The Qualcomm order offers useful lessons for anyone involved in big-case litigation. Subject to the important qualification that this list might be different if we knew the privileged information we do not know, here are my candidates:
1. Division of labor matters. As clients rationally demand lower costs for large-scale discovery, it will become more and more common for someone other than an employee of outside counsel to collect documents and review them for production. That might take the form of sourcing outside both the client and the firm or, as apparently was the case with Qualcomm (given what we know), substantial insourcing or document collection by the client. Where outside counsel are not directly responsible for discovery, they must take steps to protect themselves.
2. Responsibility should follow authority. Firms that do not have responsibility for the actual collection and/or selection of documents should not sign discovery responses relating to document production or make representations about the completeness of discovery. If the client insources, as Qualcomm seems to have done, or outsources to a firm it selects, then it should take responsibility for the completeness of discovery. To the extent outside counsel can do so, they should insist on that division.
If a client demands that a lawyer outsource document production to a cheap third-party provider, the lawyer should either demand that the client select and deal with production or choose a firm it trusts enough to treat as a sub-agent—i.e., one for whom it is responsible. If neither option is available, it shouldn't outsource.
3. If labor is divided, demand an advance privilege waiver for communications relevant to any discovery disputes. I have a lot of sympathy for the inability of the Day, Casebeer lawyers fully to defend themselves. If you don't want to be in their position, ask for a waiver up front. (Off the top of my head I can't recall authority upholding such waivers, but I would uphold them by analogy to advance conflict waivers, so long as the scope was reasonable and set out plainly in advance.)
4. Don’t trust standard procedures or formal systems. (I.e., you have to pull discovery; don't trust employees to push it.)
I’m guessing here, but suppose something like this happened. Suppose Qualcomm had a clear policy that said all business-related material had to be backed up to Qualcomm servers, or preserved in some other way under Qualcomm's control. Under the hypothetical policy, there should never be a situation where an individual employee had Qualcomm-related data on his or her laptop that was not also centrally preserved in a way that the in-house legal department could access it without physically chasing down an employee. Suppose further that Qualcomm sent reminders to people who might have litigation-relevant data, telling them to back up (or centrally deposit) whatever they have that might be relevant.
Such a policy would make a lot of sense. It would make discovery cheaper and more rational. I can (almost) see an in-house person believing in it. But it’s not realistic. Individual employees do strange things. They pile up e-mails they have ignored and forget about them. They forget about back-ups. They think they have got rid of stuff they have. They delete reminder e-mails as annoying stuff from paranoid lawyers. They're human.
To do discovery properly, you have to (figuratively) grab all the relevant people, walk into their offices, open all their drawers, pick them up, turn them upside down, shake them, and then go after their computers, especially laptops, which often serve hybrid functions mixing personal and business data. Don't trust, verify.
Annoying, expensive, and often pointless? Yes. Necessary? Read the opinion.
5. Try to de-bias yourself. When Day, Casebeer lawyers discovered 20-odd e-mails in the middle of trial, they at first declined to produce them on the ground that they were not relevant. They missed the point: The fact that the e-mails existed was relevant even if the content was not.
I don't think the initial decision to withhold necessarily implies bad faith. It might be the understandable reaction of people so focused on the merits of the trial that their perspective had narrowed too much. It might also be a by-product of someone who had spent so much time with their own arguments that they talked themselves into believing them, thus missing the importance a neutral observer might attribute to the document.
How do you fight natural psychological phenomena? It's not easy. One suggestion would be to appoint one member of the team a devil's advocate. Give them the documents and ask them to write a couple of paragraphs making the most damaging argument they can think of as to why the documents have to be produced. Another is to write a list of conditions that would have to exist for the document to be unresponsive, and then asks if all the necessary conditions do in fact exist.
6. Err in favor of production. There are four types of documents: irrelevant, relevant and helpful, relevant and neutral, and relevant and harmful. You don't have to produce type one, but by definition there is no harm to doing so. (Sure, a document might make the other side think of a new theory, but that is a kind of relevance; people can amend pleadings, remember.) You want to produce type 2. You have to produce type three and have no reason not to. You have to produce type four even though you don't want to. When you look at the possible choices, in marginal cases it is better to err in favor of production than in favor of withholding.
7. McGowan's inverse plausibility theorem: The longer an explanation is, the less likely it is true. Application: The more you have to explain a decision to withhold, the less likely a court is to buy it.
8. McGowan's inverse incentive theorem: The more you don't want to produce a document, the more important it is to produce it.
9. Be prepared to walk away. That is one of the five rules of survival I teach in PR. Note 10 of the opinion goes to this point. A relevant part:
"If Leung was unable to get Qualcomm to conduct the type of search he deemed necessary to verify the adequacy of the document search and production, then he should have obtained the assistance of supervising or senior attorneys. If Mammen and Batchelder were unable to get Qualcomm to conduct a competent and thorough document search, they should have withdrawn from the case or taken other action to ensure production of the evidence. See The State Bar of California, Rules of Professional Conduct , Rule 5-220 (a lawyer shall not suppress evidence that the lawyer or the lawyer’s client has a legal obligation to reveal); Rule 3-700 (a lawyer shall withdraw from employment if the lawyer knows or should know that continued employment will result in a violation of these rules or the client insists that the lawyer pursue a course of conduct prohibited under these rules)."
10. Associates: You have to look out for yourselves. I know this is totally unrealistic advice given large firm structures, but there are two associates in the sanctioned group, one of whom, Bier, did very little and tried to kick the problem to a senior lawyer. If you feel strongly about production, fight for it. If you don't, and you wind up in an opinion like this, at least you'll have the cold comfort of knowing it was your own decision and not passive acquiescence in somebody else's call.
11. If you mess up, fess up. Qualcomm defended its document review even after it became clear that it had failed to find and produce documents the court thought were not only relevant, but vitally important. In hindsight, that defense made Qualcomm appear churlish at best and corrupt at worst. It was the sort of defense that makes a judge think the lawyer or client just doesn't get it, and needs a corrective 2 x 4 across the brow. Or close to $9 million in a fee award. People make mistakes. It happens. When it does, it's better to admit it, apologize, and try to make it right. (Funny how things your mother told you work so well in discovery.)