Matrix metrics – the seven calculation modes
The metric decides what the color intensity of a cell represents. All seven metrics work on the same data, they only show it from different angles.
For all metrics with a color: the fuller the color, the higher the value. In the light theme a cell gets darker, in the dark theme it gets lighter.
When you hover over a cell, the tooltip shows the metric values for that cell. The color matches the value of the selected metric.
Row share
Default. How much of the activity in this row falls on this column?
- Formula:
cell value / row total - Color: fuller = higher share of the row. The shares of a row add up to 100 %.
- Use: highlights where a source or topic had its focus. A source with five posts and one with a hundred look alike when the distribution is similar. If a row spreads over many columns, its cells stay pale.
Column share
How much of the activity in this column falls on this row?
- Formula:
cell value / column total - Color: fuller = higher share of the column. The shares of a column add up to 100 %.
- Use: shows what a month or a topic is made of. Answers “who contributed the most Updates in March?”. In grouped views only the parent rows count toward the column total.
Total share
How much of the visible grand total falls on this cell?
- Formula:
cell value / grand total - Color: fuller = larger share of the whole
- Use: shows where the most is happening overall. Small activity fades in color, the hotspots stand out.
Category share
Available only in views with category rows (for example Sources (grouped by categories) over time).
- Formula:
cell value / value of the parent category in the same column - Color: fuller = larger share of the category
- Use: answers “which source dominates inside its category?”. On flat views it would coincide with Total share, there it is not selectable.
Count
The raw number, no share.
- Formula:
cell value - Color: fuller = more Updates. The cell with the most Updates is fully saturated.
- Use: when frequency matters, not distribution. A post is a post, regardless of what happens around it.
Compared to reference period
Compares the current value with the same cell value from a reference period.
- Formula:
(cell value − reference cell value) / reference cell value, that is, the relative change - Color: green = more than in the reference period, red = less, neutral = similar
- Use: trend question. Where is there more going on today than three months ago? Where has a source gone quieter?
The reference period is picked automatically by default: same length, immediately before the current period. Under Filters, at Reference period, you can untick Automatic and choose your own Reference range.
Market comparison
Only available when no option is ticked under Ownership in Filters, that is, when your own and competitor sources count together.
- Formula: normalized comparison
own activity per own sourcevs.competitor activity per competitor source - Color: green = your sources post more per source than the market, red = the market posts more per source
- Use: answers “are we more visible than our competitors, in proportion?”. Cancels out the effect that you have fewer of your own sources than the market has competitor sources.
Which metric when
| Goal | Recommended metric |
|---|---|
| Where did whom put the focus? | Row share |
| Who contributed the most in a period or topic? | Column share |
| Where is the most going on in the market? | Total share |
| Who dominates inside their category? | Category share |
| How often — plain numbers | Count |
| Where is activity shifting? | Compared to reference period |
| Are we more visible than the market? | Market comparison |
Cells with value zero
Zero-value cells are rendered differently depending on the metric:
- Share metrics (Row, Column, Total and Category share): without color
- Count: without color
- Δ metrics: neutral for similar values, colored for larger deviations
A completely empty row (a source with no activity in the period) disappears from the Matrix. If the whole Matrix is empty, take a look under Filters: usually the time range is too tight or a filter is too restrictive.