How to Run a Mock Jury (Without a $25,000 Budget)

Why the traditional model prices out most cases

A conventional mock jury exercise involves recruiting a demographically matched panel, renting a facility, hiring a moderator, and running the group through your case live — typically for one deliberation, once, on one day. The fixed costs (recruiting fees, facility rental, moderator time, travel) are largely the same whether the case is worth $150,000 or $15 million, which is why the traditional format tends to get reserved for the largest matters on the docket.

That math leaves a large tier of civil and personal-injury cases — the ones that settle on the courthouse steps, or don't settle at all because nobody tested the theme first — without any structured read on how a neutral panel receives the story. Attorneys either try the case cold or rely on gut instinct from depositions and prior verdicts, which is useful but not the same as watching independent people argue about your facts.

An AI-run mock jury doesn't replace the value of watching real humans in a one-way mirror room. What it does is make the exercise cheap and fast enough to run on ordinary cases, and to run more than once — before and after a key deposition, before and after you rewrite your theme of the case.

Step 1: Build a case file that teaches the panel something

The single biggest driver of a useful mock jury — traditional or simulated — is the quality of the case file you feed it. A thin file (a demand letter and a couple of exhibits) produces a thin, generic reaction. A file that mirrors what a real jury would actually receive — key liability facts, damages evidence, the strongest defense arguments, and the specific themes you intend to argue at trial — produces a reaction you can act on.

In practice that means including both sides fairly: your strongest liability theory and the defense's best counter, your damages narrative and the defense's likely attack on it (prior injuries, gaps in treatment, credibility issues), and any facts a jury would find surprising or hard to accept. Simulated jurors, like real ones, are only as sharp as the record they're given. If you only feed it your best evidence, you'll get an optimistic verdict that tells you nothing about your vulnerabilities.

This is also the moment to decide what you're actually testing. A case file built to test liability theory looks different from one built to test a damages number or a specific witness's credibility. Narrow the question before you seed the file, and the deliberation will actually answer it.

Step 2: Size the panel to your venue

Jury size is not a cosmetic detail — a 6-person panel and a 12-person panel don't just differ in scale, they differ in how deliberation unfolds. Smaller panels reach consensus faster and are more sensitive to a single strong personality; larger panels surface more competing readings of the facts and take longer to converge, which is closer to what many civil trial lawyers actually see in state court. Testing your case against a panel sized to your actual venue — 6, 8, or 12, depending on jurisdiction and case type — matters more than testing against a generic default.

JuryFish seats a panel sized to your venue and runs each juror as an independent agent: each one reads the case file, forms an initial view, and then reads what the other jurors have argued before deciding whether to move. That mirrors the mechanic of real deliberation — persuasion happens juror to juror, not just from the evidence to each person in isolation — across multiple rounds rather than a single up-or-down vote.

Running the same case file against both a 6-person and a 12-person panel, where your venue allows either, is a cheap way to see whether your theme's strength depends on panel size — a useful thing to know before you're picking an actual jury.

Step 3: Read the deliberation, not just the verdict line

The least useful part of any mock jury report is the topline number. A single verdict figure or split tells you almost nothing about why the panel landed there, and it's the why that you can actually use in voir dire, opening, and closing. The valuable part of the output is the deliberation itself: which juror moved, on what argument, in which round, and which arguments got raised and then dropped because nobody could answer them.

A report that shows a juror starting skeptical of causation and then moving after another juror restates the treating physician's testimony in plain language tells you something concrete — that argument works, and it works from a lay mouth, not just from counsel. A report that shows three jurors independently raising the same doubt about a damages number, and nobody in the room rebutting it, tells you exactly where the defense will go and exactly what you haven't answered yet.

JuryFish's report packages this into case vulnerabilities and voir-dire targeting alongside the directional damages range, precisely because the deliberation detail is where the actionable insight lives. To be clear about what this is and isn't: JuryFish is an argument-testing tool that simulates how a plausible panel might react to your themes. It is a directional aid — not legal advice and not a prediction of any real jury or verdict — and the value is in what it surfaces about your case, not in treating the number as a forecast.

Step 4: Act on the report before a real jury sees the case

A mock jury that doesn't change anything about your trial preparation wasn't worth running, cheap or expensive. The practical move is to take the specific vulnerabilities the panel surfaced and build them into your case in the places they belong: fold the strongest juror-generated rebuttals into your closing, use the sticking points as a checklist for voir dire questions designed to find and strike jurors likely to hold the same view, and, where the damages range came in lower than expected, figure out whether that's a presentation problem or a genuine exposure problem before you're in a settlement conference.

Because the cost structure runs on usage credits — JuryFish calls them fish food, with refill packs starting at $49 and no per-case or per-seat fee — it's realistic to run the case file more than once: once early to find the weak points, and again after you've rewritten the opening or added a witness, to see whether the rewrite actually fixed what the panel flagged. That iterative loop is difficult to justify at $10,000 to $25,000 a run; it's straightforward at the cost of a refill pack.

The report is a rehearsal, not a verdict. Treat every output as a hypothesis about how your themes land with a particular kind of panel, test it against your own read of the judge, the venue, and opposing counsel's likely strategy, and update the case file for the next round rather than taking any single run as the final word.

When you still need the traditional focus group

None of this replaces a full traditional mock jury for the cases that warrant one — high-exposure matters where a client or carrier needs the weight of an in-person panel, live moderator probing, and a facilitator experienced in reading body language and group dynamics that a written record can't fully capture. For a nine-figure exposure case heading to trial in three months, the $10,000 to $25,000 spend and the multi-week timeline are often still the right call.

Where the simulated approach earns its keep is everywhere below that threshold, and as a first pass even above it: testing themes before you commit to a traditional panel's limited number of runs, iterating on a damages narrative across depositions, and giving cases that would otherwise never get tested at all a structured read before they go to a real jury.

Frequently asked