is the insight that changes how you think about this permanently.

The four layers are multiplicative. Not additive. Not averaged. Multiplied.

This is why the number is always lower than anyone expects. A company can be reasonably competent at each layer individually and still convert barely a third of its human capital investment. The layers compound each other's losses. A 10% improvement in Layer 1 only matters if Layers 2, 3, and 4 can carry the improved signal through to output. If Layer 3 is bottlenecked, improving hiring quality at intake produces no change in composite conversion. The better talent enters the system and hits the same information friction as everyone else.

This is also why most HR initiatives fail to produce measurable business impact. They improve one layer without measuring the others. The intervention is real. The composite doesn't move. The CHRO cannot explain why. The board loses confidence. The cycle repeats.

The question is never "are our people good enough." In twenty years of measurement, we have never found an organization where the people were the problem. The question is: what percentage of their capability is the system converting into value, and where, specifically, is the rest going?

What we have found

Across deployments spanning moonshot laboratories, state energy enterprises with 300,000+ employees, unicorn SaaS companies, global consumer brands, PE portfolio companies, and Series B startups, we have observed composite OCE scores ranging from 22% to 41%.

The median is approximately 33%.

We have never encountered an organization running above 41% on first measurement.

These numbers are consistent across industries, geographies, and scale regimes. A 200-person startup and a 45,000-person industrial conglomerate can run at the same composite OCE. The sources of loss differ. The structural reality does not. Most organizations are converting roughly a third of the talent they pay for into proportional value.

The remaining two-thirds is not waste in the sense of people doing nothing. It is waste in the sense of a thermodynamic system running below its theoretical efficiency. People are working. They are working inside a system that is failing to convert their effort, judgment, and capability into output at anything close to the rate the capability distribution would predict.

What 5 points of OCE is worth

For the 5,000-person company at $150K average total comp, total human capital investment is $750M per year. At a composite OCE of 33%, approximately $502M annually is failing to convert into proportional value.

A 5-point improvement in composite OCE, from 33% to 38%, represents roughly $37.5M in additional value conversion per year.

Most HR transformation budgets run $2M to $5M. The return on a correctly targeted OCE intervention is an order of magnitude larger than the investment. But the key phrase is "correctly targeted." Improving the wrong layer, or improving the right layer in the wrong sequence, moves nothing. A perfectly executed talent acquisition overhaul has zero composite impact if Layer 3 is the bottleneck. The diagnostic must precede the intervention. Without knowing which layer is the binding constraint, any HR initiative is a guess with a budget attached.

Why this instrument didn't exist before

Three reasons.

First, the measurement requires data that lives in four different systems that almost never talk to each other: hiring process data, performance and capability data, communication and collaboration metadata, and compensation and incentive data. Most organizations have all four. None of them have connected them into a single measurement instrument. The data exists. The integration does not.

Second, the multiplicative structure is counterintuitive. Every instinct, every business review, every HR dashboard is built on additive logic. Improve hiring by 10%, get 10% better outcomes. The reality is that improving hiring by 10% while running 75% information flow efficiency yields a 7.5% improvement, not 10%. Organizations do not naturally think in multiplicative terms about their human systems. They do about their financial systems. They do about their engineering systems. They do not about their people systems. This is a blind spot at the level of organizational epistemology.

Third, HR as a function has historically measured activity, not conversion. Programs delivered, surveys completed, positions filled, training hours logged. All inputs. None of these measure the rate at which the system converts those inputs into value output. The measurement tradition of the function is not equipped to see what OCE makes visible. The instrument had to be built from outside the tradition.

Every company has this number. Almost none of them know what it is. The first step is measuring it. The second step is accepting what it reveals.