Over the last few years, courts have paid more attention to the early stages of litigation and the steps litigants must take to ensure that their documents are available for discovery. Historically, that meant making sure the parties had collected things from their file cabinets. Today, the exercise is a more complicated combination of document retention policies, litigation holds and coordination with employees and others.

At a high level, a company’s document retention policy should retain only e-mails with business record significance, to avoid the dangers associated with disclosing damaging information that might include personal communications. Such a system should include “litigation holds” to prevent destruction of documents related to ongoing or anticipated litigation. That general pronouncement is often of little use in real world cases, as litigants in recent cases have discovered.

Today, counsel must be able to effectively implement these litigation holds. The basics include issuing a hold notice, identifying the right custodians (or key players), coordinating data identification and preservation, monitoring the implementation of the hold, and then releasing the hold. Issuing that hold is the critical first step in satisfying the preservation obligation. Once that hold notice has issued, the company must suspend its document retention policies to prevent discarding data and notify employees to refrain from deleting e-mails or other computer documents.

The dangers in this area have proven quite significant, and the challenge to determine proper scope of litigation holds is hardly new, although it has changed over time. For years, companies have understood that failing to implement and monitor document retention programs effectively can result in severe consequences, even without intentional wrongdoing. In the 2008 California case Keithley v. The HomeStore.com Inc., the court sanctioned the defendant, HomeStore.com, more than $1.4 million and entered multiple adverse inferences due to its “lackadaisical attitude with respect to discovery.” The Keithley defendant conceded that no litigation hold policy existed during the relevant time periods underlying the claims. And as recently as October 2012, the federal government was sanctioned in United States ex rel. Baker v. Community Health Systems, Inc., for failing to issue a timely and adequate litigation hold.

Knowing when a litigation threat is “real” and requires a hold is subjective — and much easier to determine in retrospect. This decision requires analyzing overall claim risk, the scope of claim knowledge within the company, and the risk of data destruction absent a legal hold. No single (or easy) answer exists. The economic situation facing most companies, and the expense associated with establishing and maintaining a litigation hold, creates significant pressure to respond only when the litigation threat is absolutely concrete.

As an initial matter, a potential litigant should assess the valuation of the potential risk associated with the claim in question, in other words, the cost to meet the demand. Part of that claim assessment must consider the cost and scope of the litigation hold, including actual dollars and employee hours lost. The potential cost of not instituting a legal hold when one is needed must be part of the calculus as well, including the risk of potential sanctions associated with failing to institute a proper legal hold. For each of these, proper documentation of the decisions (and the inputs into these decisions) could be the difference between sanctions and justification before the court.

Another factor to consider is the knowledge base within the company concerning the potential claim. If only one or two people have any relevant knowledge or data, it may be relatively inexpensive to preserve the data without too much effort to conduct a more thorough analysis of the claim. However, if a large number of employees have knowledge and data that would be caught up in the litigation hold, it might be necessary to conduct a more detailed analysis of whether litigation is “reasonably anticipated.”

Yet, all is not lost. Courts are generally sympathetic to these challenges as companies attempt to determine when the preservation obligation arises. For example, in a non-ediscovery case, the Texas Supreme Court held that a producing party does not abuse the discovery process (by failing to produce) unless the opposing party proves that the non-producing party had a duty to preserve the evidence at issue. In Wal-Mart Stores Inc. v. Johnson, a Wal-Mart employee stocking merchandise accidentally knocked a decorative reindeer onto Johnson’s head and arm. The court held that nothing in the investigation surrounding the accident gave Wal-Mart notice of the Johnson’s intent to sue or that there was a substantial chance that the Johnson would sue. Thus, Wal-Mart had no duty to preserve the reindeer because Wal-Mart did not anticipate litigation.

The bottom line is that what triggers a litigation hold is highly case specific. No formal trigger is necessary, such as the filing of a complaint. In the end, if an issue arises that could end up as some type of litigation; further case-specific analysis is required. Such analysis initially should include considerations of risk and cost among other factors. And the results of those initial determinations will necessarily impact how much additional research is needed about the reality of the litigation risk. That said, no matter how difficult the analysis may be, failing to undertake these steps is a far more treacherous path.