Outlier deals are opportunities whose value or cycle length sits far from the rest of the distribution. Reporting that averages across them without showing the spread produces a number that describes no actual deal, and planning built on that number sets targets nobody can hit.
The problem is structural rather than statistical. B2B SaaS pipelines are right-skewed. A small number of large deals carry a large share of the total, so the mean sits above the median in almost every pipeline.
The gap that gives it away
ORM points to a specific version of this. A pipeline can carry an average deal size of $80,000 while closed-won deals average $40,000. That gap has a direct reading: deals are entering the pipeline at values they do not close at, and every forecast weighted on pipeline amount inherits the error.
The gap has two candidate sources. Deals get discounted on the way to signature, and the size bands may convert at different rates. Split win rate by size band in your own data to see which is driving it.
| What to compare | What a gap tells you |
|---|---|
| Pipeline mean vs. pipeline median | Concentration in a few large opportunities |
| Pipeline mean vs. closed-won mean | Deals close below their entered value |
| Cycle mean vs. cycle median | A long tail of deals that never resolve |
Handling them in reports
Show the median next to the mean on any deal size or cycle length figure. That single change makes skew visible without removing anything, and it stops the average from being read as typical.
Segment by size band rather than filtering. A pipeline split into bands with counts and win rates per band tells you where the concentration sits and how each band actually converts. A filtered report tells you what the pipeline looks like with the interesting part removed.
Flag concentration explicitly. When the top three opportunities carry a large share of the quarter, that is a risk statement, and it belongs on the report as a number rather than as a footnote someone notices in week eleven.
How forecasting should treat them
Applying one close rate and one expected value across a whole pipeline assumes every deal behaves the same way. Outliers are the proof that it does not. ORM groups each opportunity with a machine learning model and predicts a separate close curve for each group, with curves running from 1 to 80 weeks, most of the expectation landing before week 12, and very few groups carrying expectation past 52 weeks. Grouping first means a large slow deal is scored against deals that behave like it rather than against the pipeline average.
That is also why an unweighted total misleads and why crude probability weighting misleads differently. We covered the mechanics in weighted pipeline. The same logic sits under sales forecasting generally: the unit of prediction is the deal group, not the pipeline.
Watch outliers for deal slippage most closely. Large deals have more approval layers, and a pushed close date on one of them moves the quarter by itself.
Frequently Asked Questions
How do you identify an outlier deal?
Compare the median to the mean. If the mean deal size sits well above the median, a small number of large deals is pulling the average up, and those deals are the outliers. The same check works on cycle length, where a long right tail drags the average duration past what a typical deal takes.
Should you exclude outlier deals from sales reports?
Report both figures rather than choosing. Show the total including every deal, because that is the revenue, and show the median alongside the mean so the shape of the distribution stays visible. Excluding a real deal from a revenue report understates results, and averaging it in without context misleads planning.
Do outlier deals break forecast models?
They break naive models that apply one conversion rate and one expected value to every opportunity. A large deal has a different close probability and a different cycle than a typical one. Grouping opportunities by their behavior before applying a close curve handles this correctly.
What is the difference between an outlier and a bad forecast entry?
An outlier is a real deal with unusual size or duration. A bad entry is a placeholder amount or a close date nobody believes. Outliers should be modeled separately. Bad entries should be corrected at the source, because they inflate coverage while contributing nothing.
Put these metrics to work
ORM builds custom revenue forecast models that turn concepts like outlier deals in sales reporting into prescriptive action for your team.
Schedule a Demo