Apportionment, EMVR, and the Patent Damages Expert: A Field Guide

Apportionment is the default rule for patent damages. The entire market value rule is the exception, and the exception is narrow enough that very few damages experts should be invoking it. Experts who treat EMVR as a presumption rather than as a narrow doctrinal carve-out continue to be excluded under Rule 702, and the en banc decision in EcoFactor, Inc. v. Google LLC (Fed. Cir. 2025) tightens the front-end gatekeeping further.

This is a field guide to the doctrine, the categorical failure modes that have produced exclusions on WitnessLens, and the drafting choices that put a damages report on the right side of the line.

What apportionment requires

The apportionment principle is older than the modern patent damages framework. Garretson v. Clark, 111 U.S. 120 (1884), stated the rule that the patentee must "give evidence tending to separate or apportion the defendant's profits and the patentee's damages between the patented feature and the unpatented features." That formulation has been the operative requirement for 140 years.

The modern articulation requires the damages expert to identify the patented feature within the accused product, identify the value contribution of that feature relative to the rest of the product, and base the royalty calculation on the apportioned value rather than on the product as a whole. The methodologies for performing the apportionment vary. Some experts apportion on the royalty base by identifying the smallest salable patent-practicing unit. Others apportion on the royalty rate by adjusting the rate downward to reflect the patented feature's share of value. Some apportion through both base and rate.

What courts require is not a particular methodology. It is that the methodology be tied to the patented feature, that the analytical steps be explicit, and that the conclusion follow from the steps. A report that names the patented feature, calculates a royalty against the entire accused product's revenue, and then asserts the result is "apportioned" because the expert says so is not a report that survives Daubert.

When EMVR applies

The entire market value rule is the exception that allows a damages expert to base the royalty on the full value of the accused product, without apportionment, in a narrowly defined circumstance. LaserDynamics, Inc. v. Quanta Computer, Inc., 694 F.3d 51 (Fed. Cir. 2012), gave the modern articulation. EMVR applies only where the patented feature is "the sole driver of customer demand or substantially creates the value of the component parts." The Federal Circuit was explicit that the EMVR is "appropriately applied only in the limited circumstances" where that demand-driver showing is made.

The evidentiary burden for the showing is significant. Surveys, sales data correlated with the introduction of the patented feature, and customer testimony are the typical proofs. General assertions that the patented feature is "important" or "valued by customers" do not carry it. The Federal Circuit has reversed verdicts where the EMVR was invoked on a record that would have supported only an apportioned analysis.

Provisur Technologies, Inc. v. Weber, Inc. (Fed. Cir. 2024) is a recent application of the same line. The court continued to require that the EMVR be supported by evidence that the patented feature drives demand for the full product, and continued to reject the rule's application where the record supported only contributory rather than driving value.

Comparable-license analysis under EcoFactor

EcoFactor (Fed. Cir. 2025), decided en banc, is now the operative authority on what a damages expert may do with a comparable license. The court held that a per-unit royalty derivation from lump-sum settlement licenses is not admissible where the per-unit conversion is not grounded in what the parties to the license actually agreed to.

The interaction with apportionment is direct. A damages expert who relies on settlement licenses as comparables to derive a per-unit rate, and then applies that rate to a royalty base that has not been apportioned to the patented feature, is failing on both axes. The per-unit derivation has to be supported under EcoFactor, and the royalty base has to be apportioned under Garretson and LaserDynamics. Failure on either axis is independently sufficient for exclusion. Many of the recent partial-exclusion orders tracked on WitnessLens reach both findings.

Common failure modes

Five failure modes recur across the damages-exclusion orders tracked on WitnessLens.

The first is invoking EMVR without the demand-driver showing. The expert assumes the rule's availability and bases the royalty on the full product revenue. The fix is either to develop the demand-driver evidence or to apportion.

The second is the "smallest salable patent-practicing unit" shortcut. The expert identifies an internal component of the accused product, calls it the SSPPU, and applies the rate against the SSPPU's revenue without further apportionment. The Federal Circuit has repeatedly held that the SSPPU is a starting point for apportionment, not a substitute for it. If the patented feature is a sub-component of the SSPPU, further apportionment is required.

The third is reliance on a portfolio license without apportioning across the licensed patents. A portfolio license that covers ten patents and pays a lump-sum cannot, without more, support a royalty for one of the patents at the full portfolio rate. The expert has to apportion across the patents in the portfolio, and the apportionment has to be supported.

The fourth is the post hoc per-unit derivation from a lump-sum settlement that EcoFactor now squarely addresses. This was a tolerated methodology by some district courts before May 2025. It is not now.

The fifth is methodology by ipse dixit. The expert states a percentage apportionment, attributes it to "training and experience" or to a generic Georgia-Pacific weighing, and provides no underlying calculation or data. Courts continue to find this insufficient under Rule 702.

A drafting checklist

A damages report that engages each of the recurring failure modes on its face reduces to a short discipline.

State the apportionment methodology explicitly and identify whether apportionment is on the base, on the rate, or on both. Identify the patented feature and the evidentiary basis for distinguishing its value from the unpatented features. If EMVR is invoked, develop the demand-driver evidence on the record and address why the patented feature is the sole or substantial driver. If comparable licenses are relied on, identify the licenses, identify the per-unit or lump-sum structure of each, and explain how the structure supports the rate conclusion. For settlement licenses, address EcoFactor on the face of the report. For portfolio licenses, apportion across the licensed patents. Avoid stating a percentage or a rate without an underlying calculation tied to the record.

None of this is a higher methodological bar than the Federal Circuit has been requiring for the last decade. What has changed since December 2023 under amended Rule 702, and what has changed again since May 2025 under EcoFactor, is the front-end gatekeeping rigor. The discipline is now an admissibility requirement, not a weight question.

A note on the data

The exclusion-pattern observations and the recurring failure modes described above are drawn from district court Daubert orders on WitnessLens addressing patent damages experts. Case citations are to the Federal Circuit decisions identified by name and year. The order-classification methodology is described on the methodology page.