The Benchmark view is the workspace’s “freeform exploration” tool. Pick any two numeric fields, plot one against the other, optionally colour by a categorical field, and look for patterns.
What this page answers
Are our high performers also our high earners? Does utilisation correlate with sentiment? Are certain departments clustered in one corner of the financial / performance map?
What’s on the page
The scatter plot
A single large interactive scatter chart. Each point is an employee, with:
X position from the field you pick on the X axis
Y position from the field you pick on the Y axis
Colour from the categorical field you pick for “Colour by”
Hover a point to see the employee’s name, the X and Y values, and the Colour By value (Performance Level by default). Click a point to open the employee’s detail popup.
The chart shows a subtitle with how many employees are plotted vs the total.
Configuring the axes
In the sidebar:
X Axis: a numeric field. Defaults to the first financial field if one exists, otherwise Sentiment Score.
Y Axis: a numeric field. Defaults to Sentiment Score.
Colour By: a categorical field (or None). Defaults to AI Performance Level.
The numeric fields available for the axes come from two sources:
Sentiment Score: Compass’s AI-extracted sentiment value
Financial fields: from the financial data imported into your firm (revenue, billings, utilisation, etc.)
Colour By fields include AI Performance Level, Department, and any rating dimension from the review.
Overlays
The sidebar also offers four overlay toggles:
Trend Line: a best-fit line across all plotted points, with the R² shown in the legend
Quadrants: shade the chart into four quadrants, split at the average of each axis
Averages: horizontal and vertical lines at the mean of each axis
Labels: show employee names on each point
Where the data comes from
This is the most “data-direct” page in PI — there’s relatively little AI happening on the page itself. The chart pulls:
Financial values from your firm’s imported financial data
AI fields from review extraction (sentiment, quality, etc.)
Rating fields from review responses
The AI’s job here is upstream: it produced the sentiment, quality, and performance scores during extraction. This view just plots them.
How to read it
Strong upward slope: positive correlation. High X tends to come with high Y.
No discernible pattern: whatever you’re plotting on X has no obvious relationship to Y in this cohort.
Clusters by colour: if colouring by department, distinct clusters mean different departments behave differently. A single tight cluster means departments are similar on these dimensions.
Outliers: points sitting away from the cluster are usually the most interesting people: the high performer on low revenue, the high earner with poor sentiment.
When to use it
Benchmark is the answer to specific “I wonder if…” questions:
“Is leadership rating correlated with team revenue?”
“Do under-raters bill more than over-raters?”
“Where’s the practice that’s high billing, low sentiment?”
You shouldn’t open Benchmark with no question in mind — there are too many combinations of fields for that to be productive. Come with a hypothesis.
Tips
Try Performance Level for colour first. Then change the X and Y axes and watch how the colour clusters move — it’s a fast way to find which financial fields correlate with performance.
Quadrants + Averages is the most useful overlay combo for talent-grid-style thinking.
Use hover instead of the labels overlay. The labels overlay clutters the chart quickly; leave it off and hover to identify points.
Empty page? If your firm doesn’t have imported financial data, you’ll see a notice and the axis options are limited to Sentiment Score. Import financial data to unlock the financial × performance combinations.