Dividing Work With the Damages Expert

The rule of thumb, stated plainly

Where search traffic is the injury, two experts are usually required and the line between them is not a matter of taste. The defensible split: the search expert opines on causation and the traffic or visibility counterfactual, and stops. The financial expert takes that traffic delta as a stated assumption and converts it into revenue and incremental profit.

The search expert establishes:

• what changed technically, and on what date;
• whether the change is attributable to the conduct rather than to an algorithm update, seasonality, competition, or the party's own actions;
• the counterfactual traffic or visibility baseline, with confounds controlled and the residual stated;
• the provenance, integrity, and completeness of the underlying data, which also supplies the foundation for the exhibits;
• the reliability and limits of each data source.

The financial expert establishes:

• revenue per session, or the value of a converted lead;
• incremental margin and the costs avoided on the sales not made;
• the conversion of the traffic delta into a revenue delta;
• mitigation, present-value discounting, and any growth assumptions;
• the damages figure and its sensitivity to each assumption.

Neither expert should be able to produce the other's number without the other's report. That is the test of whether the line was drawn correctly.

Why the search expert stops at traffic

The reason is not modesty. The financial half of the work has established methods a technical expert is not using, and courts notice.

In Zenith Electronics Corp. v. WH-TV Broadcasting Corp., 395 F.3d 416 (7th Cir. 2005), the Seventh Circuit affirmed exclusion of a damages expert and rejected the substitution of the company's internal projections. The language is worth having exactly: "A witness who invokes 'my expertise' rather than analytic strategies widely used by specialists is not an expert as Rule 702 defines that term." And: "Reliable inferences depend on more than say-so, whether the person doing the saying is a corporate manager or a putative expert." The court pointed to multivariate regression as what was missing. The opinion is reported in full.

Now picture the search expert who has done careful cohort work, established the counterfactual traffic series, and then multiplies lost sessions by a conversion rate and an average order value. That step is a financial opinion produced by say-so. It is the one paragraph an opposing damages expert can dismantle without knowing anything about search, and a motion aimed at it reaches the technical work too.

The companion warning is LifeWise Master Funding v. Telebank, 374 F.3d 917 (10th Cir. 2004), affirming exclusion of an "S-curve" growth model that "was not in regular usage for predicting future profits, was not peer reviewed, … and is capable of manipulation to achieve virtually any desired result." Two details are often reported loosely and matter: the witness was not a retained expert but LifeWise's own chief executive, and he was separately held unqualified, having never used the methodology and having no training in damages analysis. Growth projections in this field are S-curves, routinely drawn by people with no damages training.

Why the financial expert cannot simply assume the traffic loss

The error runs the other way too, and it is at least as common. A damages model opening with "the site lost 40% of its organic traffic because of the defendant" and proceeding to monetize that figure has an unsupported input. Rule 702(b) requires testimony based on sufficient facts or data, and an assumption is not data unless someone competent established it.

Two propositions hide in that sentence, and they need different witnesses:

• "Organic traffic fell 40%" — a measurement, with provenance questions. In what unit, from which source, over which date range, against which baseline?
• "…because of the defendant" — an opinion about causation, requiring the confounds excluded and the mechanism identified.

A financial expert can take both as assumptions, but only if a qualified witness is offering them and the report says so. If the search opinion is excluded, an assumption-driven damages model collapses with it.

The related failure is subtler: a financial expert who takes the traffic input from a third-party estimator supplied by the client rather than from an analysis of first-party data. Those are modeled estimates from proprietary panels, and they will not carry a damages model on their own.

What the search expert should actually hand over

The handoff should be a document, not a conversation. What the damages expert needs:

1. A dated counterfactual series, not a single number. Monthly or weekly, by cohort, actual against counterfactual across the damage period.
2. The segmentation stated precisely. Which URLs, which query set, which channel definition, and which pages were excluded and why.
3. The unit named without ambiguity. Search Console clicks are not analytics sessions and neither is an order. A model that silently swaps units produces an unreconstructable number.
4. The confound controls applied, and the residual left unexplained. If a dated core update accounts for part of the movement, the series should show it separated out.
5. A range or sensitivity, not a point estimate. The financial model can carry uncertainty forward if it is given it, and cannot invent it later.
6. The source files and the date each was pulled.

The working test I apply: the financial expert should be able to rebuild the input from the exhibits without asking me a question. A question that has to be asked becomes a communication, and communications routed through counsel raise disclosure questions a well-drafted exhibit avoids.

Interlocking without overlapping

The two opinions have to fit together without covering the same ground, and each expert must be able to say which assumptions came from the other. The practical device is an assumptions ledger in each report: every input taken from another witness, with its source, its date, and the person who supplied it. It costs a page and answers the first ten minutes of a deposition.

The overlaps to avoid recur:

• Two experts opining on the same quantity in different numbers. The traffic figure in the damages model must equal the figure in the search report. Any discrepancy becomes a demonstrative for the other side.
• Mismatched date ranges. A damage period running from the conduct date in one report and the complaint date in the other is a gift.
• Inconsistent channel definitions. If the traffic exhibit counts a channel one way and the revenue model another, the reconciliation happens on cross.
• Double counting of paid search. Advertising bought to replace lost organic traffic is either a mitigation cost or an avoided cost, and it must be treated once, by the expert who owns it.

The gaps matter as much: where each expert assumes the other addressed mitigation, nobody has, and the model claims revenue as profit.

The disclosure dimension

The division of labor is also a disclosure question. FRCP 26(a)(2)(B) requires a signed report containing a complete statement of all opinions and their basis and reasons, the facts or data considered, the exhibits, qualifications including publications from the previous ten years, a list of cases in which the witness testified in the previous four years, and a statement of compensation.

Rule 26(b)(4)(B) protects drafts of any required report regardless of the form recorded. Rule 26(b)(4)(C) protects attorney communications with a report-writing expert, except to the extent they (i) relate to compensation, (ii) identify facts or data the attorney provided and the expert considered, or (iii) identify assumptions the attorney provided and the expert relied on.

That third carve-out is the one this page runs into. Where the traffic counterfactual passes to the financial expert through counsel and is relied on, the rule text points toward its being discoverable. The practice follows from the rule rather than from clever drafting: state the assumption in the report, attribute it to the other expert by name and date, and keep the transfer inside the served documents rather than an email chain.

Timing follows the same logic. Absent stipulation or court order, expert disclosures are due at least 90 days before trial, and rebuttals within 30 days after the disclosure being rebutted. The financial model cannot be finished until the traffic counterfactual is, so the technical work must complete first even though both reports are served together.

What sits in the gap, and who takes it

Several inputs live on the boundary. Assigning them deliberately prevents the overlap and the gap alike.

Conversion rate. The search expert can establish what the analytics record shows about how each cohort converted, which is a provenance and segmentation question. Deciding which rate to apply to the counterfactual traffic, and whether to hold it constant, belongs to the financial expert.

Mitigation. What recovery was achievable, by what means, and on what timeline is a search question, and one I am asked to answer for either side. Pricing that recovery, and deciding whether the party's efforts were reasonable, is not.

Paid search substitution. The search expert can establish the cost per click and impression share available for the affected query set. Whether that spending is an avoided cost, a mitigation cost, or a separate head of damages is not a search question.

Rebuttal work. An expert rebutting a damages model can attack the traffic input — its unit, its baseline, its uncontrolled confounds — without opining on the financial model built on it, which keeps the rebuttal inside the witness's own discipline.

The failure modes worth watching for

These are the patterns I see when the division was never explicitly set, in roughly the order of how much damage they do.

• One expert doing both halves. A technical witness computing the dollars, or a financial witness reasoning about rankings. Either produces one motion that reaches the whole opinion.
• A financial model whose traffic input came from a third-party estimator nobody validated against first-party data.
• Two reports with different damage periods, channel definitions, or units.
• A point estimate handed across with no range, producing a damages figure whose apparent precision cannot be defended.
• An orphan assumption — the financial expert believes it came from the search expert, who believes it came from counsel, and the question is first asked in a deposition.
• Nobody addressing avoided costs, so the model treats lost revenue as lost profit.
• A claimant's in-house witness supplying the growth curve, which is the fact pattern in which an S-curve model drawn by an untrained witness was excluded.

Where both kinds of evidence are material, counsel settles this allocation at retention. It is much harder to unwind after two reports have been served.

Frequently Asked Questions

Published 2026-08-11