A recurring tension in patent damages law arises when a patented feature is embedded in a product whose market value derives substantially from unpatented components. The patentee wants to recover damages calculated on the full value of the product; the accused infringer wants to limit damages to the value of the patented feature alone. The entire market value rule (EMVR) mediates that tension by permitting recovery on the full product value — but only when the patented feature is shown to drive consumer demand for the product as a whole. After the Federal Circuit’s decisions in Uniloc USA, Inc. v. Microsoft Corp. and LaserDynamics, Inc. v. Quanta Computer, Inc., establishing that predicate has become the central battleground in complex patent damages cases, and survey evidence has emerged as both the most direct proof and the most contested methodology for establishing it.
The Demand-Driver Predicate
The doctrinal foundation of the EMVR traces to cases in which the entire value of an accused product was attributable to a single patented component. In Aro Manufacturing Co. v. Convertible Top Replacement Co., 377 U.S. 476 (1964), the Supreme Court held that the measure of damages must reflect the value of the infringing features rather than the total product value. The Federal Circuit refined this into an affirmative rule: to use the entire market value as the royalty base, the patent holder must demonstrate that the patented feature is the basis for customer demand. Panduit Corp. v. Stahlin Bros. Fibre Works, Inc., 575 F.2d 1152, 1157 (6th Cir. 1978), articulated the lost profits corollary — that demand for the patented product, not merely for the product category, must be shown. The EMVR applies that logic to reasonable royalty calculations by requiring a showing that would justify treating the entire product as the proper royalty base.
What “drives consumer demand” means in practice has never been fully resolved. At minimum it requires something more than showing that the patented feature is a feature of the accused product. The patented component must be, in some meaningful sense, the reason customers select the product over alternatives. Whether that requires proof that the patented feature is the primary reason, the but-for reason, or merely a substantial reason has divided courts and commentators.
Origins and Doctrinal Development of the EMVR
The EMVR developed primarily in cases involving combination products where the patented element was physically separable from the rest of the accused article. Early Federal Circuit cases — notably Rite-Hite Corp. v. Kelley Co., 56 F.3d 1538 (Fed. Cir. 1995) (en banc) — extended the rule to functionally related products, holding that lost profits on unpatented components could be recovered when those components were functionally inseparable from the patented item. The court in Rite-Hite demanded that the patented and unpatented components “function together as a unit,” and that the unpatented components be sold “as a result of the demand for the patented apparatus.” This was a comparatively permissive standard.
The Federal Circuit tightened the rule significantly through the 2000s, particularly as cases involving standard-essential patents and consumer electronics reached the court with royalty bases drawn from enormously valuable end products. In Cornell University v. Hewlett-Packard Co., 609 F. Supp. 2d 279, 286 (N.D.N.Y. 2009), Judge Rader, sitting by designation, refused to apply the EMVR to an instruction reorder buffer patent claimed against an entire server, holding that the smallest salable patent-practicing unit — not the entire product — was the appropriate royalty base absent proof that the patented feature drove demand for the complete system. The “smallest salable patent-practicing unit” (SSPPU) concept became a counterweight to the EMVR, giving accused infringers an alternative framing for royalty base disputes.
The Post-Uniloc Tightening
The Federal Circuit’s 2011 decision in Uniloc USA, Inc. v. Microsoft Corp., 632 F.3d 1292 (Fed. Cir. 2011), represents the most emphatic judicial statement of the rule’s limits. Uniloc had obtained a jury verdict of approximately $388 million for infringement of a software activation patent, with the royalty base calculated on the entire revenue from Microsoft Office and Windows. The Federal Circuit vacated the damages award. The court held that the use of the entire market value was legally impermissible because Uniloc had not established that the patented activation feature was the basis for demand for Office or Windows. More significantly, the court held that reference to the total revenue figure — even as an anchoring point for a subsequent apportionment calculation — was reversible error, because such references “cannot help but skew the damages horizon for the jury.”
LaserDynamics, Inc. v. Quanta Computer, Inc., 694 F.3d 51 (Fed. Cir. 2012), extended that reasoning. LaserDynamics held a patent on a method for disc-type identification in optical disc drives. It sought damages on the entire laptop computer value. The Federal Circuit rejected the entire-computer base, holding that LaserDynamics had failed to show that the disc-identification feature drove consumer demand for laptops. Crucially, the court rejected the argument that showing the feature was “important” or “valuable” was sufficient; demand-driver status requires that the feature be the basis for consumer choice, not merely a welcome attribute. The SSPPU framing prevailed: the royalty base should be the optical disc drive, not the laptop.
By 2014, when Ericsson, Inc. v. D-Link Systems, Inc., 773 F.3d 1201 (Fed. Cir. 2014), reached the Federal Circuit, the court was wrestling with the specific complications that arise with standard-essential patents. Ericsson declined to impose a categorical rule requiring the SSPPU as the royalty base in FRAND cases, acknowledging that the smallest salable unit may not itself reflect the true incremental value of the standard-essential feature. This created an opening for patentees: if the SSPPU is itself a complex multi-component article that incorporates value independent of the patented feature, apportionment still applies, and the question of how to perform that apportionment remains contested.
Survey Evidence as Proof of Consumer Demand
Against this doctrinal background, patentees in post-Uniloc litigation have increasingly turned to consumer surveys as the most direct way to establish the demand-driver predicate. The intuition is straightforward: if courts require proof that consumers buy the product because of the patented feature, the most logical evidence is survey data in which consumers describe their purchasing motivations. Courts have accepted this logic in principle. Survey evidence of consumer demand is admissible — the question is whether the specific survey adequately captures the phenomenon the doctrine requires.
Traditional consumer preference surveys, however, face a fundamental methodological mismatch. A survey question asking “how important was Feature X to your decision to purchase Product Y?” measures subjective importance, not demand causation. Importance and causation diverge wherever the feature, though rated highly, would not have changed the consumer’s choice absent its presence. A laptop buyer who rates the disc-identification function as “very important” may nonetheless have purchased the same laptop if the function were absent, because the function is near-universal in the laptop market. Importance questions, moreover, are susceptible to social desirability bias and hypothetical response inflation: respondents rate features as more important in surveys than their actual behavior reflects.
Methodological Challenges: Conjoint Analysis and Willingness-to-Pay
Conjoint analysis — a family of survey methodologies borrowed from market research and adapted for litigation contexts — has emerged as the primary vehicle for patentee-side experts seeking to quantify the incremental value of a patented feature. In a conjoint survey, respondents choose among product configurations that vary across multiple attributes simultaneously. Statistical analysis of those choices allows the researcher to estimate the utility weight associated with each attribute, which can then be converted into a willingness-to-pay estimate or a market-share simulation. Because conjoint forces respondents to make trade-offs rather than merely rate features in isolation, it is less susceptible to importance inflation than direct importance questions.
Courts have nonetheless excluded conjoint surveys in patent damages cases on multiple grounds. First, a survey that presents the patented feature as a discrete attribute may mischaracterize how consumers actually perceive the product if the feature is invisible or bundled with others. Second, conjoint surveys conducted on a population of current owners may over-sample respondents whose purchasing decisions were made in a market that included the patented feature as a standard offering, inflating demand-driver estimates. Third, the conversion of conjoint utility weights into a royalty rate can introduce additional methodological steps — market simulation assumptions, competitive response assumptions, demand-elasticity assumptions — that are independently contestable.
In Commonwealth Scientific & Industrial Research Organisation v. Cisco Systems, 809 F.3d 1295 (Fed. Cir. 2015), the Federal Circuit signaled that patentees need not always reduce the royalty base to the SSPPU if they can demonstrate through reliable evidence that the incremental value of the patented feature as a share of the full product is properly calculated. This is a demand-driver question framed as an apportionment question — and it is precisely the question that well-conducted conjoint analysis is designed to answer, provided the methodology survives Daubert scrutiny.
Implications for Patent Damages Reform
The survey-evidence debate exposes a deeper structural tension in the EMVR doctrine. Courts have tightened the demand-driver predicate in response to concerns about damages windfalls in complex-product litigation, particularly in standard-essential patent disputes. That tightening is defensible on policy grounds: a royalty base inflated by the entire value of an iPhone or a laptop bears no relationship to the incremental contribution of any single patented feature, and damages divorced from incremental contribution are both economically inaccurate and strategically manipulable.
At the same time, a doctrinally rigid preference for the SSPPU creates its own distortions. The SSPPU may itself be a commodity whose market price reflects none of the demand-generating value that the patented feature contributes to the end product. In that scenario, the SSPPU royalty base produces an artificially depressed royalty that undercompensates the patentee and sends the wrong signal to future inventors. Surveys — and specifically conjoint surveys — can bridge the gap by directly measuring the value contribution of the patented feature within the context of the product as consumers actually experience it. Their admissibility should turn on the rigor of the methodology applied to the specific facts, not on a categorical rule that disfavors all survey evidence in EMVR litigation.
The Federal Circuit’s movement toward a case-by-case apportionment framework, visible in Ericsson and foreshadowed in CSIRO v. Cisco, is the more promising path. It demands that patentees prove incremental value through reliable evidence, imposes discipline on survey methodology through Daubert, and permits the royalty base to reflect the actual economics of the accused product rather than a legal proxy. Whether courts will develop a consistent methodology for evaluating conjoint surveys under this framework remains to be seen — but the demand-driver predicate, interpreted as a question about value contribution rather than consumer psychology, can be answered with the tools empirical economics has already developed.
Related articles: Shifting the Burden: Proving Infringement and Damages · Software Patents and Extraterritorial Protection · IP Privateering
Frequently Asked Questions
What is the entire market value rule in patent law?
The entire market value rule (EMVR) permits a patentee to calculate reasonable royalty or lost profits damages based on the total revenue of a multi-component product, rather than only the revenue attributable to the patented component. To invoke the rule, the patentee must demonstrate that the patented feature drives consumer demand for the entire product — that purchasers buy the product because of the patented feature. After Uniloc USA v. Microsoft (Fed. Cir. 2011) and LaserDynamics v. Quanta (Fed. Cir. 2012), courts have applied this predicate rigorously, and the smallest salable patent-practicing unit has emerged as the default royalty base absent proof of demand-driving status.
How do courts determine whether a patented feature drives consumer demand?
Courts assess the demand-driver predicate through a fact-intensive inquiry examining marketing materials, internal documents, and expert testimony about purchasing motivation. Consumer surveys are sometimes offered as direct evidence, but courts have been skeptical of surveys using generic willingness-to-pay questions that do not isolate the patented feature’s incremental contribution. Conjoint analysis — which forces respondents to choose among product configurations varying across multiple attributes — has gained traction as a more rigorous methodology, though specific surveys have been excluded under Daubert where underlying assumptions were found unreliable.
What role does survey evidence play in patent damages calculations?
Survey evidence serves two functions in patent damages. First, it can establish the demand-driver predicate required to invoke the entire market value rule by showing that the patented feature drove consumer purchasing decisions. Second, survey data — particularly conjoint analysis — can estimate the incremental value of the patented feature as a share of total product value, providing a basis for either the royalty base or the royalty rate. Both uses have been accepted in principle; admissibility turns on whether the specific methodology is reliable under the Daubert standard and Federal Rule of Evidence 702.