Make discovery earn its keep

Effective discovery is measured not by volume, but by what can be used at trial

2026 October

“Facts are stubborn things.” John Adams said that more than two centuries ago, and trial lawyers still live by that truth. Every case ultimately asks the factfinder to reach conclusions: the defendant was negligent, the company had notice, the collision caused the injury, or the defense expert’s opinion is unreliable. But those conclusions must rest on specific, admissible facts. Discovery is where counsel builds that foundation – one fact, one admission, and one piece of evidence at a time – so the ultimate conclusions can be persuasively established at trial.

That distinction becomes clearest when preparing a witness examination. A trial outline stating “Defendant was negligent” offers little help on cross. The facts beneath that conclusion do: the defendant saw traffic slowing ahead, the view was unobstructed, the roadway was dry, the brakes worked, nothing prevented the defendant from braking, and the defendant did not brake in time. With those facts, counsel has a cross-examination.

Focused discovery for a small law firm

This article is based on a simple premise: stop chasing conclusions and start building facts. It proceeds in four parts. First, it works backward from the verdict to identify the facts that matter. Second, it shows how written discovery and depositions can secure those facts. Third, it addresses expert foundations, admissibility, and trial-file organization. Finally, it considers technology and resource allocation for solo and small-firm lawyers.

Here is a simple discipline to remember: one question, one fact, one admission.

Begin with the verdict and work backward

Focused discovery begins with the eventual trial. Before serving written discovery or taking an important deposition, identify the conclusion that the factfinder must reach, then work backward to the individual facts needed to support it.

Suppose the proposition is that a driver had a reasonable opportunity to avoid a collision. That is the conclusion, not the proof. The proof may include evidence that the driver saw traffic slowing, had an unobstructed view, was traveling on a dry roadway, had functioning brakes, had time to react, and faced nothing that prevented braking. Those facts shape the discovery plan, then the examination outline, and eventually the closing argument.

The same method applies in other cases. In a premises liability case, “the store had notice” is a conclusion. The supporting facts may be that an employee passed through the area shortly before the incident, inspections were required every 30 minutes, no inspection was documented during the preceding hour, and warning cones were available but never placed.

In an injury case, “the event caused the condition” is a conclusion. The proof may include the absence of similar pre-incident complaints, the onset of symptoms after the event, a mechanism capable of producing the condition, and treatment consistently directed to the same complaints. Headings such as “Negligence,” “Notice,” and “Causation” identify the issues, but a heading in a trial notebook does not give counsel questions to ask. Trial preparation should begin before discovery is served, not after it closes.

Use written discovery to separate
the facts

Written discovery is often the most economical way to establish facts, narrow disputes, identify evidence, and decide where more expensive discovery is justified. Requests for admission are particularly valuable because their function extends beyond information gathering.

California law permits requests concerning the genuineness of documents, facts, opinions relating to fact, and the application of law to fact. (Code Civ. Proc., § 2033.010.) As the California Supreme Court explained, requests for admission are designed not merely to gather information, but to resolve issues so they need not be tried. (Cembrook v. Superior Court (1961) 56 Cal.2d 423, 429.)

Consider a request asking the defendant to “Admit that YOU were negligent.” The request may have strategic value, but its likely response of “Deny” does little to develop the case. A more useful approach separates the conclusion into component facts:

• The defendant saw the plaintiff’s vehicle before impact.
• Nothing obstructed the defendant’s view.
• The roadway was dry.
• The defendant’s brakes were functioning properly.
• The defendant saw traffic slowing before impact.
• The defendant did not apply the brakes upon first seeing traffic slow.
• The defendant’s vehicle struck the rear of the plaintiff’s vehicle.

Now the defense must take a position on the facts that make up the liability theory. Drafting the requests this way also tracks the statutory requirement that each request be separately stated, complete in itself, and not compound, conjunctive, or disjunctive. (Code Civ. Proc., § 2033.060, subds. (c), (d) & (f).)

If the propositions are admitted, they become fixed points in the case. A matter admitted in response to a request for admission is conclusively established against the admitting party in the pending action unless the court permits withdrawal or amendment. (Code Civ. Proc., § 2033.410, subd. (a).) A clean admission can later support a motion, trial brief, examination, or closing argument. If a request is denied, counsel has still learned where the dispute lies. A denial to one compound request, such as a request to admit that the defendant saw traffic slowing, had sufficient time to stop, and could have avoided the collision, reveals almost nothing. Separate requests show whether the dispute concerns perception, timing, vehicle condition, braking, or something else. Precision in drafting produces precision in the case.

Treat denials as a roadmap

A denial can be nearly as useful as an admission if it directs the next step. Judicial Council Form Interrogatory 17.1, when used with requests for admission, requires the responding party to identify the facts, witnesses, and documents supporting each response that is not an unqualified admission.

Suppose a defendant denies that nothing obstructed the view of the plaintiff’s vehicle during the five seconds before impact. The denial raises concrete questions: What caused the obstruction? Where was it? How long did it interfere with the defendant’s view? Who else saw it? Is there a photograph, recording, or other evidence supporting the contention?

The answers may identify a legitimate issue requiring further discovery. Or they may show that the denial rests on little more than a possibility. That is useful information either way. Not every denial warrants a deposition, subpoena, or expert analysis; written discovery should help counsel decide which disputes deserve the investment.

Take depositions for the trial transcript

Depositions require substantial time and expense. An important deposition should do more than uncover information; it should produce testimony capable of later use at trial. Before taking a deposition, ask: If this witness gives a different answer at trial, what testimony should be available in the transcript?

The most useful answers usually concern facts rather than conclusions. Asking a defendant to agree that the defendant “failed to exercise reasonable care” invites disagreement and argument. Establishing the underlying facts is more productive:

“You saw traffic slowing ahead?”

“Nothing obstructed your view?”

“Your brakes were working?”

“Nothing prevented you from applying them?”

“You did not brake when you first saw traffic slowing?”

Each question proves one point, and each answer can stand alone. A question that combines several propositions such as, “You saw traffic slowing, had enough time to stop, and nothing prevented you from braking, correct?” surrenders that clarity. A “No” leaves counsel, and later the jury, uncertain about what the witness disputed.

The method also improves control. A narrow question generally calls for a narrow answer, while a compound question gives the witness room to qualify, explain, or reject only part of its premise. What matters is not a dramatic question, but a sequence the jury can follow.

California’s rules governing the use of depositions make this discipline especially important. Deposition testimony may be used to contradict or impeach the deponent, and an adverse party may use the deposition of a party for any purpose, subject to the rules of evidence. (Code Civ. Proc., § 2025.620, subds. (a) & (b).) A deposition transcript is not merely a record of discovery. It may become part of the trial.

Stop before the witness reclaims
the story

One of the most common mistakes in deposition and cross-examination is asking the ultimate question after obtaining the necessary admissions.

Suppose the defendant has admitted that the view was unobstructed, the road was dry, the brakes worked, traffic was visibly slowing, and the defendant did not brake in time. Counsel may be tempted to ask, “So this collision was preventable, correct?” In most cases, that is one question too many. It gives the witness an opportunity to reclaim the narrative: the plaintiff stopped suddenly, another vehicle blocked an escape route, there was nowhere to go, everything possible was done. Counsel had the facts and then invited an argument about the conclusion.

There was no need. Jurors can decide what the established facts mean. The adverse witness need not endorse counsel’s theory.

Draft questions that will age well

If testimony may be used months or years later, significant questions should be complete enough to stand on their own. Questions such as “And that didn’t happen?” or “That was before?” may be clear in the moment but frustratingly vague when counsel prepares an impeachment or deposition designation.

Instead of asking, “And nobody did?” ask, “No employee inspected Aisle 7 between 10:00 a.m. and the time plaintiff fell?” Instead of “That was before?” ask, “You received the customer complaint before plaintiff fell?”

A useful transcript should not require the reader to reconstruct what “that” meant on page 146. Complete questions create admissions counsel can actually use. For the lawyer who may handle the matter from intake through verdict, that small discipline saves time.

Test the expert’s foundation

Expert discovery calls for the same discipline. An opposing expert is paid to express an opinion, and trying to force the expert to abandon it often consumes time without producing useful testimony.

If a defense physician opines that the incident did not cause the plaintiff’s condition, asking the physician to concede causation will rarely help. Testing the opinion’s foundation may. Did the plaintiff report the same symptoms before the incident? Did the symptoms begin afterward? Is the mechanism capable of producing the condition? Are the findings consistent with the complaints? Was treatment directed to those complaints? Can the expert identify another traumatic event that produced the onset? Is there a pre-incident medical record documenting the same problem?

The expert may keep the ultimate opinion. But the cross-examination may establish no similar complaints before the incident, symptoms afterward, a medically plausible mechanism, and no identified alternative event. The jury then has a fair basis to evaluate the opinion.

Other disciplines call for the same scrutiny. With an accident reconstructionist, identify and test each assumed speed, distance, coefficient, and reaction time. With an economist or life care planner, examine the source for every rate, duration, and projected cost. Often the productive target is not the opinion itself, but the assumptions holding it up.

Match the discovery device to the task

Written discovery and depositions serve related but distinct purposes. California permits interrogatories asking whether a party makes a contention and seeking the facts, witnesses, and writings on which it is based. (Code Civ. Proc., § 2030.010, subd. (b).) That does not mean broad contention questions should simply be transferred into a deposition outline.

In Rifkind v. Superior Court (1994) 22 Cal.App.4th 1255, 1259-1263, the court distinguished permissible questions about facts within a deponent’s knowledge from deposition questions requiring a represented party to marshal all facts, witnesses, and documents supporting a legal contention. The latter are properly pursued through interrogatories.

The practical division is straightforward. Written discovery can define the opposing party’s contentions and identify the supporting universe of facts, witnesses, and documents. The deposition can then focus on what the witness personally saw, heard, received, said, knew, and did. Counsel, not the witness, can assemble those facts into the legal theory.

Plan the evidentiary path

An important fact uncovered in discovery is not necessarily an admissible fact at trial. As each significant point emerges, ask: How will this be proved?

If a video exists, who possesses it and who can authenticate it? If a photograph matters, who took it and when? If a text message is important, who sent or received it? If a business record contains the key fact, what foundation is required? If the proof depends on a third-party witness, can that person be located, subpoenaed, and presented? If an expert relies on particular data, how will the underlying evidence reach the jury? These questions should not wait until the eve of trial. By then, a witness may be unavailable, a foundation may be incomplete, or an exhibit may be difficult to authenticate. Every material discovery point should ultimately answer two questions: What does this prove and how will it be admitted?

That is the bridge from discovery to evidence.

Build the trial file as the case develops

One of the simplest and most valuable habits for a small litigation team is to organize admissions by the proposition they prove, not merely by their source. Instead of storing a useful answer only under “Defendant Deposition,” place it under a trial-oriented heading such as “Defendant Had a Clear Opportunity to Avoid the Collision.” Beneath that heading, record:

• Saw traffic slowing [Def. Depo at 42:7-12]
• View unobstructed [Def. Depo at 44:3-8]
• Roadway dry [Def’s resp to RFA No. 12]
• Brakes functioning [Def’s resp to RFA No. 15]
• Nothing prevented braking [Def. Depo at 51:4-9]
• Did not brake upon first seeing traffic slow [Def. Depo at 53:11-17]

The discovery file is now becoming the examination outline. It can also feed the statement of facts, motions in limine, exhibit list, settlement presentation, and closing argument. This practice avoids an expensive ritual familiar to trial lawyers: reopening thousands of pages shortly before trial to rediscover what was established months earlier. The lawyer who creates a citation-ready, issue-based record during discovery performs the work once and reuses it when time is most scarce.

Use technology without surrendering judgment

Transcript platforms, document-review systems, litigation databases, and artificial intelligence can help smaller firms manage discovery at a scale once reserved for larger teams. These tools can build chronologies, locate testimony on a disputed issue, compare accounts across witnesses, identify potential inconsistencies, and organize material by element of proof. They can also help turn a large record into a first-pass deposition or trial outline.

Technology earns its place when it leads counsel back to the source material needed to make a judgment. A search for every reference to notice, prior complaints, braking, symptom onset, or a particular exhibit can reduce hours of manual review. But every resulting quotation, characterization, citation, and page-and-line reference should be checked against the original.

Technology also introduces professional obligations. The State Bar of California’s 2026 guidance on generative AI emphasizes that lawyers retain responsibility for professional judgment and must understand a tool’s material capabilities and risks, protect client confidences, and review and verify AI-generated work. (State Bar of Cal., Standing Com. on Prof. Responsibility & Conduct, Practical Guidance for the Use of Generative Artificial Intelligence in the Practice of Law (2026), pp. 3-6, 9.) Before uploading discovery, counsel should understand how the provider stores, uses, and shares data, limit access appropriately, and avoid placing confidential information into a system lacking adequate safeguards.

The tools will change, but the lawyer’s job will not. Technology should leave counsel with more time to decide how the evidence fits the case and how best to present it.

Make every discovery dollar serve
the trial

A smaller budget calls for disciplined choices, not thin discovery.

Before noticing another deposition, serving another set of requests, or spending another hour with an expert, ask what trial value the effort is likely to produce. If an issue has already been established through admissions, another deposition may add little. If a corporate representative has established the governing policy, responsibility, and notice, multiple employee depositions may be cumulative. If an expert has supplied the concessions needed for cross-examination, another hour on marginal subjects may not materially improve the case.

Some of the most valuable admissions do not directly prove an element; they eliminate excuses. A dry roadway removes one explanation. Functioning brakes remove another. An unobstructed view removes a third. Proof that the defendant saw traffic slowing and faced nothing preventing a response narrows the field further. Six carefully selected facts may do more trial work than 20 peripheral ones.

Before spending the next discovery dollar, ask whether the proposed work will make the case materially stronger, more provable, or more efficient at trial.

Let the factfinder finish the job

Discovery is not an end in itself. Its purpose is not to accumulate paper, take depositions by habit, or persuade adverse witnesses to accept counsel’s theory. Its purpose is to develop the facts and evidence necessary to prove the case.

For solo and small firm trial lawyers, that requires focus at every stage. Written discovery should define disputes and secure usable admissions. Depositions should produce clear testimony for examination and impeachment. Expert discovery should expose the assumptions beneath an opinion. Evidence planning should ensure that important facts can actually be presented. Technology should make the record easier to search, organize, and use without replacing the lawyer’s judgment.

The adverse witness does not have to admit negligence. The corporate representative does not have to concede notice. The expert does not have to abandon the opinion. Those are conclusions, and adverse witnesses are rarely eager to provide them. Fortunately, they do not have to.

Begin with the conclusion the factfinder must reach and work backward to the facts that establish it. Secure those facts one question and one admission at a time. Preserve them in a form that can be used at trial, then organize them into the story the evidence tells.

Get the facts. Secure the admissions. And let the factfinder finish the job.

Christopher Kim is the founder and principal attorney of Kim Law APC, where he represents plaintiffs in personal injury and employment matters. He currently serves as secretary of the Orange County Korean American Bar Association and president of the OCKABA Foundation, a nonprofit committed to expanding access to justice and supporting future legal professionals. He earned his J.D. from Loyola Law School, Los Angeles, and his B.A. from the University of California, San Diego.