PERSPECTIVE
From Top-line to Truth: Why Your Unit Economics Are Wrong
Sitelines
Across consumer, services, venture-backed tech, healthcare, finance, and manufacturing, leaders still celebrate top-line growth and sanity-check it with “unit economics.” That is the 2002 playbook. In 2025, it is incomplete. The unit itself has changed, and the economics around it now include AI fragility, compliance drag, supply volatility, and human fatigue that GAAP and classic MBB lenses do not surface.
Pattern signals we see repeatedly:
- Revenue up, cash conversion negative. Deposits rise while obligations outpace collections.
- Contribution that looks fine until reality is priced in. After AI rework, compliance overhead, volatility reserves, and human drag, “healthy” units can quietly slip underwater.
- Portfolio sprawl and misallocated capacity. More SKUs, projects, or customers than the model can profitably carry. Headcount grows with volume, not contribution.
- Decision rights without guardrails. Additions and scaling decisions made on revenue, not adjusted economics.
The takeaway: revenue is noise unless the unit works, and most leaders are not measuring the real unit.
What traditional unit economics get right, and what they miss
Still true: if each incremental unit destroys value, growth multiplies loss.
What is outdated: the definition of the unit and the cost model ignore hidden costs of modern businesses:
- AI as a double-edged cost curve
Automation lowers apparent cost-to-serve, then introduces rework, reliability, compliance, and IP risk that do not sit cleanly on the P&L. - Data as capital, not overhead
Storage, governance, lineage, privacy, and model risk management are economic. Most firms bury them in G&A. - Regulated and clinical realities
In healthcare and finance, the unit is a single encounter, claim, or loan. Yet, most unit economics don't factor how outcome variability, liability, and audit friction change the math downstream. - Manufacturing and supply chain entropy
Carbon disclosure, climate shocks, re-shoring, and insurance premiums make volatility a structural cost. - Human fatigue as economic drag
Burnout, turnover, and low AI-augmentation efficacy do not appear in GAAP, yet in many organizations they silently crush realization and margin in real time.
The Harms lens: modernize the unit, then the economics
We translate the unit to the operating reality, then price modern drag into contribution.
- Retail and consumer: unit = SKU or basket. Price contribution with spoilage, promo dilution, labor alignment, and store ops drag.
- B2B services: unit = hour, project, or client cohort. Price realization, span of control, rework, bench time, and AI-assist efficacy.
- SaaS and tech: unit = customer, seat, or ARR block. Price CAC and LTV with AI oversight cost, model risk, privacy, and security lift.
- Healthcare: unit = encounter, procedure, or claim. Price reimbursement versus outcome variability, compliance, and malpractice cover.
- Finance: unit = loan, origination, or portfolio slice. Price yield versus WAC, loss reserves, and verification or KYC friction.
- Manufacturing: unit = widget or order. Price COGS versus volatility reserve, carbon and resilience premiums, and warranty tail.
Principle: do not replace GAAP, augment it. Use a supplemental view that turns hidden costs into unit-level reality leaders can act on.
A simple try-it-yourself exercise
Give your team a flashlight, then call us for the X-ray.
- Define your unit. SKU, seat, claim, project, loan, widget.
- Start with standard contribution. Revenue per unit minus direct costs.
- Layer four hidden cost categories. Proxies are fine.
- AI and tech fragility: percent of outputs needing rework times hours times loaded rate.
- Compliance and liability: annual regulatory or legal overhead divided by total units.
- Resilience and volatility: supplier price variance plus carbon or insurance reserve divided by units.
- Human drag: turnover cost-plus lost productivity plus low AI-assist efficacy divided by units.
- Recalculate adjusted contribution per unit. The delta versus standard contribution is your erosion zone.
Illustrative example:
Standard contribution per unit = $50
Adjustments (governance proxies): AI rework $5, Compliance $3, Volatility/Resilience $4, Human drag $2 Adjusted contribution = 50 − (5 + 3 + 4 + 2) = $36
Erosion % = (50 − 36) ÷ 50 = 28%
Definitions:
Threshold: direct overhead costs per unit + a 10–20% cushion.
Erosion: the % gap between standard and adjusted contribution.
If adjusted contribution wipes out more than 10 to 15% of contribution, or drops below your threshold to fund overhead and growth, you do not have a revenue problem.
You have an architecture problem.
Conversely, if erosion is above 10% treat it as a signal to tune price, cost, or design. If it stays above 15% for two consecutive quarters, act immediately: reprice, bundle, redesign, or pause.
How we implement this with clients
We will outline the approach so leaders understand the lift, without giving away our instrumentation.
1) Blind-spots map against the P&L, balance sheet, and cash flow. Annotate missing reality and tie each blind spot to a unit-level allocation.
2) Build the Adjusted Unit Economics Statement as a supplemental schedule beside your P&L:
- define the unit and its recognition rules
- standard contribution by product, service, or cohort
- adjustments with documented proxies and sources
- net adjusted contribution with confidence bands
- decision flags to prune, bundle, reprice, or re-sequence capacity
3) Quantify via guardrail proxies that are good enough to govern, then refine.
4) Stand up a live dashboard that shows standard versus adjusted contribution, the erosion delta, trend, and attribution by product, region, or cohort.
5) Embed decision guardrails into the operating rhythm:
- Gate new units: Only launch if the base-case adjusted contribution meets or exceeds your threshold.
- AI deployments: Approve only when the net adjusted ROI clears your hurdle (for example, 20%).
- Hiring and capacity: Add only when adjusted utilization and contribution support it.
- Capital allocation: Re-weight quarterly toward the strongest adjusted contributors; prune the bottom decile.
- Review Cadence: Monthly operations and finance reviews, quarterly portfolio pruning.
- Exceptions: Strategic loss leaders only allowed with explicit off-setting gains, or a time bound plan with a named owner to sunset.
6) Change architecture: decision rights, authority levels tied to thresholds, lightweight pricing or bundling tests, and quarterly portfolio pruning driven by dashboard deltas.
Decision thresholds to start
- Early warning: Erosion > 10% in a period, fix within 1–2 quarters.
- Action: If erosion > 15% for two consecutive periods or the adjusted contribution drops below threshold in any period, reprice, bundle, redesign, or pause.
- Hard stop: If adjusted contribution fails to cover direct overhead in any period, stop and redesign. If it stays below threshold for two periods, retire or pause.
- Scale signal: Concentrate capacity and capital on top-quartile adjusted contributors with stable or falling erosion.
What “good” looks like after implementation
- Clarity: Finance, leadership, and ops speak the same unit economics language.
- Discipline: Portfolio sprawl contained.
- Resilience: Cash conversion improves because volatility and fragility are priced in.
- Confidence: Boards and investors see economics that hold under modern constraints.
- Momentum: You scale what actually builds enterprise value.
Time to shift: 2 to 4 weeks to baseline the supplemental statement and dashboard. One to two quarters to embed guardrails and show compounding effects.
Author: Tisha Hartman, CEO, Harms Advisory Group.
Human-led. AI-assisted. Always accountable.