Consulting Versus Testifying Experts

Two roles, one rule, very different exposure

The distinction the federal rules draw is between an expert retained in anticipation of litigation who is expected to testify and one who is not. Rule 26(b)(4)(D) provides that ordinarily a party may not, by interrogatories or deposition, discover facts known or opinions held by an expert who has been retained or specially employed by another party in anticipation of litigation or to prepare for trial and who is not expected to be called as a witness at trial — except as provided in Rule 35(b), or on a showing of exceptional circumstances under which it is impracticable for the party to obtain facts or opinions on the same subject by other means. The text is on Cornell's copy of Rule 26.

Behind it sits Rule 26(b)(3), the general work-product doctrine, protecting documents and tangible things prepared in anticipation of litigation by or for a party or its representative, with near-absolute protection for opinion work product.

The practical difference is large. A testifying expert's opinions, basis and reasons, facts or data considered, qualifications, publication and testimony lists, and compensation are disclosed under Rule 26(a)(2)(B), and the expert sits for a deposition. A consulting expert is ordinarily not subject to interrogatories or deposition at all. The same person doing the same work is either a disclosed witness or an undisclosed advisor, depending on a designation decision often made in the first week by someone who does not yet know what the data shows.

What “exceptional circumstances” means, and where search matters sit

The standard on the face of the rule is impracticability: it must be impracticable for the party seeking discovery to obtain facts or opinions on the same subject by other means. Not inconvenient. Not expensive. Not merely duplicative of effort.

The classic occasion is evidence that no longer exists in a form the other side can examine — something tested destructively, a condition observed once, a scene altered before anyone else could inspect it.

A website is an unusually good candidate for that argument, because a website is a condition rather than an object, it changes continuously, and it is not preserved by default. A site rebuilt on a new platform cannot be re-examined as it was. A consultant who crawled the old site before the rebuild holds observations nobody can now reproduce, and the same is true of an analyst who exported performance data before a rolling retention window closed over it.

State the limit honestly. I have not located a reported decision applying the exceptional-circumstances standard to a website since rebuilt or to search data since expired. The argument follows from the rule text and from how the technology behaves; it is not a holding I can point you to. The reasoning also runs both ways: if the consultant's early crawl is the only surviving record of the disputed condition, that is what makes the other side's motion plausible.

The conversion runs one way

A consulting expert can be designated as testifying at any point before the disclosure deadline. A testifying expert who has been designated and whose report has been served cannot be returned to consulting status and withdrawn from discovery. You cannot un-ring that bell, even if you later conclude the person would have been more useful as an advisor.

That asymmetry should drive the sequencing, and the reasoning is arithmetic rather than tactics. You can always promote; you can never demote. Starting as a consultant and converting later costs a designation and a report. Starting as a testifying expert and wishing you had not costs a deposition of a witness whose opinion you did not want, taken by counsel who now knows it exists.

Note what a formal withdrawal does not cure. Once a report has been served, its contents are known to the other side, and the opinions in it are on the table whatever becomes of the witness. This is one of the few genuinely irreversible decisions in a case.

What the early work in a search matter actually is

Before anyone can sensibly decide whether there is a case, someone has to answer questions that are entirely technical and that no lawyer can answer from the file:

• What does the site look like now. Crawl it and record status codes, canonical directives — the rel="canonical" element telling a search engine which of several similar URLs it should treat as the one to index — noindex directives, redirect chains, and internal link structure.
• What data still exists. Google Search Console holds sixteen months of performance data on a rolling basis. Analytics retention depends on how the property was configured. Server logs are commonly rotated on a cycle of weeks. The first question is not what the data shows; it is what survives to be looked at.
• What changed, and when. Reconstruct the timeline from archive snapshots, third-party index records with their first-seen and last-seen dates, the client's own deployment history, and whatever log data is still on disk.
• Does the data support the theory at all. The client's account of the loss is a hypothesis, not a finding. Frequently the numbers say something else: the decline began before the conduct complained of, or tracks a known algorithm update, or is an artifact of a tracking tag that stopped firing, or repeats a seasonal pattern visible in each of the prior three years.

That fourth answer is the one worth paying for early, and the one you least want to receive after a testifying designation has been made and a report is due in three weeks.

Why consulting-first is the right default

Retaining the technical expert initially as a consultant, evaluating the merits, and designating as testifying only once the opinion is known to be supportable is better practice, for reasons unrelated to gamesmanship. The merits question gets answered before the answer is discoverable. Counsel and the analyst can work through a developing theory without every abandoned line of inquiry becoming an exhibit. If the analysis comes back unfavorable, the party learns it while it can still act.

The objection is that this looks like a way to hide a bad opinion. It is not. A consultant who reports that the data does not support the claim has done the most useful thing anyone can do in the first month of a matter, and the alternative — finding out at a deposition — is worse for everyone including the client. An expert whose opinion was formed before designation is also a more credible witness, because the sequence shows the analysis preceded the assignment.

One caution. Nothing about the consulting phase excuses a lower standard. Consulting work frequently becomes the foundation of a report, and the crawl you ran in month one is the crawl you will be cross-examined about in month fourteen. Documenting it as though it will be disclosed is the only sensible assumption.

What changes at the moment of designation

Designation switches on a defined set of obligations, and it is worth having them in front of you before the decision rather than after:

• A signed written report under Rule 26(a)(2)(B), containing all six required items.
• Production of the facts or data considered — considered, not merely relied upon.
• The ten-year publication list and the four-year testimony list.
• A statement of the compensation to be paid for the study and testimony.
• A deposition, which under Rule 26(b)(4)(A) may be taken only after the report is provided.
• Draft protection under Rule 26(b)(4)(B), and communications protection under 26(b)(4)(C) subject to its three carve-outs.
• Fee-shifting under Rule 26(b)(4)(E), under which the party seeking the discovery ordinarily pays the expert's reasonable fee for time spent responding.

One wrinkle is particular to technical matters. Where the consultant and the testifying expert are the same person, the consulting-phase work log is the record of the facts and data considered, and it is going to be produced. Where they are different people, counsel faces a distinct decision about what the testifying expert is given, because material handed to an expert by counsel is a Rule 26(b)(4)(C)(ii) item regardless of who originally generated it.

Two categories that are not consulting experts

Two kinds of technical people get treated as consulting experts and are not, and the mistake is expensive because it is usually discovered after the documents have been created.

The party's own employees. An in-house SEO manager, a developer, or a marketing director who works on the matter is a fact witness who happens to be technical. Rule 26(b)(4)(D) speaks to an expert “retained or specially employed” in anticipation of litigation; an employee performing ordinary duties is neither, their files are the party's files, and their knowledge is ordinary discovery. Where such a person is later offered to give opinion testimony, the disclosure route is Rule 26(a)(2)(C) — subject matter and a summary of the facts and opinions — rather than a report. Where a particular employee's role actually sits is a question for counsel, but the assumption that internal analysis is protected because it was done “for the lawyers” is a bad one to rely on.

The vendor whose work is in dispute. The agency that performed the search work, ran the migration, or built the links is a party or a witness. It cannot be the neutral evaluator of its own work, and its internal files, tickets, and messages are ordinary discovery targets.

How this decision is usually made badly

The recurring failures, in rough order of frequency:

1. Designating a testifying expert at retention, before anyone knows what the data shows, because the scheduling order has a date on it and the date feels like an instruction.
2. Having the client's incumbent agency perform the analysis, then discovering that its files, drafts, and messages are ordinary discovery — and that it has an interest in the answer.
3. Treating the consulting phase as informal, keeping no work log, then having to reconstruct months of data pulls when the same person is designated.
4. Assuming that withdrawing a designation withdraws a served report.
5. Bringing in a technical expert so late that the preservation questions have answered themselves: the sixteen-month performance window has rolled, the logs have rotated, and the site has been rebuilt.

The fifth is the one I encounter most often and the only one nobody can repair. The others cost money or position. That one costs evidence, and once it is gone the best expert in the field is reduced to reasoning from substitutes.

Frequently Asked Questions

Published 2026-08-11