Module 6
Benchmarks and outliers
By the end of this module you can summarise a field with its median and quartiles, say where a proposed term sits in your portfolio, flag extreme values with a fence rule, and treat each one as a question to check rather than an error to delete.
Part 1. The lecture
Why benchmark
“Is 60 days a normal payment term for us?” is a question a portfolio can answer. A benchmark shows where a clause you are being offered, or a term you want to propose, falls compared with what you have actually agreed before. It turns “that feels long” into “that would be longer than 90% of our contracts”.
The middle of the data
- Median: the middle value when the list is sorted. Half the contracts are above it and half below. For payment terms of 15, 30, 30, 45, 45, 60, 90 it is 45.
- Average (mean): the sum divided by the count. It is pulled by extreme values, so for skewed data such as contract values it can describe nobody’s actual contract.
- Quartiles: Q1 has a quarter of the values at or below it, Q3 has three quarters. The span from Q1 to Q3, the IQR, holds the middle half of the contracts.
- Percentile rank: the share of contracts at or below a given value. A value at the 90th percentile is higher than nine in ten.
Outliers are questions
A common rule flags any value more than 1.5 × IQR beyond the quartiles: the fences are Q1 − 1.5 × IQR and Q3 + 1.5 × IQR. A flagged value is unusual. It is not necessarily wrong. A payment term of 450 days is probably a typing slip for 45. A liability cap of 12 times fees may be a real deal done for a strategic customer. A $2.4 million contract may be a genuine anchor client.
The honest response to an outlier is the same each time: check the source contract, then correct it or keep it with a note. Deleting a value because it spoils the picture hides either an error or a real deal your reader ought to know about.
What a benchmark can and cannot tell you
- It describes your own portfolio, which reflects your past bargaining position. It is not a survey of the market.
- Compare like with like. A lease and a software licence have different norms.
- State the count. A benchmark from eight contracts moves a long way when a ninth arrives.
- Say what was left out. If blanks are excluded, say how many.
Part 2. Lab: Benchmark Finder
Pick a field. The chart shows how the contracts are spread, and the summary gives the middle and the range of the typical values. Then test a position against the portfolio, and decide what to do with the extreme values.
Blank payment terms are left out. 57 of 60 contracts have a value.
- 01
- 159
- 3019
- 4510
- 6010
- 907
- 4501
Bars in red lie outside the fences below. The unit is days.
Median
30
Average
49.74
Middle half
30 to 60
Fences (1.5 × IQR)
0 to 105
86% of the contracts with this field have a value at or below 75. That is within the usual range.
The extreme values
C-017 (Underhill Packaging Ltd): 450 days
Part 3. Check your understanding
Ten questions drawn from a larger pool. In this pilot your score is shown but not saved.
This course teaches how to turn contracts into reliable data, using invented contracts, fictional parties and made-up figures. No real contracts or tools are used, and nothing here is legal advice. Never paste a real client’s confidential contracts into a public tool.
