Estimating impact

The chance tells you how often. The impact tells you how much it hurts when it does. For a legal problem, impact is rarely only the amount in the claim. Build the list in groups:

• Direct money: damages or a settlement, the other side’s costs if you must pay them, interest, fines.
• Your own costs: legal and expert fees, court costs, and the time of your own people.
• Knock-on effects: a customer lost, a higher insurance premium, a delayed project, a stopped sale.
• Things that resist a price: reputation, morale, a regulator’s attention. Give them a rough range and say so, rather than leaving them out.

For every item, ask for three figures: a low (things go well), a most likely, and a high (things go badly). Then work out a weighted average that leans toward the likely figure: (low + 4 x likely + high) divided by 6. This is the three-point estimate, used in project planning for decades.

Example: legal fees to trial with a low of $25,000, a likely $40,000 and a high $70,000 give an expected figure of $42,500. It sits a little above the likely value because the high end is further away than the low end. That is the point: costs are rarely symmetrical.

• Do not add all the highs. Every cost reaching its high at once is rare. The total of the highs is a ceiling, not a forecast.
• Use a planning range. If the items move independently, the total’s spread is smaller than the sum of the spreads, because some items come in high while others come in low. The lab shows a range that covers about four outcomes in five.
• But watch for linked costs. One bad turn, such as a damaging document, can push several items up together. If so, widen the range.
• Never count a loss twice. “Lost sales” and “lower profit” for the same deal are one loss. Prefer lost profit to lost revenue.

In order of preference:

1. The firm’s or client’s own records of similar past matters.
2. Published figures and benchmarks, checked for date and relevance.
3. Estimates from people who have done the work, ideally asked separately and then compared.
4. A clearly labelled guess, with a deliberately wide range.

When you know little, widen the range. A narrow range should be earned by evidence.

• Listing only the claim amount and forgetting costs, time and lost business.
• Giving one figure per item, so the range of the total cannot be seen.
• Leaving out management time because no extra salary is paid.
• Ignoring delay: a payment a year late has a cost.
• Counting the same loss under two names.