What is apportionment trying to fix?
The mismatch between a claim to one feature and a product containing thousands of them. A modern device embodies an enormous number of separately patentable contributions, and a patent covering one of them is not a patent on the device. Apportionment is the requirement that damages reflect the value attributable to the patented contribution rather than to everything around it.
Economists perform the final allocation, but the material they need is technical: which components actually practice the claim, what the accused functionality does for the product and for the customer, and whether the patented feature drives purchasing at all. That is why apportionment is an evidence problem long before it is a calculation.
When can whole-product revenue be used?
Where the evidence shows the patented feature drives demand for the entire product. That is a genuine standard rather than a formality — not that the feature is important, useful, or one selling point among several, but that it is the reason customers buy the product.
For a component inside a complex product, that bar is high and rarely cleared. Where it cannot be met, the analysis works from a narrower base and explains the choice. Where a party uses whole-product revenue anyway, it faces both a challenge to the base and the separate argument that the jury was improperly anchored to a large number that was never a legitimate starting point.
What if the accused feature is not sold separately?
That is the common case in integrated products, and it does not remove the apportionment requirement — it means the apportionment has to be built from evidence other than a component price. This is where teams frequently give up too early.
Engineering cost allocation, development effort, usage and telemetry data, feature-level demand research, and the cost and availability of non-infringing alternatives all contribute. The analysis is harder and entirely doable. What is not doable is treating the absence of a component price as permission to use the whole product, which is the inference courts have specifically rejected.
Does the royalty base need a causal connection to the accused functionality?
Yes, and this has become the specific ground on which bases are attacked. The question is why the chosen metric moves with the accused functionality rather than merely being available in the company’s financial reporting. A base needs a demonstrated link, not a convenient one.
The 2026 decision reversing an exclusion turned on exactly this: a usage-based metric was held to establish a sufficient causal connection to the accused functionality, so the expert should have been permitted to rely on it. The lesson runs in both directions — a base tied to actual use of the accused feature is defensible, and a base tied to total revenue because that was the number to hand is not.
How much does conjoint analysis actually help?
It can be genuinely valuable, and it is a target. Well-designed conjoint studies isolate what customers will pay for a specific feature, which is directly responsive to the apportionment question and hard to answer any other way for consumer products.
Poorly designed ones are attacked on their design rather than their conclusions — leading attribute descriptions, unrealistic feature bundles, unrepresentative respondents, prices outside any plausible range. A study that does not survive that attack is worse than no study at all, because it becomes the story of the damages case. If conjoint is going to be used, its design needs to anticipate its own Daubert motion from the first draft of the questionnaire.
Does apportionment apply in trade secret cases?
The doctrine developed in patent law, and the underlying logic travels even though the framing differs. Where a defendant’s product or profit stream reflects far more than the misappropriated information, tying recovery to the contribution of what was actually taken is both a live legal dispute and a matter of basic persuasion in front of a fact-finder.
The measures available in trade secret cases differ from the patent measures, and which of them a claim permits is a question of law. What a measure yields belongs to our Economic Damages Institute, which covers the trade secret measures and the royalty machinery properly. This Institute covers whether the evidence underneath will hold.