Most civil cases never reach a jury. They settle, usually at mediation. The only jury most cases ever face is a hypothetical one: the panel both sides picture when they decide what the case is worth. It is the most consequential jury in the file, and almost nobody studies it.
How the number gets set today. On both sides, most mediation numbers rest on two inputs. The first is the published record of jury awards in cases that look roughly like this one, usually pulled from verdict reporting services, what the trade calls verdict reports. The second is instinct: what counsel, together with the claims professional when an insurer is involved, believe a jury will do with these facts. Both are legitimate. Both have a blind spot. Verdict reports describe the small share of cases that did not settle, often because one side misjudged the jury. And instinct, however seasoned, is a read on jurors in general, not on the people this venue would actually summon this year.
The usual structured alternative is a mock trial. Recruit a panel in the venue, present abbreviated plaintiff and defense cases, and watch the jury deliberations. Done well, it is the best evidence available short of a verdict, and it is priced accordingly: a one-day mock commonly runs $30,000 to $50,000 in fees, and multi-panel projects reach $75,000 and higher, as set out in The True Cost of a Jury Consultant. Online mock panels and AI courtroom simulations have brought the price down. But a recruited panel is a dozen or two residents reacting to one presentation of the case, and a simulated panel is not residents at all. Neither gives a structured read of the full population a venue draws its juries from. For most of the docket, the settlement number still gets set with no structured read of the jury at all.
The verdict before the verdict
Our session at the PLDF Annual Meeting this September took its title from this idea. Every mediation is a negotiation over a verdict that will never be rendered. Each side's number is a forecast of how a group of strangers would assign responsibility and damages for what happened. The side with the better-grounded read of those strangers negotiates from a stronger position, whether or not the case ever comes near a courtroom.
That read does not require a mock trial, real or simulated. It requires answering three questions about the jury pool before mediation rather than before trial, using the consumer, occupational, and behavioral data that already describe the people in a venue who are eligible to serve.
1. Who is likely in the pool. Not a census snapshot of the county, but the behavioral makeup of the people a venue tends to seat: their relationship to institutional authority and expertise, whether they see fairness as an individual or a collective responsibility, and how they process risk. Two venues with nearly identical demographics can lean in opposite directions on all three.
2. How they tend to assign responsibility. The same fact lands differently with different jurors. An insurer's documented claims process reads as diligence to one juror and as paperwork built to say no to another. A plaintiff's own choices in the moments before the injury read as understandable caution to one juror and as the real cause to another. In contract and business litigation, a party's reliance on the plain terms of an agreement reads as fairness to one segment and as the stronger party hiding behind fine print to another. Knowing which reading a pool leans toward tells you how hard a damaging fact is likely to hit, how much a sympathetic fact is likely to carry, and which themes are worth building a case around or defending against.
3. What that does to a defensible number. A structured read of the pool does not produce a number. It produces reasons: which facts carry the most risk, or the most leverage, with this pool; where the story is strongest on either side; and where the gap between the two sides' assumptions about the jury is widest. A number built on those reasons, a demand or a range, can be explained to a case-evaluation committee or a claims committee, defended to a client, and adjusted when the facts move.
What that looks like in practice. Take a hypothetical premises liability claim against a national retailer: a customer injured by a falling display, a store incident report filed the same day, and a safety policy the store manager admits was not followed that week. Now place the same file in two counties with nearly identical census profiles. In the first, the pool leans toward trusting institutions and judging people by their individual choices. The incident report reads as a company doing what it should, and the open question becomes the customer's own attention, a pool the defense should lean into and the plaintiff should prepare to work around. In the second, the pool leans skeptical of corporate process and sees safety as a shared obligation. The same report reads as paperwork, and the unfollowed policy becomes the case, a pool that rewards the plaintiff for building around the unfollowed policy and leaves the defense needing a different anchor. Neither read says what a jury would award. Both tell either side which fact will carry the most weight, which theme is worth building or defending, and why the number in the second county should be built differently from the first.
Why this matters now
Insurers are being pushed to decide earlier, and the data gap between the two sides has been closing for years, just not evenly. Plaintiffs' firms have invested in verdict data and venue analytics longer and harder than the defense bar, and it shows in verdicts. But that lead is a head start, not a finish line. Most plaintiff workups still run on the same verdict reports and case-type analytics anyone can buy, not a structured read of how this specific venue's pool leans on the themes this case will actually turn on. Risk advisors are telling insurers to stress-test early claim valuations against current venue data, and to spot serious exposure while there are still options to preserve. Whichever side brings the more specific read of the actual pool negotiates from a stronger position.
Litigation funding changes the calculus for everyone. Third-party funding means a plaintiff no longer has to settle for cash-flow reasons alone, and an EY model presented to casualty actuaries in 2025 put the most likely cost of funders' returns paid by casualty insurers at $13 billion to $18 billion over five years. For insurers, a nuisance offer no longer moves a funded case. For plaintiff firms, the same funding raises the bar on their own number: a funder backing a case wants a demand grounded in something more specific than what juries in this venue tend to award on average, the same gap a structured venue read is built to close. Either way, a number has to be defensible on the merits, and the merits include the jury.
The tail keeps getting wider. Marathon Strategies counted nearly 200 verdicts of $10 million or more against corporate defendants in 2025, up 40.7% from the prior year and a record in its tracking. Most cases will never come close. For an insurer, a number that ignores the jury pool is quietly assuming its case is not one of them. For a plaintiff firm, the same trend is only leverage if the number behind it can be explained as more than a headline, which is exactly what a structured venue read provides.
A read, not a forecast
It is not a verdict prediction, and it is not a dollar figure. That is a deliberate choice. Some tools in this market now advertise accuracy rates for simulated verdicts, and a claims committee, or a firm's own case-evaluation team, should ask what those rates are measured against. No method can tell you what a specific jury will award, and a number built on a claim that it can is harder to defend, not easier. A structured read does not replace the judgment of the claims professional, the plaintiff attorney, or defense counsel, who own the number and the decision. Nor does it replace the mock trial for a matter whose stakes, for either side, justify it, where live deliberations are worth what they cost.
It is what belongs between instinct and a $60,000 research project: a structured, repeatable read of the likely jury pool, applied to the matters that would otherwise get none. The same behavioral framework behind VoirPro's trial work, described in Two Identical Jurors, Opposite Verdicts, goes to work long before a panel is summoned.
For a claims organization, or a plaintiff firm with a concentrated venue practice, the value compounds. Neither mediates one case. An insurer works the same handful of venues through the same panel counsel, matter after matter; a plaintiff firm with a regional docket files in the same courthouses again and again. A venue read built for one matter is already in hand for the next one filed there, and each matter sharpens it. Across a docket, that turns jury intelligence from a cost reserved for the few largest files into a standing view of the venues where the exposure, or the opportunity, actually sits. It also gives leadership on either side something it rarely has: a consistent basis for comparing numbers across adjusters or case teams, panel firms, and regions.
Meet the jury before you mediate
Mediation prep usually asks what the case is worth. The better question is: worth to whom? The jury you will never seat is still the audience both sides are negotiating in front of. A team that has met that audience, even on paper, walks in with a number it can explain. A team that has not is negotiating against a forecast it never examined.
VoirPro provides this read per matter, ahead of mediation, for insurers, claims teams, defense and plaintiff counsel, as part of its basic license and no mock trial required.
The method behind that read is the VoirPro FINDS™ framework. How insurers and claims teams put it to work across a docket is on the Insurance & Risk Professionals page; how defense and plaintiff firms build the same read into their dockets are on the Defense and Plaintiff Counsel pages.