Skip to content
    Home » Articles » Decision Analysis for Corporate Counsel – Intro

    Decision Analysis for Corporate Counsel – Intro

    This is the first article in our series on Decision Analysis for Corporate Counsel. You can find the other articles here.

    In high-stakes disputes, too many legal teams still rely on instinct, vague descriptors, or a single “percentage chance of winning” when advising their clients. For GCs and board-level decision makers that’s no longer sufficient.

    Decision analysis is a technique long used in industries like energy, healthcare and finance that offers a structured, quantitative way to assess litigation risk. It’s not a silver bullet, but it can help to transform uncertainty into a map of possible outcomes with clear probabilities and financial impacts. This provides for informed decision making and good corporate governance for resolving litigation.

    What is Decision Analysis in Litigation?

    In the disputes context, decision analysis involves breaking a case into its key decision points and uncertainties, then mapping out the possible outcomes as a decision tree. Each branch reflects a specific scenario, such as prevailing on liability but with reduced damages, and assigns it a probability and a financial value. It is still legal counsel or experts who estimate the probabilities and damages (not AI, for example), but it provides a structured way to combine these estimates into a quantitative and visual picture of risk and likely outcomes.

    Unlike a gut-feel assessment, this approach forces counsel to articulate assumptions, account for multiple possible outcomes, and show how costs, interest, and counterclaims affect the final picture. The result is a visualization of the range of potential results and how likely they are, along with a risk-weighted average, or “expected value” (EV). This range of likely results and EV are very useful inputs to deciding a reasonable settlement value.

    As Marc B. Victor, who pioneered Litigation Risk Analysis in the 1970s, notes, “A decision tree is a picture of your case, showing the sequence of decisions and uncertainties that will be encountered and their possible consequences” [Victor, Interpreting a Decision Tree Analysis of a Lawsuit].

    Why the Traditional Approach Falls Short

    In surveys of corporate counsel, 70% report dissatisfaction with external lawyers’ ability to quantify litigation risk in actionable terms (ACC Litigation Risk Management Survey, 2021). Common problems include:

    • Single-point estimates (“70% chance of winning”) hide the fact that multiple outcomes are possible, each with very different financial implications. Winning and receiving full damages is very different from winning with a minimal damage award.
    • Vague language (“likely,” “strong case”) is highly prone to misinterpretation, even among trained professionals. Studies show that terms like “likely” are interpreted anywhere from 55% to 90% probability depending on the listener [Kent, 1964; Zonination, Perceptions of Probability Terms].
    • Cognitive biases skew judgment: research shows lawyers are on average 11–15 percentage points overconfident in predicting case outcomes [Goodman-Delahunty et al., 2010; Jeklic, 2023].
    • Omission of financial mechanics – Traditional assessments often fail to factor in the direct and indirect costs of litigation, possible cost-shifting, or the impact of interest. Decision analysis can integrate these elements into a model, weighting them alongside damages to provide a more complete financial picture.

    Traditional approaches often focus on a comprehensive legal analysis, analyzing the important legal arguments and their strengths and weaknesses. But they leave the decision maker to combine and weigh all the different factors themselves. This task is far from simple, even for experienced litigation managers and relatively simple cases.

    Why Hasn’t This Been Widely Adopted in Litigation?

    Despite its clear benefits, decision analysis faces cultural and practical barriers in the legal industry:

    • Tradition over transformation – Litigation has long been driven by narrative persuasion and precedent, not quantitative modelling. Many lawyers are unfamiliar with probability-based reasoning or view it as outside their professional skill set.
    • Perceived complexity – Some see decision trees as too technical or time-consuming, especially under pressure to act quickly. In reality, basic models can be built in minutes and refined over time. A lack of simple tools has contributed, giving the impression that lawyers need to do the math themselves.
    • Fear of accountability – Assigning numbers to probabilities feels risky; if the case outcome diverges from the model, counsel may fear criticism. Yet in corporate risk management generally, modelling is standard despite inherent uncertainty.
    • Lack of client demand – Until recently, few GCs have insisted on this level of rigour from outside counsel, so law firms have had little incentive to invest in capability.
    • Counter-productive for external counsel – Robust risk assessments can indicate that the best course is to settle earlier, which would reduce the potential billable work on a matter. This dynamic means firms may have limited incentive to adopt techniques that could reduce billing in the short-term.

    Similar barriers once slowed adoption in industries like energy and pharmaceuticals—until competitive advantage and investor expectations made quantitative risk analysis the norm.

    The Core Tool

    Decision Trees visually map out how the litigation is likely to proceed, with nodes for each decision and outcome. Probabilities are assigned based on legal judgment and evidence; financial values incorporate damages, costs, interest, and recoveries.

    Simple decision tree showing a liability node and damages node

    These models can be built with pen and paper, spreadsheets, or tools like Eperoto, TreeAge or PrecisionTree. Some can produce rich vizualisations of the likely financial outcomes, giving a clear, visual overview of litigation risk and the distribution of likely financial outcomes.

    What matters most is the quality of the inputs: clear assumptions, realistic probabilities, and good cost/quantum estimates. But even early on when these figures are broad guesstimates, this process still gives a much better ballpark range than simply a best- vs worst-case assessment. 

    Visualizations show the results of the modelling in different ways. Taking the simple tree above and applying interest & legal costs would give results like below. They can also show the case from the other party’s perspective, which can be very insightful, e.g. in this case the expected value is only $35,000 for the claimant despite it being -$640,000 for the defendant. Such insights could lead to productive settlement discussions for both parties.

    The Payoff for General Counsel

    1. Clarity for Decision-Makers – An EV calculation supported by a probability distribution tells the board not just “we think we’ll win” but, for example, “there’s a 20% chance of losing $50m and a 10% chance of breaking even.” This aligns with corporate risk management standards in other domains.
    2. Better Settlement Decisions – By modelling both sides’ positions, you can estimate the Zone of Possible Agreement (ZOPA). For example, if your EV is +$3M but the other side’s EV is –$1M, there’s room to settle anywhere between those numbers. Clearly other factors come in to play, but the model provides a reference point to considering the other factors. For example, a case on paper might be a $1M liability but you may settle it at $1.5M, effectively paying a $500,000 premium to avoid the non-financial repercussions like risk of reputation damage.
    3. Budgeting and Reserves – Litigation reserves are a financial reporting requirement for many companies. Decision analysis provides a defendable basis for these numbers, which auditors increasingly expect [PwC, Legal Contingencies and Disclosures, 2022].
    4. Bias Mitigation – Structured models counter overconfidence and framing effects by making assumptions explicit and testable. Sensitivity analysis (“what we drop the probability of liability from 75% to 50%?”) can show how which points are critical, driving litigation and settlement strategy.

    Real-World Adoption and Success Stories

    Decision analysis is gaining traction:

    • Insurance – Underwriters have for decades used probability-weighted loss modelling to set premiums and reserves for litigation risk [Celona, 2016]. HSF Kramer’s biggest tree was a $1.3B insurance claim, where the analysis aided the board’s settlement strategy [HSF Kramer, 2023].
    • Energy & Infrastructure – ConocoPhillips’ legal department integrated LRA with outside counsel for major disputes, reporting faster settlements and better prioritisation of evidence [Victor, ACC Annual Meeting Panel, 2018].
    • Law Firms – A small but significant number of firms, from AmLaw 100 to litigation boutiques in USA, Canada & Europe, use decision analysis tools to deliver quantitative risk assessments to clients. [Refer articles below]
    • Technology Sector – A global software company used decision analysis in a $200m IP dispute to model likely claim outcomes; the model revealed a high-value counterclaim risk that prompted an early settlement saving an estimated $50m in potential loss.
    • Mediators – Some commercial dispute mediators use decision analysis to show sides the value of their case, as a way to bring them closer together. For example, Podrebarac Mediation uses it in the pre-mediation discussions for every dispute.

    These examples show that the barrier is not whether it works, but whether clients ask for it.

    The Data-Driven Future

    Businesses increasingly expect legal to operate with the same quantitative discipline as other functions. In a 2020 survey, 82% of CFOs said they wanted legal risk presented in numerical form to support enterprise risk dashboards (EY Law, Reimagining the Legal Function).

    Decision analysis won’t eliminate uncertainty, but it makes the uncertainty visible and manageable. As statistician George Box famously put it:

    “All models are wrong, but some are useful.”

    For GCs, a well-constructed litigation decision analysis is one of the most useful models you can have, helping you decide whether to fight, settle, or change strategy with eyes wide open.

    At Eperoto we love to talk decision analysis 🙌. Get in touch if you’d like to discuss anything from the articles or you’d like to learn more about Eperoto.

    References and Further Reading